Robert Sobot

Engineering
Mathematics
by Example
Engineering Mathematics by Example
Robert Sobot

Engineering
Mathematics by
Example

123
Robert Sobot
École nationale supérieure de l’électronique et de
ses applications
Paris, France

ISBN 978-3-030-79544-3 ISBN 978-3-030-79545-0 (eBook)

https://doi.org/10.1007/978-3-030-79545-0

© Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the
whole or part of the material is concerned, specifically the rights of reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher
nor the authors or the editors give a warranty, expressed or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To my math teacher Mr. Miloš Belušević
Preface

This tutorial book resulted from my lecture notes developed for undergrad-
uate engineering courses in mathematics that I have been teaching over the
last several years at l’École Nationale Supérieure de l’Électronique et de ses
Applications (ENSEA), Cergy in Val d’Oise department, France.
My main inspiration to write this tutorial-type collection of solved prob-
lems came from my students who would often ask, “How do I solve this? It is
impossible to find the solution”, while struggling to logically connect all the
little steps and techniques that are required to combine before reaching the
solution. In the traditional classical school system, mathematics used to be
thought with the help of systematically organized volumes of problems that
help us develop “the way of thinking”, in other words, to learn how to apply
the abstract mathematical concepts to everyday engineering problems. The
same is true for music; it is also true for mathematics that in order to reach
high level of competence, one must put daily effort in studying typical forms
over a long period of time.
In this tutorial book, I choose to give not only the complete solutions to
the given problems but also guided hints to techniques being used currently.
Therefore, problems presented in this book do not provide review of the
rigorous mathematical theory, instead the theoretical background is assumed,
while this set of classic problems provides a playground to play and to adopt
some of the main problem-solving techniques.
The intended audience of this book are primarily undergraduate students
in science and engineering. At the same time, my hope is that students of
mathematics at any level will find this book a useful source of practical
problems for practice.

Île-de-France, France Robert Sobot

November 30, 2020

vii
Acknowledgements

I would like to acknowledge all those classic wonderful collections of

mathematical problems that I grew up with and used as the source of my
knowledge, and to say thank you to their authors for providing me with
thousands of problems to work on. Specifically, I would like to acknowledge
classic collections written in ex-Yugoslavian, Russian, English and French
languages, some are listed in the bibliography. Hence, I do want to acknowl-
edge their contributions, which are clearly visible throughout this book and
are now being passed on to my readers.
I would like to thank all of my former and current students who I had
opportunity to tutor in mathematics since my student years. Their relentless
stream of questions posed with unconstrained curiosity forced me to open my
mind, to broaden my horizons and to improve my own understandings.
I want to acknowledge Mr. Allen Sobot for reviewing and verifying some
of the problems, for very useful discussions and suggestions, as well as his
work on technical redaction of this book.
My sincere gratitude goes to my publisher and editors for their support
and making this book possible.

ix
Contents

Part I Algebra and Analysis

1 Basic Number Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.1 Basic Number Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3 Linear Equations and Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Exponential and Logarithmic Functions . . . . . . . . . . . . . . . . . . . . 51
4.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5 Trigonometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6 Complex Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
7 Linear Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
8.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
9 Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
9.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
10 Function Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
10.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

xi
xii Contents

11 Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
11.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
12 Multivariable Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
12.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
13 Complex Functions in Engineering and Science . . . . . . . . . . . . . 295
13.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
14 Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
14.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

Part II Mathematics for Signal Processing

15 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
15.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
16 Special Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
16.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
17 Convolution Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
17.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
18 Discrete Convolution Sum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
18.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
19 Fourier Transformation Integral . . . . . . . . . . . . . . . . . . . . . . . . . . 413
19.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
20 Discrete Fourier Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 437
20.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
Acronyms

f (x), g(x) Functions of x

f  (x), g  (x) First derivatives of functions of f (x), g(x)
f (g(x)) Composite function
sin(x) Sine of x
cos(x) Cosine of x
tan(x) Tangent of x
arctan(x) Arctangent of x
sinc (t) Sine cardinal of t
log(x) Logarithm of x, base 10
ln(x) Natural logarithm of x, base e
logn (x) Logarithm of x, base n
x∗y Convolution product
i Imaginary unit, i 2 = −1
j Imaginary unit, j 2 = −1
(z) Imaginary part of a complex number z
(z) Real part of a complex number z
F(x(t)) Fourier transform of x(t)
X(ω) Fourier transform of x(t)
F
−
→ Apply property of Fourier transform
F−1 (x(ω)) Inverse Fourier transform of x(ω)
DF T Discrete Fourier transform
FFT Fast Fourier transform
δ(x) Dirac distribution
Λ(x) Triangular distribution
Π(x) Square (rectangular) distribution
XT (t) Dirac comb whose period is T
Γ (t) Step function of t
u(t) Step function t
r(t) Ramp function t
sign (t) Sign function t
(a, b) Pair of numbers
(a, b) Open interval
≡ Equivalence
{a1 , a2 , a3 , . . . } List of elements, vector, series
{an } List of elements, vector, series
a∈C Number a is included in the set of complex numbers
∀x For all values of x

xiii
xiv Acronyms

Angle, argument
R The set of real numbers
C The set of complex numbers
N The set of natural numbers
Q The set of rational numbers
⇒ It follows that
dy
dx
Derivative of y relative to x
ux Partial derivative of u(x, y, z, . . . ) relative to x
ux,y Second partial derivative of u(x, y, z, . . . ), relative to x
then to y
z∗ Complex conjugate of number z
|x| Absolute value of x
Am,n Matrix A whose size is m × n
In The identity square matrix whose size is n × n
|A| Determinant of matrix A
AT Transpose of matrix A
Δ Difference between two variables
Δ Main determinant of a matrix
Δx Cramer’s sub–determinant relative to variable x
a Vector a
a Vector a
∞
i=0 ai Sum of elements {a0 , a1 , . . . , a∞ }
dB Decibel
dBm Decibel normalized to “10−3 ”, i.e. “milli”
H (j x) Transfer function
 Part I
Algebra and Analysis
 Basic Number Theory
1

Important to Know

Assuming a ∈ R and a = 0 basic number operations with exponents and radicals obey the following
definitions and rules

a n+1 = a n · a
a0 = 1
def

a n a m = a m+n
a1 = a
def
 n m
a = an m
1
a −n =
def

an (a b)n = a n bn
n def √
a m = an
m
an
= a n a −m = a n−m
 am
x ; (x ≥ 0)  a n
|x| =
def
an
−x ; (x < 0) = n
b b

1.1 Basic Number Operations

1.1 Basic Calculations

1. Calculate the following numbers:

 −2  −2  −1
1 1 1
(a) (−2)−3 (b) (c) +
2 2 2
2. Calculate the following numbers:
 −3
(−2)−3 − (−3)−2 2
(a)
−1
 (−4)  3 
(b) (−2)−1 + (−3)−1 ÷ (−3)−1 − (−6)−1

© Springer Nature Switzerland AG 2021 3

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_1
4 1 Basic Number Theory

(c) −4 × 104 + 2.5 × 105

(d) 0.5−1 + 0.25−2 + 0.125−3 + 0.0625−4

3. Calculate the following numbers:

 0
3
(a) 20 (b) −
8
√
(c) ( 3 + π − 1)0 (d) (cos x)0 , (|x| < π/2)
 
3 π 0 4 7π 0
(e) + − −
2 6 27 9
4. Simplify the following rational expressions:
9ab 4ab 7a 2 b
(a) (b) (c)
18ab 6a 2 b 21ac
5. Simplify the following:
(a) 23000 × 32000 (b) 3200 × 4−300 (c) 5−2000 × 23000

6. Calculate the following:

 1  − 43
16 2 1
25− 2
3
(a) (b) (c)
25 27
 0.375  − 25
1 1
(d) 0.25−0.5 (e) (f)
256 1024
 − 32  2
1 8 3
(g) +
4 27
7. Write the following expressions in the form of radicals (x, y, z, v > 0):
x− 5 x − 20 y − 15
2 3 7 1 2
(a) (b) (c) x 8 y 25
x− 2 y 4
1 3
x− 2
5
 −0.75
z− 12
7 2 5
(d) (e) 3
(f) x 3 y 12
− 78 5
y
z v 3 4

8. Simplify the following expressions (x > 0):

  − 13 
3 2   1 − 23    2  n−1
1

÷ x −1 2
1
(a) x4 (b) x n x n−1 ÷ x n

9. Simplify the following expressions (x, y, z > 0):

x 3 x 4 x −1
2 3
(a)
3 4 5
 5 2 7
(b) x 2 y 5 z 6 ÷ x 4 y 3 z 12
  
3 2
− 13
 12 − 23
(c) x4 ÷ x −1
Solutions 5

10. Simplify the following expressions with radicals, (x, a, b, n > 0):
√ 2 √ √
xn ÷ xa
3 a n
(a) 4a 2 b (b)

√ √ √
x
√
x
√
x
(c) x x x3x
3
x x (d) a 3x−4 a 1−x a 3−x

1.2 ** Absolute Numbers, Equations, and Inequalities

1. Calculate x in the following equations:

(a) |x − 3| = 7 (b) |3 − x| > 4 (c) ||x| − 5| > 4

2. Simplify the following expressions:

|x| |x| + x |x| + 2x
(a) (b) (c) + |x| + x
x 2 3
2 2
x − 2 + |x − 2| x − 2 − |x − 2|
(d) +
2 2
3. Simplify the following expressions (x, y, z > 0):
x 3 x 4 x −1
2 3
(a)
3 4 5
 5 2 7
(b) x 2 y 5 z 6 ÷ x 4 y 3 z 12
  
3 2
− 13
 12 − 23
(c) x4 ÷ x −1

Solutions

Exercise 1.1, page 3

1. By using definition of negative powers:

 −2
−3 1 1 1
(a) 2 = 3 = (b) = 22 = 4
2 8 2
 −2  −1
1 1
(c) + = 22 + 2 = 6
2 2
6 1 Basic Number Theory

2. By obeying the order of operations we write:

(a) Negative powers of fractions and negative numbers are calculated as
 −3  3
(−2)−3 − (−3)−2 2 (−1/2)3 − (−1/3)2 3
× = ×
(−4)−1 3 − /41 2

− /8 − /9 27
1 1 1 1 27
= × = + × 4
×
−1/4 8 8 9 2

8


17 27 17 × 3 × 9
51
= × = =
72 2 9
× 8 × 2 16

(b) The division symbol is another way to write fractions

 
    1 1 1 1
(−2)−1 + (−3)−1 ÷ (−3)−1 − (−6)−1 = − − ÷ − +
2 3 3 6

5 1 5
=− ÷ − = ×6=5
6 6 6

(c) Expressions with decimal numbers may be simplified to the same powers

−4 × 104 + 2.5 × 105 = −4 × 104 + 25 × 104 = 21 × 104 = 2.1 × 105

(d) Decimal numbers may be converted into their respective fractional forms. In addition, the
powers of two are extensively used in informatics.

0.5−1 + 0.25−2 + 0.125−3 + 0.0625−4

 −1  −2  −3  −4
1 1 1 1
= + + +
2 4 8 16
= 2 + 42 + 83 + 164 = 2 + (22 )2 + (23 )3 + (24 )4
= 2 + 24 + 29 + 216 = 2 + 16 + 512 + 65536 = 66066

3. By definition, x 0 = 1 for x = 0, thus

(a) 1 (b) 1 (c) 1 (d) 0

4. Factorization of numbers and variables:

9
1 
ab 1 
4
2 
ab 2 7
1 a 2 b ab
(a) = (b) = (c) =

 
18 2 
ab 2 6
3 a  b

2 3a  3 ac
21
3c
Solutions 7

5. Large powers may be combined.

(a) 23000 × 32000 = (23 )1000 × (32 )1000 = 81000 × 91000 = 721000
  2 100  200
(32 )100 9 100 3 3
(b) 3200 × 4−300 = 3 100 = = 2
=
(4 ) 64 8 8
3 1000  1000
(2 ) 8
(c) 5−2000 × 23000 = 2 1000 =
(5 ) 25
6. Fractional powers and radicals are interchangeable. Power of power expression is calculated as
the product of powers. Number one is neutral for the multiplication and power operations.
 1 √
16 2 16 4
(a) =√ =
25 25 5

− 32
 − 3  2
2× −3
−3 1 1
(b) 25 = 52 2 = 5  =5 = 3 =
5 125

 − 43
1  4 4
3× 4
(c) = 27 = 33 3 = 3 3 = 34 = 81
3
27
 − 12
−0.5 1 1 √
(d) 0.25 = = 42 = 4=2
4

 0.375  3  
8× 3

(e) 1 1 8 1  =1
8
= =
256 28 2 8

 − 25
1 2   25 
10 2 × 2
(f) = 1024 = 2 5 10
=2 
51 = 24 = 16
1024

 − 32  2  2  
2
1 8 3 3 2 3 3  3 2 3×

3
+ = 42 + = 22 2 +
4 27 33 3
(g) 
4 4 8×9+4 76 
2× 3

=8+ = =2 = = 19  +
2
9 9 9 4
√
7. Rational powers are another way of writing radicals, x m/n = n x m .
 1

− 2 1 5 1
(a) x 5 = =√
x 2 5
x2
 1  1
1 20 1 15 1
(b) x − 20 y − 15 =
3 7
= √ 
x3 y7 20
x 15 y 7
3

1 2 √ 
(c) x 8 y 25 = 8 x 25 y 2
1  
− 12 34 √ 4 y3 √8 7 4
x y x z y3
(d) 5 = = √ √
z− 8 v 3
7
1 √ 3
3
x v5
√8 7
v 5
z
x− 2
5
1
× z− 12 = √ √
7
(e) 3 4
√
y 4 x 5 x 3 12 z7
8 1 Basic Number Theory

 
 2 5 −0.75  2 5 − 34 − 2 1 
3
− 5

31

= x 3 y 12 =x
4
x 3 y 12  
3 42 y

12 4

(f)
1
= x − 2 y − 16 = √ 
1 5

x 16 y 5
√
8. Rational powers are another way of writing radicals, x m/n = n
xm.
(a) Division of two fractions is transformed into product

 2   
− 13
 12 − 23 a c a d  3 − 23   1  23
x −1 x −1 2
3
x4 ÷ = ÷ = = x4
b d b c
1
= x − 12 x − 6 = x − 2 − 3 = x − 6 = √
6 2 1 1 1

6
x

(b) Abstract power expressions follow the same rules as “regular” power expressions.

   2  n−1
1    2 − n−1
1
n2
= x n+ n−1 x − n−1
1 1 1
x n x n−1 ÷ x n = x n x n−1 x n

x − n−1 = x

= x −1 = 1
n2 −n+1 n2 n2 −n+1−n2
=x n−1 n−1
x

9. Simplify and convert into the equivalent radicals if there is an opportunity, (x, y, z > 0):
√
(a) x 3 x 4 x −1 = x 3 + 4 −1 = x 12 = x 5
2 3 2 3 5 12

 5 2 7 √  √
(b) x 2 y 5 z 6 ÷ x 4 y 3 z 12 = x 2 − 4 y 5 − 3 z 6 − 12 = x 4 y 15 z 4 = 4 y 15 y 2 4 z
3 4 5 3 5 4 2 5 7 1 2 1

(c) By definitions of rational division, sum, and negative powers, powers of powers, etc., we
write

 3 2 − 13   12 − 23  3 − 23   1  23
x −1 x −1 2 = x − 12 x − 6
6 2
x4 ÷ = x4

1
= x− 2 − 3 = x− 6 = √
1 1 5

6
x5

10. By rearranging the expression of radical’s argument we find.

√ 2 √ √  √
3
4a 2 b = 24 a 4 b2 = 2 · 23 a 3 a b2 = 3 2(2a)3 a b2 = 2a 2ab2
3 3 3
(a)
√ √ n2 −a 2
x n ÷ x a = x a ÷ x n = x a x − n = x an
n a n a
(b) a n
Solutions 9

(c) Nested radicals are converted to fractional powers and “unfolded” from within.

√ √ 1 3 1 4 3 3
x x=x xx2 = x
3
x x x3x x xx3 x x3 x2


42 1
2
3 1
5 1 5 1 1
=x xx 3
 x 3 = x
21 x3 x2 = x x3 2 x2

14 7
√
= x 1+ 6 + 2 = x 6 3 = x 3 = x 2+ 3 = x 2 3 x
5 1 7 1

(d) “x-th” radicals obey the same rules as the “ordinary” radicals.

√ √ √ 
x

a 3x−4 a 1−x a 3−x = a x a x a x = a  x  = a x = a

x x x 3x−4 1−x 3−x 3x−4+1−x+3−x

Exercise 1.2, page 5

1. Absolute numbers are defined as |x| = x, (x ≥ 0) or, |x| = −x, (x < 0), that is to say that
result of the absolute operation is always a positive real number. It is important to realize that
an equation with one absolute expression is actually a compact syntax to write two equations.
Consequently, the presence of multiple absolute expressions inside a single equation means that
we must systematically write separate equations for each combination of positive/negative cases.
(a) One abs function generates two equations, thus two solutions
|x − 3| = 7
x − 3 ≥ 0 ∴ |x − 3| = x − 3 ∴ x − 3 = 7 ∴ x1 = 10
x − 3 < 0 ∴ |x − 3| = −(x − 3) ∴ −(x − 3) = 7 ∴ x2 = −4
Therefore, there are two discrete solutions, x1 = 10 and x2 = −4, see Fig. 1.1.

Fig. 1.1 Example 1.2-1(a)

−4 10 x

(b) However, an inequality with abs function generates two inequalities, thus two intervals as
the solutions. First, we calculate the interval boundaries

|3 − x| > 4
3 − x ≥ 0 ∴ |3 − x| = 3 − x ∴ 3 − x > 4 ∴ x < −1
3 − x < 0 ∴ |3 − x| = −(3 − x) ∴ −(3 − x) > 4 ∴ x > 7
10 1 Basic Number Theory

Therefore, there are two intervals x ∈ [−∞, −1] and x ∈ [7, +∞], see Fig. 1.2

Fig. 1.2 Example 1.2-1(b) x < −1 x>7

−1 7 x

(c) Due to two abs functions, in total there are four possible inequalities to solve

||x| − 5| > 4

where, the four possible inequalities are found as,

x>0:

|x| = x ∴ |x − 5| > 4 if: x − 5 ≥ 0 ∴ x − 5 > 4 ∴ x > 9

if: x − 5 < 0 ∴ −x + 5 > 4 ∴ x < 1
x<0:
|x| = −x ∴ | − x − 5| > 4 if: −x − 5 ≥ 0 ∴ −x − 5 > 4 ∴ x < −9
if: −x − 5 < 0 ∴ x + 5 > 4 ∴ x > −1
Four inequalities must be satisfied at the same time, we summarize them as (also, see Fig. 1.3)

x ∈ [−∞, −9] and − 1 ≤ x ≤ 1 and x ∈ [9, +∞]

Fig. 1.3 Example 1.2-1(c) x >−1

x <−9 x <1 x >9

−9 −1 1 9 x
Solutions 11

Plot of function ||x| − 5| vs. “4” clearly illustrates the intervals where ||x| − 5| > 4, see Fig. 1.4.

Fig. 1.4 Example 1.2-1(c)

||x| −5|

x
0

−9 −1 1 9

2. Some typical expressions with abs function that are extensively used in discrete mathematics for
signal processing.
(a) Signs of abs function’s argument are
|x| x
x > 0 : |x| = x ∴ = =1
x x
|x| −x
x < 0 : |x| = −x ∴ = = −1
x x
Indeed, these two |x|/x = x/|x| = ±1 ratios are known as sign function.
(b) Signs of abs function’s argument are
|x| + x x+x 2x
x > 0 : |x| = x ∴ = = =x
2 2 2
|x| + x −x + x 0
x < 0 : |x| = −x ∴ = = =0
2 2 2
This is one of the possible definitions for the function known as the ramp function.
(c) Similarly,
|x| + 2x x + 2x
x > 0 : |x| = x ∴ + |x| + x = + x + x = 3x
3 3
|x| + 2x −x + 2x x
x < 0 : |x| = −x ∴ + |x| + x = −x+x =
3 3 3
12 1 Basic Number Theory

(d) Signs of abs function’s argument are

x − 2 + |x − 2| 2 x − 2 − |x − 2| 2
x−2>0: ∴ +
2 2

x−2+x−2 2
+
x − 2 − x :

2
2 0
= + 
2  2
2x − 4 2
= = (x − 2)2
2
x − 2 + |x − 2| 2 x − 2 − |x − 2| 2
x−2<0: ∴ +
2 2

x − 2 −
x 2
+ :

2 0
x−2+x−2 2
=   +
 2 2
2
2x − 4
= = (x − 2)2
2
3. Simplify and convert into the equivalent radicals if there is an opportunity, (x, y, z > 0):
√
(a) x 3 x 4 x −1 = x 3 + 4 −1 = x 12 = x 5
2 3 2 3 5 12

 5 2 7 √  √
(b) x 2 y 5 z 6 ÷ x 4 y 3 z 12 = x 2 − 4 y 5 − 3 z 6 − 12 = x 4 y 15 z 4 = 4 y 15 y 2 4 z
3 4 5 3 5 4 2 5 7 1 2 1

(c) By definitions of rational division, sum, and negative powers, powers of powers, etc., we
write

 3 2 − 13   12 − 23  3 − 23   1  23
x −1 x −1 2 = x − 12 x − 6
6 2
x4 ÷ = x4

1
= x− 2 − 3 = x− 6 = √
1 1 5

6
x5
 Polynomials
2

Important to Know

Basic polynomial transformations:

a(b ± c) = ab + ac
(a + b)(c + d) = ac + ad + bc + bd
(a ± b)2 = a 2 ± 2ab + b2
(a 2 − b2 ) = (a − b)(a + b)
(a 3 ± b3 ) = (a ± b)(a 2 ∓ ab + b2 )

Factor theorem: reminder r of division P (x) ÷ (x − a) is equal to r = P (a). Consequently, if

r = P (a) = 0, then P (x) is divisible by (x − a). That being the case, binomial (x − a) is one of the
P (x) factors and x = a one of its roots.

2.1 Exercises

2.1 Multiplication, Identity

1. Given P (x) = x 2 + x + 1 and Q(x) = x 3 − 1, calculate

(a) P (x) + Q(x) (b) P (x) − Q(x) (c) 2xP (x) − 2Q(x)

2. Multiply the following polynomials:

(a) (x − 1)(x 2 + x + 1) (b) (x + 3)(x 2 − 3x + 9)
(c) (3x 2 − 5x + 6)(2x − 7) (d) (x 4 + x 3 + 2)(x 2 − 3x + 4)
(e) (3y 2 + 2x − 6xy)(2x 2 + 4xy − 2y 3 ) (f) (x − y)(y − z)(z − x)

© Springer Nature Switzerland AG 2021 13

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_2
14 2 Polynomials

3. Calculate parameters a, b, c so that P (x) = Q(x), if

(a) P (x) = x 3 − 2x 2 + 3, Q(x) = (x + 1)(ax 2 + bx + c)
(b) P (x) = 2x 3 − x 2 + x + 4, Q(x) = (x − 2)(ax 2 + bx + c)

4. Given P (x) = x 3 +ax 2 +bx +c calculate parameters a, b, c so that it is divisible by the following
binomials: (x + 1), (x − 1), and (x + 2).
5. With the help of Pascal’s triangle, calculate
(a) (x − 1)6 (b) (x + 2)5 (c) (2x + 1)4

2.2 Factor Theorem

1. With the help of the factor theorem, answer the following questions.
(a) Calculate the reminder in (x 4 − 2x 3 + 3x 2 − 4x + 1) ÷ (x − 1)
(b) Calculate the reminder in (x 3 + 1) ÷ (x + 1)
(c) Calculate parameter n so that P (x) = 4x 5 + nx 4 + 8x 3 + 5x 2 + 3x + 2 is divisible by x + 2.

2.3 *** Factor Theorem

1. Using the factor theorem, answer the following questions.

(a) Calculate parameter n so that
P (x) = x 3 − 3nx + 4(n2 + 1)x − (n3 + 5) is divisible by (x − 1)
(b) Calculate reminder in P (x)/Q(x), given
P (x) = x 1965 − 256x 1961 + 1 and Q(x) = x 2 − 4x
(c) Calculate reminder in P (x)/Q(x), given that
P (x) = x 2021 − 2x 2020 − 1 and Q(x) = x 2 − 3x + 2

2.4 Common Factors

1. Factorize the following polynomials by finding the common factors:

(a) 5a + 5x (b) 2a − 2 (c) 7a − 14
(d) 3a + 9
2
(e) 3a + 6b + 9 (f) 6x + ax + bx
(g) 9a − 6a + 12
2
(h) a −a
2 3
(i) 3a 2 − 6a
(j) x 3a2 − x 3 (k) 3a 3 + 2a 2 + a (l) 4x 2 − 2x + xy

2. Factorize the common terms:

(a) x3y3 − x3y + x4y3 (b) x3y3 − x2y8
(c) 6x 2 y 2 − 4xy 3 (d) 5x 3 − 15x 2 y 3
(e) 6x 3 y − 9x 2 y 2 + 3x 3 y 2 (f) x 3 − x 7 − 2x 5
2.1 Exercises 15

(g) a 3 b2 + 2a 4 b2 − 4ab5 (h) 3a 3 b3 − 9a 2 b4 + 12a 5 b4

(i) a(m + n) + b(m + n) (j) m(a − b) + n(a − b)
(k) 7q(p − q) + 2p(q − p) (l) 3(x + y) + (x + y)2

3. Factorize the following polynomials by finding the common terms:

(a) am − an + bm − bn (b) am − an − bm + bn
(c) ab + ay − bx − xy (d) an − ab − mn + mb
(e) 5ax + 5ay − x − y (f) 2x 2 + 2xy − x − y
(g) 4ym − 4yn − m + n (h) ax 2 − bx 2 − bx + ax − a + b
(i) 6by − 15bx − 4ay + 10ax (j) 5ax 2 − 10ax − bx + 2b − x + 2

4. Factorize the following polynomials by finding the common terms:

(a) xyz + x 2 y 2 − 3x 4 y 5 − 3x 3 y 4 − xy − z
(b) m2 x 4 − mnx 3 + 2mx 2 − 2nx + n − mx

5. Factorize the following expressions:

(a) a 2n + a n (b) a 3x − 2a 2x bx
(c) 2x m+n + 6x n (d) a 3x + 3a 2x + 5a x
(e) x 2n+2 − 2x n+1 + 1

2.5 Polynomial Identities: a 2 − b2 = (a + b)(a − b)

1. Factorize by using the difference of squares identity:

(a) x 2 − 49 (b) a 2 − 36 (c) 16x 2 − 9
(d) 9x 2 − 49 (e) 25 − x 2 (f) 9 − x4
1 x2 4 9x 2 4y 2
(g) x2 − (h) − (i) −
49 4 9 4 9
49x 2
(j) − 9y 2 (k) x 2 − 0.36 (l) x 2 − 0.0009
25
(m) 0.04x 2 − 0.25 (n) 0.01x 2 − 0.04y 2 (o) x 4 y 2 − 0.01
(p) 0.25x 2 y 2 − 0.0001

2. Factorize the following expressions:

(a) (x − 3)2 − 4 (b) (a + 5)2 − 9
(c) y 2 − (x − y)2 (d) x 2 − (x + y)2
(e) (x + 2)2 − 4x 2 (f) 9x 2 − (x − 1)2
(g) (x − y)2 − 16(x + y)2 (h) (x + 2y)2 − 9(x − 2y)2
(i) 4(x − y)2 − 25(x + y)2 (j) 36(x − 2)2 − 25(x + 1)2
(k) (x + y − z)2 − (x − y + z)2 (l) (x + y − 3)2 − 25(x + 2)2
16 2 Polynomials

3. Calculate the following without using a calculator:

(a) 98 × 102 (b) 99 × 101 (c) 83 × 77
(d) 79 × 81 (e) 18 × 22 (f) 201 × 199
(g) 1.05 × 0.95 (h) 1.01 × 0.99 (i) 9.9 × 10.1

2.6 The Binomial Square (ax)2 ± 2abx + b2 = (ax ± b)2

1. Factorize the following quadratic and bi-quadratic polynomials:

√
(a) x 2 − 2x + 1 (b) x 2 − 6x + 9 (c) x 2 − 2 2x + 2
(d) 4x 2 − 12x + 9 (e) x 4 − 2x 2 + 1 (f) x 2n+2 − 2x n+1 + 1

2.7 * Viète’s Formulas: x1 + x2 = −(b/a), x1 x2 = (c/a)

1. Factorize the following quadratic polynomials:

(a) x 2 − 6x + 5 (b) x 2 − 9x + 14
(c) x 2 − 6x + 8 (d) 2x 2 + 3x + 1
(e) 3x 2 − 14x + 8 (f) −2x 2 + x + 3
(g) −x 2 + 5x − 4 (h) −6x 2 + 5x + 4
√ √ √
(i) x 2 + ( 2 + 3)x + 6

2.8 * Bi-quadratic Form: ax 2n + bx n + c

1. Factorize the following bi-quadratic polynomials:

(a) x 4 − 13x 2 + 36 (b) x 4 − 10x 2 + 9 (c) x 6 − 2x 3 + 1
(d) x 6 + 3x 3 + 2 (e) x 6 − 9x 3 + 8 (f) 2x 2m+2 − 11x m+1 + 9

2.9 ** Perfect Square (x + a)2 = b

1. Completing the square technique exploits binomial square form.

(a) x 2 + 6x − 7 = 0 (b) 2x 2 − 10x − 3 = 0
(c) −x 2 − 6x − 3 = 0 (d) x 2 + 3x
(e) 2x 2 + 6x + 2 (f) x 2 − x − 6

2.10 * Long Division

1. Do the following division:

(a) (x 4 − 2x 3 − 7x 2 + 8x + 12) ÷ (x − 3)
(b) (6x 3 + 10x 2 + 8) ÷ (2x 2 + 1)
(c) (4x 3 − 7x 2 − 11x + 5) ÷ (4x + 5)
(d) (x 2 + x − 5) ÷ (x + 3)
2.1 Exercises 17

2. Factorize:
(a) P (x) = x 3 + x 2 − 4x − 4 (b) Q(x) = x 3 − 1
(c) Q(x) = x 3 − y 3 (d) Q(x) = x 3 + y 3

2.11 ** Sum of Cubes a 3 ± b3 = (a ± b)(a 2 ∓ ab + b2 )

1. Factorize:
(a) Q(x) = x 3 − 1 (b) Q(x) = x 3 − y 3 (c) Q(x) = x 3 + y 3

2.12 **** Multiple Factorization Techniques

1. Factorize the following expressions.

(a) x 4 + 4 (b) x 4 + x 2 + 1
(c) x 5 + x + 1 (d) (x + y + z)3 − x 3 − y 3 − z3
(e) (x + 1)(x + 3)(x + 5)(x + 7) + 15 (f) (x 2 + x + 1)(x 3 + x 2 + 1) − 1

2.13 *** Multiple Factorization Techniques

1. Simplify, (x = 0):
x + x2 + x3 + · · · + xn
x n+2 x 3n+2 + x 3n+1
(a) (b) (c) 1 1 1 1
2x x n x 2 x 3n + 2 + 3 + ··· + n
x x x x

2.14 *** Partial Fractions

1. Derive the partial fraction form of the following rational functions

x x+2 2x 2
(a) (b) (c)
(x + 1)(x − 4) x3 − 2x 2 x4 − 1
x 3 + x 2 − 16x + 16 1 4
(d) (e) (f)
x 2 − 4x + 3 x3 − 1 x4 + 1
x +x−1
3
(g)
(x 2 + 2)2
18 2 Polynomials

Solutions

Exercise 2.1, page 13

1. Given P (x) = x 2 + x + 1 and Q(x) = x 3 − 1, we write

(a) P (x) + Q(x) = (x 2 + x + 1) + (x 3 − 1) = x 3 − 1
+ x 2 + x + 1

= x(x 2 + x + 1)
(b) P (x) − Q(x) = (x 2 + x + 1) − (x 3 − 1) = −x 3 + 1 + x 2 + x + 1
= −x 3 + x 2 + x + 2
(c) 2x3 + 2x 2 + 2x − 
2xP (x) − 2Q(x) = 2x(x 2 + x + 1) − 2(x 3 − 1) =  2x3 + 2
= 2x + 2x + 2 = 2(x + x + 1)
2 2

2. Multiply the following polynomials:

(a) (x − 1)(x 2 + x + 1) = x 3 + x 2 + x
− x 2 − x
− 1
= x3 − 1
(b) 3x2 + 
(x + 3)(x 2 − 3x + 9) = x 3 −  +
9x 3x2 − 
 + 27
9x
=x −3
3 3

(c) (3x 2 − 5x + 6)(2x − 7) = 6x 3 − 21x 2 − 10x 2 + 35x + 12x − 42

= 6x 3 − 31x 2 + 47x − 42
(d) (x 4 + x 3 + 2)(x 2 − 3x + 4) = x 6 − 3x 5 + 4x 4
+ x 5 − 3x 4 + 4x 3
+ 2x 2 − 6x + 8
= x 6 − 2x 5 + x 4 + 4x 3 + 2x 2 − 6x + 8
(e) (3y 2 + 2x − 6xy)(2x 2 + 4xy − 2y 3 )
= 6x 2 y 2 + 12xy 3 − 6y 5 + 4x 3 + 8x 2 y − 4xy 3 − 12x 3 y − 24x 2 y 2 + 12xy 2
= −18x 2 y 2 + 8xy 3 − 6y 5 + 4x 3 + 8x 2 y − 12x 3 y + 12xy 4
(f) (x − y)(y − z)(z − x) = (xy − xz − y 2 + yz)(z − x)
=  − xz2 − y 2 z + yz2 − x 2 y + x 2 z + xy 2 − 
xyz 
xyz
= x 2 z + yz2 + xy 2 − xz2 − y 2 z − x 2 y

3. Two polynomials are equal if each of their corresponding monomials are equal. The product of
(x + 1) binomial and parametric quadratic polynomial is therefore
Solutions 19

(a) Q(x) = (x + 1)(ax 2 + bx + c) = ax 3 + bx 2 + cx + ax 2 + bx + c

= ax 3 + (a + b)x 2 + (b + c)x + c
∴
Q(x) = P (x) ∴ ax 3 + (a + b)x 2 + (b + c)x + c = x 3 − 2x 2 + 3
x 3 terms: ∴ a = 1
x 0 terms: ∴ c = 3
x 2 terms: a + b = −2 ∴ 1 + b = −2 ∴ b = −3
x 1 terms: b + c = 0 ∴ −3 + 3 = 0 
In conclusion, P (x) = Q(x) ∴ x 3 − 2x 2 + 3 = (x + 1)(x 2 − 3x + 3)

(b) Q(x) = (x − 2)(ax 2 + bx + c) = ax 3 + bx 2 + cx − 2ax 2 − 2bx − 2c

= ax 3 + (b − 2a)x 2 + (c − 2b)x − 2c
∴
Q(x) = P (x) ∴ ax 3 + (b − 2a)x 2 + (c − 2b)x − 2c = 2x 3 − x 2 + x + 4
∴
x 3 terms: ∴ a = 2
x 0 terms: − 2c = 4 ∴ c = −2
x 2 terms: b − 2a = −1 ∴ b − 4 = −1 ∴ b = 3
x 1 terms: c − 2b = 1 ∴ −2 − 6 = 1 ✗
In conclusion, there is no Q(x) that satisfies this P (x) = Q(x) equation.

4. Third order polynomial Q(x) that is divisible by (x + 1), (x − 1), and (x + 2) must have form

Q(x) = (x + 1)(x − 1)(x + 2) = (x 2 − 1)(x + 2) = x 3 + 2x 2 − x − 2

Therefore, we conclude,

P (x) = Q(x) ∴ x 3 + ax 2 + bx + c = x 3 + 2x 2 − x − 2
∴
a = 2, b = −1, and c = −2
5. Formal expression for binomial expansion is
n 
 n 

n n
(a + b)n = a n−k bk = a k bn−k (2.1)
k=0
k k=0
k

and it is used to generate the polynomial development of (a +b)n . Visually, due to their symmetry,
the polynomial coefficients are easily arranged in the form of Pascal’s triangle, Fig. 2.1. Each
20 2 Polynomials

row lists polynomial coefficients of the corresponding n-th binomial power. Note that each row
coefficient is calculated as the sum of two neighbouring coefficients left and right in the row
above.
(a) Given n = 6 and a = x and b = −1, in Pascal’s triangle we find the following coefficients:
1, 6, 15, 20, 15, 6, 1. Note how, in accordance to (2.1), the powers of a and b are systematically
written in the falling and increasing orders, respectively.

(x − 1)6 = 1 x 6 (−1)0 + 6 x 5 (−1)1 + 15 x 4 (−1)2 + 20 x 3 (−1)3 +

15 x 2 (−1)4 + 6 x 1 (−1)5 + 1 x 0 (−1)6
= x 6 − 6x 5 + 15x 4 − 20x 3 + 15x 2 − 6x + 1
(b) Given n = 5 and a = x and b = 2, in Pascal’s triangle we find the following coefficients:
1, 5, 10, 10, 5, 1, thus we write

(x + 2)5 = 1 x 5 20 + 5 x 4 21 + 10 x 3 22 + 10 x 2 23 + 5 x 1 24 + 1 x 0 25
= x 5 + 10x 4 + 40x 3 + 80x 2 + 80x + 32
(c) Given n = 4 and a = 2x and b = 1, in Pascal’s triangle we find the following coefficients:
1, 4, 6, 4, 1, thus we write

(2x + 1)4 = 1 (2x)4 10 + 4 (2x)3 11 + 6 (2x)2 12 + 4 (2x)1 13 + 1 (2x)0 14

= 16x 4 + 32x 3 + 24x 2 + 8x + 1

Fig. 2.1 Example 2.1-5 n=0 1

n=1 1 1
n=2 1 2 1
n=3 1 3 3 1
n=4 1 4 6 4 1
n=5 1 5 10 10 5 1
n=6 1 6 15 20 15 6 1
n=7 1 7 21 35 35 21 7 1
···

Exercise 2.2, page 14

1. The factor theorem states that a polynomial P (x) has a binomial factor (x − n) if and only if
P (n) = 0, that is to say n is one of the P (x) roots. Otherwise, if the value P (n) = 0 we say that
r = P (n) is the reminder of division P (x) ÷ (x − n).
(a) Given divisor (x − 1), i.e. n = 1, we find
r = P (1) = (1)4 − 2(1)3 + 3(1)2 − 4(1) + 1) = −1
Therefore n = 1 is not the root of P (x).
Solutions 21

(b) Given divisor (x + 1), i.e. n = −1, we find

r = P (−1) = (−1)3 + 1 = 0
Therefore n = −1 is one of the roots of P (x), in other words, P (x) is divisible by (x +1). Indeed,

x3 + 1
= x2 − x + 1
x+1

(c) For P (x) to be divisible by (x + 2) it is necessary that r = P (−2) = 0. Therefore, we write,

r = P (−2) = 4(−2)5 + k(−2)4 + 8(−2)3 + 5(−2)2 + 3(−2) + 2

= −128 + 16k − 64 + 20 − 6 + 2 = 0 ∴ 16k − 176 = 0
176
∴ n= = 11
16

Indeed,

4x 5 + 11x 4 + 8x 3 + 5x 2 + 3x + 2
= 4x 4 + 3x 3 + 2x 2 + x + 1
x+2

Exercise 2.3, page 14

1. The factor theorem states that a polynomial P (x) has a binomial factor (x − n) if and only if
P (n) = 0, that is to say n is one of the P (x) roots. Otherwise, if the value P (n) = 0 we say that
r = P (n) is the reminder of division P (x) ÷ (x − n).
(a) Given binomial (x − 1), the condition of divisibility for P (x)/(x − 1) is that r = P (1) = 0,
therefore we write,

r = P (1) = (1)3 − 3n(1) + 4(n2 + 1)(1) − (n3 + 5)

= 1 − 3n + 4n2 + 4 − n3 − 5
= −n3 + 4n2 − 3n = −n(n2 − 4n + 3) = −n(n2 − n − 3n + 3)
= −n[n(n − 1) − 3(n − 1)] = −n(n − 1)(n − 3) = 0
22 2 Polynomials

Therefore, there are three values of n that produce r = 0, i.e. n = 0, 1, 3. Indeed,

P (x) x 3 + 4x − 5
n=0: = = x2 + x + 5
x−1 x−1
P (x) x 3 + 5x − 6
n=1: = = x2 + x + 6
x−1 x−1
P (x) x 3 + 31x − 32
n=3: = = x 2 + x + 32
x−1 x−1

(b) Given Q(x) = x 2 − 4x = x(x − 4), we calculate the division reminders for both (x − 0)
and (x − 4) factors. There are multiple x to calculate, thus we must calculate the reminders for
each of them P (0) and P (4), i.e. r(x) as

:0
 :0

(0)
r(0) = P (0) =  1965
(0)
− 256 1961
+1=1
r(4) = P (4) = (4)1965 − 256(4)1961 + 1 = 41965 − 44 (4)1961 + 1

=  − 41965

41965  +1=1
In conclusion, reminder is r = 1. In other words P (x) = Q(x)f (x) + 1, where f (x) is a 1963th
order polynomial.

(c) Given
Q(x) = x 2 − 3x + 2 = x 2 − x − 2x + 2 = x(x − 1) − 2(x − 1)
= (x − 1)(x − 2)
we calculate the division reminders for both (x − 1) and (x − 2) factors. There are multiple x to
calculate, we must calculate the reminders for each of them P (1) and P (2), i.e. r(x) as

:1
 :1

(1)
r(1) = P (1) =  2021
(1)
− 2 2020
− 1 = −2

r(2) = P (2) = (2)2021 − 2(2)2020 − 1 =   − 22021


22021  − 1 = −1
Because r(1) = r(2) and there are only two points in consideration, x = 1 and x = 2,
we conclude that reminder must be a first order polynomial r(x) = ax + b, where its linear
coefficients are calculated as
 
x = 1 : r(1) = −2 = a(1) + b a + b = −2
∴ ∴ a = 1, b = −3
x = 2 : r(2) = −1 = a(2) + b 2a + b = −1
In conclusion, the reminder in division P (x)/Q(x) is r(x) = ax + b = x − 3.
Solutions 23

Exercise 2.4, page 14

1. We find the common factors:

(a) 5a + 5x = 5(a + x) (b) 2a − 2 = 2(a − 1)
(c) 7a − 14 = 7(a − 2) (d) 3a 2 + 9 = 3(a 2 + 3)
(e) 3a + 6b + 9 = 3(a + 2b + 3) (f) 6x + ax + bx = x(6 + a + b)
(g) 9a − 6a + 12 = 3(3a − 2a + 4)
2 2
(h) a 2 − a 3 = a 2 (1 − a)
(i) 3a 2 − 6a = 3a(a − 2) (j) x 3 a 2 − x 3 = x 3 (a 2 − 1) = x 3 (a + 1)(a − 1)
(k) 3a 3 + 2a 2 + a = a(3a 2 + 2a + 1) (l) 4x 2 − 2x + xy = x(4x − 2 − y)

2. We search for the highest order common factors:

(a) x 3 y 3 − x 3 y + x 4 y 3 = x 3 y y 2 − x 3 y + x 3 x y y 2 = x 3 y(y 2 − 1 + xy 2 )
(b) x 3 y 3 − x 2 y 8 = x 2 y 3 (x − y 5 )
(c) 6x 2 y 2 − 4xy 3 = 2xy 2 (3x − 2y)
(d) 5x 3 − 15x 2 y 3 = 5x 2 (x − 3y 3 )
(e) 6x 3 y − 9x 2 y 2 + 3x 3 y 2 = 3x 2 y(2x − 3y + xy)
(f) x 3 − x 7 − 2x 5 = x 3 (1 − x 4 − 2x 2 )
(g) a 3 b2 + 2a 4 b2 − 4ab5 = ab2 (a 2 + 2a 3 − 4b3 )
(h) 3a 3 b3 − 9a 2 b4 + 12a 5 b4 = 3a 2 b3 (a − 3b + 4a 3 b)
(i) a(m + n) + b(m + n) = (m + n)(a + b)
(j) m(a − b) + n(a − b) = (a − b)(m + n)
(k) 7q(p − q) + 2p(q − p) = (p − q)(7q − 2p)
(l) 3(x + y) + (x + y)2 = (x + y)(3 + x + y)

3. We factorize in two steps by grouping binomials with the same factors:

(a) am − an + bm − bn = a(m − n) + b(m − n) = (m − n)(a + b)
(b) am − an − bm + bn = a(m − n) − b(m − n) = (m − n)(a − b)
(c) ab + ay − bx − xy = a(b + y) − x(b + y) = (b + y)(a − x)
(d) an − ab − mn + mb = a(n − b) − m(n − b) = (n − b)(a − m)
(e) 5ax + 5ay − x − y = 5a(x + y) − (x + y) = (x + y)(5a − 1)
(f) 2x 2 + 2xy − x − y = 2x(x + y) − (x + y) = (x + y)(2x − 1)
(g) 4ym − 4yn − m + n = 4y(m − n) − (m − n) = (m − n)(4y − 1)
(h) ax 2 − bx 2 − bx + ax − a + b = x 2 (a − b) + x(a − b) − (a − b)
= (a − b)(x 2 + x − 1)
(i) 6by − 15bx − 4ay + 10ax = 2y(3b − 2a) − 5x(3b − 2a)
= (3b − 2a)(2y − 5x)
(j) 5ax 2 − 10ax − bx + 2b − x + 2 = 5ax(x − 2) − b(x − 2) − (x − 2)
= (x − 2)(5ax − b − 1)
24 2 Polynomials

4. We factorize in two steps by grouping binomials with the same factors:

(a) xyz + x 2 y 2 − 3x 4 y 5 − 3x 3 y 4 z − xy − z
= xy(xy + z) − 3x 3 y 4 (xy + z) − (xy + z)
= (xy + z)(xy − 3x 3 y 4 − 1)
(b) m2 x 4 − mnx 3 + 2mx 2 − 2nx + n − mx
= mx 3 (mx − n) + 2x(mx − n) − (mx − n)
= (mx − n)(mx 3 + 2x − 1)

5. We search for the highest order terms:

(a) a 2n + a n = a n a n + a n = a n (a n + 1)
(b) a 3x − 2a 2x bx = a 2x (a x − 2bx )
(c) 2x m+n + 6x n = 2x n (x m + 3)
(d) a 3x + 3a 2x + 5a x = a x (a 2x + 3a x + 5)
 2  
(e) x 2n+2 − 2x n+1 + 1 = x 2(n+1) − 2x n+1 + 1 = x n+1 − 2 x n+1 + 1
 2
= x n+1 − 1

Exercise 2.5, page 15

1. Using the difference of squares identity, we simply write

(a) x 2 − 49 = x 2 − 72 = (x + 7)(x − 7)
(b) a 2 − 36 = a 2 − 62 = (a + 6)(a − 6)
(c) 16x 2 − 9 = (4x)2 − 32 = (4x + 3)(4x − 3)
(d) 9x 2 − 49 = (3x)2 − 72 = (3x + 7)(3x − 7)
(e) 25 − x 2 = 52 − x 2 = (5 + x)(5 − x)
(f) 81 − x 4 = 92 − (x 2 )2 = (9 + x 2 )(9 − x 2 ) = (9 + x 2 )(3 + x)(3 − x)
 
1 1 1
(g) x2 − = x+ x−
49 7 7
   2  
x2 4 x 2 2 x 2 x 2
(h) − = − = + −
4 9 2 3 2 3 2 3
2 2  
9x 4y 3x 2y 3x 2y
(i) − = + −
4 9 2 3 2 3
 2  
49x 2 7x 7x 7x
(j) − 9y 2 = − (3y)2 = +3 −3
25 5 5 5
 2  
9

36 3 3 3
(k) x 2 − 0.36 = x 2 −  = x2 − = x+ x−

100 25 5 5 5
(l) x 2 − 0.0009 = x 2 − 0.032 = (x + 0.03) (x − 0.03)
(m) 0.04x 2 − 0.25 = (0.2x)2 − 0.52 = (0.2x + 0.5)(0.2x − 0.5)
1
= (2x + 5)(2x − 5)
100
Solutions 25

(n) 0.01x 2 − 0.04y 2 = (0.1x)2 − (0.2y)2

= (0.1x + 0.2y)(0.1x − 0.2y)
1
= (x + 2y)(x − 2y)
100
 2
(o) x 4 y 2 − 0.01 = x 2 y − 0.12 = (x 2 y + 0.1)(x 2 y − 0.1)
(p) 0.25x 2 y 2 − 0.0001 = (0.5xy + 0.01)(0.5xy − 0.01)

2. After using the difference of squares identity we simplify if possible:

(a) (x − 3)2 − 4 = (x − 3 + 2)(x − 3 − 2) = (x − 1)(x − 5)
(b) (a + 5)2 − 9 = (a + 5 + 3)(a + 5 − 3) = (a + 8)(a + 2)
(c) y 2 − (x − y)2 = (y
+ x − y)(y

− (x − y)) = x(2y − x)
(d) x 2 − (x + y)2 = (x + x + y)(x
− x
− y) = −y(2x + y)
(e) (x + 2)2 − 4x 2 = (x + 2 + 2x)(x + 2 − 2x) = (3x + 2)(2 − x)
(f) 9x 2 − (x − 1)2 = (3x + x − 1)(3x − (x − 1)) = (4x − 1)(2x + 1)
(g) (x − y)2 − 16(x + y)2 = (x − y + 4x + 4y)(x − y − 4x − 4y)
= (5x + 3y)(−3x − 5y) = −(5x + 3y)(3x + 5y)
(h) (x + 2y)2 − 9(x − 2y)2 = (x + 2y + 3x − 6y)(x + 2y − (3x − 6y))
= (4x − 4y)(8y − 2x) = 4(x − y) 2(4y − x)
= 8 (x − y)(4y − x)
(i) 4(x − y) − 25(x + y) = (2x − 2y + 5x + 5y)(2x − 2y − 5x − 5y)
2 2

= (7x + 3y)(−3x − 7y) = −(7x + 3y)(3x + 7y)

(j) 36(x − 2)2 − 25(x + 1)2 = (6x − 12 + 5x + 5)(6x − 12 − 5x − 5)
= (11x − 7)(x − 17)
(k) (x + y − z)2 − (x − y + z)2 = (x + y
−
z + x − y
+
z)(x
+ y − z − x
+ y − z)
= 2x(2y − 2z) = 4x(y − z)
(l) (x + y − 3)2 − 25(x + 2)2 = (x + y − 3 + 5x + 10)(x + y − 3 − 5x − 10)
= (6x + y + 7)(y − 4x − 13)

3. Practical calculations exploiting on the squares difference identity:

(a) 99 × 101 = (100 − 1)(100 + 1) = 1002 − 1 = 9999
(b) 98 × 102 = (100 − 2)(100 + 2) = 1002 − 22 = 10,000 − 4 = 9996
(c) 83 × 77 = (80 + 3)(80 − 3) = 802 − 32 = 6400 − 9 = 6391
(d) 79 × 81 = (80 − 1)(80 + 1) = 802 − 1 = 6400 − 1 = 6399
(e) 18 × 22 = (20 − 2)(20 + 2) = 400 − 4 = 396
(f) 201 × 199 = (200 + 1)(200 − 1) = 40,000 − 1 = 39,999
(g) 1.05 × 0.95 = (1 + 0.05)(1 − 0.05) = 1 − 0.052 = 1 − 0.0025 = 0.9975
(h) 1.01 × 0.99 = (1 + 0.01)(1 − 0.01) = 1 − 0.012 = 1 − 0.0001 = 0.9999
(i) 9.9 × 10.1 = (10 − 0.1)(10 + 0.1) = 100 − 0.01 = 99.99
26 2 Polynomials

Exercise 2.6, page 16

1. All trinomials whose form is (ax)2 ± 2abx + b2 are factorized into (ax ± b)2 .
(a) x 2 − 2x + 1 = (1x)2 − 2(1)(1)x + 12 = (x − 1)2
(b) x 2 −6x + 9 = x 2 −2 × 3x + 32 = (x − 3)2
√ √ √ √
(c) x 2 − 2 2x + 2 = x 2 − 2 2x + ( 2)2 = (x − 2)2
(d) 4x 2 − 12x + 9 = (2x)2 − 2 × 3 × 2x + 32 = (2x − 3)2
(e) x 4 − 2x 2 + 1 = (x 2 )2 − 2x 2 + 1 = (x 2 − 1)2 = (x + 1)2 (x − 1)2
 2  
(f) x 2n+2 − 2x n+1 + 1 = x 2(n+1) − 2x n+1 + 1 = x n+1 − 2 x n+1 + 1
= (x n+1 − 1)2

Exercise 2.7, page 16

1. The relations between the sum and product of polynomial roots, attributed to François Viète, are
rather practical, among other applications in mathematics, for rapid factorization of quadratic
polynomials.
(a) Given x 2 − 6x + 5 we search for factors x1 , x2 of “+5” so that x1 + x2 = −6. Being
prime number, five is divisible only with ±1, ±5. We find that −1 and −5 factors satisfy both
conditions, i.e. their product equals +5 and their sum is −6. Thus we write,
x 2 −6x + 5 = x 2 −(1)x − 5x + 5 = x(x − 1) − 5(x − 1) = (x − 1)(x − 5)

(b) Two factors of “+14” whose sum equals −9 are −2 and −7, thus we write,

x 2 −9x + 14 = x 2 −2x − 7x + 14 = x(x − 2) − 7(x − 2) = (x − 2)(x − 7)

(c) Two factors of “+8” whose sum equals −6 are −2 and −4, thus we write,
x 2 −6x + 8 = x 2 −2x − 4x + 8 = x(x − 2) − 4(x − 2) = (x − 2)(x − 4)

(d) In case when the leading coefficient of ax 2 + bx + c does not equal one, i.e. a = 2 = 1,
there are at least two possible techniques to use,

Technique 1: first, factor the leading coefficient in trinomial as ax 2 + bx + c = a[x 2 + (b/a)x +

(c/a)], then search factors of (c/a) whose sum equals (b/a), i.e.
Solutions 27

 
3 1
2x + 3x + 1 = 2 x + x +
2 2
2 2
 
1 1 1 3
= 1× = and 1 + =
2 2 2 2
 
1 1 1
= 2 x2 + x + x + = 2 x(x + 1) + (x + 1)
2 2 2

1
= 2 (x + 1) x + = (x + 1)(2x + 1)
2

Technique 2: search factors of (ac) whose sum equals b, i.e.

 
2x 2 + 3x + 1 = ac = 2 × 1 = 2 ∴ 1 × 2 = 2 and 1 + 2 = 3
= 2x 2 +3x + 1 = 2x 2 +2x + x + 1 = 2x(x + 1) + x + 1
= (x + 1)(2x + 1)
 
3 × 8 = 24, (−12) × (−2) = 24 and
(e) 3x −14x + 8 =
2
(−12) + (−2) = −14
= 3x 2 −12x − 2x + 8 = 3x(x − 4) − 2(x − 4)
= (x − 4)(3x − 2)
 
(f) −2x 2 +x + 3 = −2 × 3 = −6, (−2) + 3 = 1
= −2x 2 −2x + 3x + 3 = −2x(x + 1) + 3(x + 1)
= (x + 1)(3 − 2x)
(g) −x +5x − 4 = −x 2 +x + 4x − 4 = −x(x − 1) + 4(x − 1) = (x − 1)(4 − x)
2

(h) −6x 2 +5x + 4 = −6x 2 −3x + 8x + 4 = −3x(2x + 1) + 4(2x + 1)

= (2x + 1)(4 − 3x)
√ √ √ √ √ √
(i) x 2 + ( 2 + 3)x + 6 = x 2 + x 2 + x 3 + 2 × 3
√ √ √
= x(x + 2) + 3(x + 2)
√ √
= (x + 2)(x + 3)

Exercise 2.8, page 16

1. After writing x 2n as (x n )2 , we convert the bi-quadratic polynomial form into quadratic relative to
(x n ) as:
(a) x 4 −13x 2 + 36 = (x 2 )2 −4(x 2 ) − 9(x 2 ) + 36 = x 2 (x 2 − 4) − 9(x 2 − 4)
= (x 2 − 4)(x 2 − 9) = (x + 2)(x − 2)(x + 3)(x − 3)
(b) x 4 −10x 2 + 9 = x 4 −x 2 − 9x 2 − 9 = x 2 (x 2 − 1) − 9(x 2 − 1)
= (x 2 − 1)(x 2 − 9) = (x + 1)(x − 1)(x + 3)(x − 3)
28 2 Polynomials

(c) x 6 − 2x 3 + 1 = (x 3 )2 − 2x 3 + 1 = (x 3 − 1)2
 
= (x 3 − 1) ÷ (x − 1) = x 2 + x + 1
= [(x 2 + x + 1)(x − 1)]2 = (x 2 + x + 1)2 (x − 1)2
= (x + 1)(x − 1)(x 2 + x + 1)2
(d) x 6 +3x 3 + 2 = x 6 + x 3 + 2x 3 + 2 = x 3 (x 3 + 1) + 2(x 3 + 1)
= (x 3 + 1)(x 3 + 2)
√ √ √
= (x + 1)(x 2 − x + 1)(x + 2)(x 2 − x 2 + 4)
3 3 3

(e) x 6 +9x 3 + 8 = x 6 +x 3 + 8x 3 + 8 = x 3 (x 3 + 1) + 8(x 3 + 1)

= (x 3 + 1)(x 3 + 8) = (x + 1)(x 2 − x + 1)(x 3 + 8)
= (x + 1)(x 2 − x + 1)(x + 2)(x 2 − 2x + 4)
 2  
(f) 2x 2m+2 − 11x m+1 + 9 = 2 x m+1 − 11 x m+1 + 9
 2    
= 2 x m+1 − 2 x m+1 − 9 x m+1 + 9
   
= 2x m+1 x m+1 − 1 − 9 x m+1 − 1
  
= x m+1 − 1 2x m+1 − 9

Exercise 2.9, page 16

1. Completing the square technique exploits binomial square form.

(a) Complete the binomial square form as follows

x 2 + 6x − 7 = x 2 + 2(3)x +32 −32 − 7 = (x + 3)2 − 16 = 0

     
(x + 3)2 −16

Therefore, the quadratic equation is solved as

(x + 3)2 = 16 ∴ x + 3 = ±4 ∴ x1,2 = −3 ± 4 ∴ x1 = 1, x2 = −7

(b) The leading coefficient (i.e. “2”) is not equal to one so it can be factored. In addition, the
linear term coefficient “5” is a prime number, consequently it is not possible to factor “2” that is
necessary for the complete square form. The workaround is to multiply and divide “5” by two.
Solutions 29

  
3 5 3
2x − 10x − 3 = 2 x − 5x −
2 2
=2 x −2
2
x−
2 2 2
⎛ ⎞
  2  2
5 5 5 3
= 2 ⎝x 2 − 2 x+ − − ⎠
2 2 2 2
! " ! "
2 2
5 25 3 5 31
=2 x− − − =2 x− − =0
2 4 2 2 4

∴
 2  2
5 31 5 31
x− − =0 ∴ x− =
2 4 2 4
∴
√
5 31 5 31 5 ± 31
x− =± ∴ x1,2 = ± =
2 4 2 4 2

(c) Factor the leading coefficient “−1” as

−x 2 − 6x + 7 = −[x 2 + 2(3)x +32 −32 − 7] = −[(x + 3)2 − 16] = 0

∴
√
(x + 3)2 − 16 = 0 ∴ x + 3 = ± 16 ∴ x1,2 = −3 ± 4
∴
x1 = 1, x2 = −7

Note, however, that although the roots of −x 2 − 6x + 7 and x 2 + 6x − 7 (Example 2.9-1(a)) are
equal, the two quadratic polynomials are not the same.
(d) Force the form of complete square as
  2  2  2
3 3 3 3 9
x + 3x = x + 2
2 2
x+ − = x+ − =0
2 2 2 2 4
∴
3 3 3 3
x+ =± ∴ x1,2 = − ± ∴ x1 = 0, x2 = −3
2 2 2 2
30 2 Polynomials

(e) Factorize the leading coefficient, as well as 6 = 2 × 3, so that

!  2  "
2
3 3 3
2x + 6x + 2 = 2 x + 2
2
x+
2
− +1
2 2 2
! "
3 2 9
=2 x+ − +1 =0
2 4

∴
√
3 5 −3 ± 5
x+ =± ∴ x1,2 =
2 4 2

 2
1 25
(f) x2 − x − 6 = 0 ∴ x− − = 0 ∴ x1 = −2, x2 = 3
2 4

Exercise 2.10, page 16

1. After rearranging all terms given polynomials in descending powers, the long division is done in
the manner similar to division of large numbers.
(a) Given two polynomials, we proceed as follows.

1. Divide only the two highest power terms (underlined) and write the result, that is to say x 4 ÷
x = x3.

(x 4 − 2x 3 − 7x 2 + 8x + 12) ÷ (x − 3) = x 3

2. Multiply x 3 with the divisor polynomial and write the product under dividend

(x 4 − 2x 3 − 7x 2 + 8x + 12) ÷ (x−3) = x 3
x 4 − 3x 3 Multiply

3. Subtract the vertically aligned terms, i.e. (x 4 − 2x 2 ) − (x 4 − 3x 2 ) = x 3 , then add the next term
from dividend, i.e. −7x 2 . Now, the problem is reduced to division (x 3 − 7x 2 ) ÷ (x − 3)

(x 4 − 2x 3 −7x 2 + 8x + 12) ÷ (x − 3) = x 3
−(x 4 − 3x 3 )
= 0 + x 3 −7x 2
Solutions 31

4 Next term in the solution polynomial is found as x 3 ÷ x = x 2 , then repeat the previous steps
until the last term in the solution is found as

(x 4 − 2x 3 − 7x 2 + 8x + 12) ÷ (x − 3) = x 3 + x 2 − 4x − 4
−(x 4 − 3x 3 )
x 3 − 7x 2
− (x 3 − 3x 2 )
− 4x 2 + 8x
− (−4x 2 + 12x)
− 4x + 12
− (−4x + 12)
=0

Because the last subtraction equals zero means that r = 0, that is to say x = 3 is zero of
P (x) = x 4 − 2x 3 − 7x 2 + 8x + 12. It is verified by setting x = 3 in P (x) as

P (3) = (3)4 − 2(3)3 − 7(3)2 + 8(3) + 12 = 81 − 54 − 63 + 24 + 12 = 0

(b) Long division gives,

−3x + 3
(6x 3 + 10x 2 +0x + 8) ÷ (2x 2 + 1) = 3x + 5 +
2x 2 + 1
  
r

6x 3
+0x + 3x
2

10x 2 − 3x + 8
10x 2 +0x + 5

−3x + 3 = 0
32 2 Polynomials

(c) Long division gives,

(4x 3 − 7x 2 − 11x + 5) ÷ (4x + 5) = x 2 − 3x + 1

4x 3 + 5x 2

− 12x 2 − 11x
− 12x 2 − 15x

4x + 5
4x + 5

=0

(d) Long division gives,

1
(x 2 + x − 5) ÷ (x + 3) = x − 2 +
x+3
  
r

x + 3x
2

− 2x − 5
− 2x − 6

=1

2. One of the techniques to factorize odd order polynomials exploits two fundamental theorems in
algebra (very loosely interpreted as):

1. The total number of polynomial roots (i.e. both real and complex) is equal to the polynomial
degree, where the complex roots come in pairs, i.e. each complex root is accompanied by its
complex conjugate pair. Consequently, in odd order polynomials there must be at least one real
root (i.e. the one that does not have its pair).
2. After multiplying the leading and zero term coefficients, polynomial roots may be found
among the factors of that product.
Solutions 33

(a) Given P (x) = x 3 + x 2 − 4x − 4, the leading coefficient is “(+1)” and the zero power term
equals “(−4)”. Factors of −4 × 1 = −4 are {±1, ±4, ±2}. Thus we calculate P (x) for each
x = ±1, ±4, ±2 as

P (−1) = ((−1)3 + (−1)2 − 4(−1) − 4) = 0 ∴ x1 = −1

P (1) = ((1)3 + (1)2 − 4(1) − 4) = −6 ∴ P (1) = 0

P (−2) = ((−2)3 + (−2)2 − 4(−2) − 4) = 0 ∴ x2 = −2

P (2) = ((2)3 + (2)2 − 4(2) − 4) = 0 ∴ x3 = 2

P (−4) = ((−4)3 + (−4)2 − 4(−4) − 4) = −36 ∴ P (−4) = 0

P (4) = ((4)3 + (4)2 − 4(4) − 4) = 60 ∴ P (4) = 0

Therefore,
P (x) = x 3 + x 2 − 4x − 4 = (x − x1 )(x − x2 )(x − x3 ) = (x + 1)(x + 2)(x − 2).
As it happens, all three roots are real.

(b) Similarly, given Q(x) = x 3 − 1, factors of its zero power term are ±1. We calculate,

Q(−1) = (−1)3 − 1 = −2 ∴ Q(−1) = 0

Q(1) = (1)3 − 1 = 0 ∴ x1 = 1

That is to say,

Q(x) = (x − 1)(ax 2 + bx + c)
∴
(ax 2 + bx + c) = Q(x) ÷ (x − x1 ) = (x 3 ) ÷ (x − 1)

By the long division technique, we write

(x 3 − 1) ÷ (x − 1) = x 2 + x + 1
−(x 3 − x 2 )
x2 − 1
− (x 2 − x)
x−1
− (x − 1)
=0
34 2 Polynomials

Therefore, Q(x) = x 3 − 1 = (x − 1)(x 2 + x + 1). Discriminant of the second order

polynomial is negative, therefore the last two roots are complex,

Δ = b2 − 4ac = 12 − 4 × 1 × 1 = −3 < 0
∴
√ √
−b ± Δ −1 ± i 3
x2,3 = =
2a 2
∴
Q(x) = x 3 − 1 = (x − 1)(x 2 + x + 1)
 √  √ 
1+i 3 1−i 3
= (x − 1) x + x+
2 2

(c) Important polynomial form is known as the difference of two cubes. Given P (x) = x 3 − y 3
by inspection we conclude that one root of this binomial is x = y, i.e. P (y) = 0. Therefore, P (x)
is divisible by (x − y), we note that Example 2.10-(b) is special case when y = 1. Using long
division technique we derive this important identity,

(x 3 +0x 2 y +0xy 2 − y 3 ) ÷ (x − y) = x 2 + xy + y 2
x3 − x2y

x 2 y+0xy 2
x 2 y − xy 2

xy 2 − y 3
xy 2 − y 3

=0

Therefore, (x 3 − y 3 ) = (x − y)(x 2 + xy + y 2 ).
(d) Important polynomial form is known as the sum of two cubes. Given P (x) = x 3 + y 3 by
inspection we conclude that one root of this binomial is x = −y, i.e. P (−y) = 0. Therefore,
P (x) is divisible by (x + y). Using long division technique we derive this important identity,
Solutions 35

(x 3 +0x 2 y +0xy 2 + y 3 ) ÷ (x + y) = x 2 − xy + y 2
x3 + x2y

− x 2 y+0xy 2
−x 2 y − xy 2

xy 2 + y 3
xy 2 + y 3

=0

Therefore, (x 3 + y 3 ) = (x + y)(x 2 − xy + y 2 ).

Exercise 2.11, page 17

1. Difference/sum of cubes is factorized as follows, see Example 2.10-2.

(a) Q(x) = x 3 − 1 = (x − 1)(x 2 + x + 1)
(b) Q(x) = x 3 − y 3 = (x − y)(x 2 + xy + y 2 )
(c) Q(x) = x 3 + y 3 = (x + y)(x 2 − xy + y 2 )

Exercise 2.12, page 17

1. Factorization of higher order polynomials usually requires multiple steps and techniques. As a
consequence there is no unique path to find the solution, thus among number of possible ways to
solve the problem we aim to find ones that are elegant and creative.
(a) x 4 + 4 = x 4 +4x 2 + 4−4x 2 = (x 2 + 2)2 − (2x)2
= (x 2 + 2x + 2)(x 2 − 2x + 2)
(b) x 4 + x 2 + 1 = x 4 +2x 2 + 1−x 2 = (x 2 + 1)2 − (x)2
= (x 2 − x + 1)(x 2 + x + 1)
(c) x 5 + x + 1 = x 5 −x 2 +x 2 + x + 1 = x 2 (x 3 − 1) + x 2 + x + 1
 3 
x − 1 = (x − 1)(x 2 + x + 1) see Example 2.10-2(b)
= x 2 (x − 1)(x 2 + x + 1) + (x 2 + x + 1)
= (x 2 + x + 1)[x 2 (x − 1) + 1]
= (x 2 + x + 1)(x 3 − x 2 + 1)
36 2 Polynomials

(d) (x + y + z)3 − x 3 − y 3 − z3 = (x + y + z)3 − x 3 − (y 3 + z3 )

 3 
a − b3 = (a − b)(a 2 + ab + b2 )
=
a 3 + b3 = (a + b)(a 2 − ab + b2 )
= (x
+ y + z − x)[(x

+ y + z)2 + (x + y + z) x + x 2 ] − (y + z)(y 2 − yz + z2 )
 
= (y + z) (x + y + z)2 + (x + y + z) x + x 2 − y 2 + yz − z2
 
= (y + z) 2xy + 2xz + x 2 +y2 + 2yz+z2 + x 2 + xy + zx + x 2 
−y2 + yz −z2
   
= (y + z) 3xy + 3xz + 3yz + 3x 2 = 3(y + z) xy + xz + yz + x 2

= 3(y + z)[x(x + y) + z(x + y)] = 3(y + z)(x + y)(x + z)

(e) (x + 1)(x + 3)(x + 5)(x + 7) + 15 = (x + 1)(x + 7) (x + 3)(x + 5) + 15
= (x 2 + 8x + 7)(x 2 + 8x + 15) + 15
= (x 2 + 8x + 7)(x 2 + 8x + 7 + 8) + 15
 
= substitution: x 2 + 8x + 7 = t
= t (t + 8) + 15 = t 2 + 8t + 15
= t 2 + 3t + 5t + 15 = t (t + 3) + 5(t + 3) = (t + 3)(t + 5)
 
= back substitution: t = x 2 + 8x + 7
 
= x 2 + 8x + 7 + 3)(x 2 + 8x + 7 + 5
 
= x 2 + 8x + 10)(x 2 + 8x + 12
  
= x 2 + 2(4)x +42 − 42 + 10 x 2 + 2x + 6x + 12
 
= (x + 4)2 − 6 (x + 2)(x + 6)
 
= a 2 − b2 = (a + b)(a − b)
√ √
= (x + 4 − 6)(x + 4 + 6)(x + 2)(x + 6)
 2  
(f) x + x + 1 x 3 + x 2 + 1 − 1 = (x 2 + 1 + x)(x 3 + x 2 + 1) − 1
  2 
multiply by x + 1
 2      
= x + 1 x3 + x2 + 1 x2 + 1 + x2 + 1 x + x4 − 1
   2     
= x2 + 1 x3 + x2 + 1 + x2 + 1 x + x2 − 1 x2 + 1
 #    $
= x2 + 1 x3 + x2  + x + x2
+1 −1
 # $   
= x 2 + 1 x 3 + 2x 2 + x = x x 2 + 1 x 2 + 2x + 1
 
= x x 2 + 1 (x + 1)2
Solutions 37

Exercise 2.13, page 17

1. x n+2 x 
x n+1 x
(a)
n
= 
n+1
=
2x x 2
x 2
x 3n+2
+x 3n+1
x
  (x + 1)
3n+1
x+1
(b) =  =
x 2 x 3n xx 3n+1 x
x + x + x + ··· + x
2 3 n
x + x2 + x3 + · · · + xn
= n−1
(c) 1 1 1 1 x + x n−2 + x n−3 + · · · + 1
+ 2 + 3 + ··· + n
x x x x  xn 
x + x2 + x3 + · · · + xn xn
= n−1
x + x n−2 + x n−3 + · · · + 1
 ((( ( 
1 +(x 2(+( x 2(+(· (
· · + x n−1 x n+1
= ( ( ( = x n+1
n−2((n−3 ( ((((
n−1
+x( +x + ··· + 1
((((
x

Exercise 2.14, page 17

1. Partial fraction form of a rational function is very useful for calculating its integral. A general case
of rational expression of two polynomials, for example, k-th order Pk (x) and m-th order Qm (x),
where m > k, may be written as

Pk (x) P (x)
=
Qm (x) (x − x1 )(x − x2 )(x − x3 )n (ax 2 + bx + c)
α β γ1 γ2 γn δx + η
= + + + + ··· + +
x − x1 x − x2 (x − x3 )1 (x − x3 )2 (x − x3 )n (ax 2 + bx + c)

where, α, β, γ , . . . are the coefficient constants to be calculated, and partial decomposition

rational forms are created as follows:

1. For each unique zero of Q(x), such as x1 , x2 , there is one rational term.
2. For zeros with n multiplicity, such as x3 , there is a series of n rational terms.
3. When quadratic term (ax 2 + bx + c) has complex zeros, then its corresponding numerator
must be linear binomial.

(a) In this example, Q(x) has two unique zeros at x = −1 and x = 4, thus we write

P (x) x A B A(x − 4) + B(x + 1)

= = + =
Q(x) (x + 1)(x − 4) x+1 x−4 (x + 1)(x − 4)
Ax − 4A + Bx + B (A + B)x + B − 4A
= =
(x + 1)(x − 4) (x + 1)(x − 4)

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as
38 2 Polynomials

x ≡ (A + B)x + B − 4A
∴
term x 1 : 1 = A + B ∴ A = 1 − B
term x 0 : 0 = B − 4A ∴ B = 4A
∴
4 4 1
B = 4(1 − B), ∴ B = ∴ A=1− =
5 5 5

Consequently,

P (x) x 1 4
= = +
Q(x) (x + 1)(x − 4) 5(x + 1) 5(x − 4)

(b) In this example, Q(x) has one unique zero at x = 2 and one double zero x = 0, thus we
write

P (x) x+2 x+2  

= 3 = 2 = Q(x) = x 2 (x − 2) ∴ x1,2 = 0, x3 = 2
Q(x) x − 2x 2 x (x − 2)
A B C Ax(x − 2) + B(x − 2) + Cx 2
= + 2+ =
x x x−2 x 2 (x − 2)
Ax 2 − 2Ax + Bx − 2B + Cx 2 (A + C)x 2 + (B − 2A)x − 2B
= =
x 2 (x − 2) x 2 (x − 2)

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as

x + 2 ≡ (A + C)x 2 + (B − 2A)x − 2B
∴
term x 2 : 0 = A + C ∴ A = −C
term x 1 : 1 = B − 2A
term x 0 : 2 = −2B ∴ B = −1
∴
B = −1, ∴ 1 = (−1) − 2A ∴ A = −1 ∴ C = 1

Consequently,

P (x) x+2 x+2 1 1 1

= 3 = 2 =− − 2 +
Q(x) x − 2x 2 x (x − 2) x x x−2
Solutions 39

(c) Denominator Q(x) has two unique real zeros at x = 1 and x = −1, as well as two complex
zeros, thus we write

P (x) 2x 2   2x 2
= 4 = a 2 − b2 = (a − b)(a + b) = 2
Q(x) x −1 (x − 1)(x 2 + 1)
2x 2 A B Cx + D
= = + + 2
(x − 1)(x + 1)(x + 1)
2 x−1 x+1 x +1
A(x + 1)(x 2 + 1) + B(x − 1)(x 2 + 1) + (Cx + D)(x − 1)(x + 1)
=
(x − 1)(x + 1)(x 2 + 1)
(A + B + C)x 3 + (A − B + D)x 2 + (A + B − C)x + A − B − D
=
(x − 1)(x + 1)(x 2 + 1)

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as

2x 2 ≡ (A + B + C)x 3 + (A − B + D)x 2 + (A + B − C)x + (A − B − D)

∴
term x 3 : 0 = A + B + C
term x 2 : 2 = A − B + D
term x 1 : 0 = A + B − C
term x 0 : 0 = A − B − D

We solve the above system of equations, for example, by the reduction method. From first
equation we have 0 = A + B + C ∴ B = −A − C, then we write

2 = A − B + D ∴ 2 = A − (−A − C) + D = 2A + C + D
0 = A + B − C ∴ 0 = A + (−A − C) − C = −2C ∴ C = 0
0 = A − B − D ∴ 0 = A − (−A − C) − D = 2A − @
C − D = 2A − D ∴ 2A = D
2 = 2A + @
C + D ∴ D = 2 − 2A
1
2A = D ∴ 2A = 2 − 2A ∴ A = ∴ D=1
2
1
B = −A − @
C ∴ B=−
2

Consequently,

P (x) 2x 2 2x 2 1 1 1
= 4 = = − + 2
Q(x) x −1 (x − 1)(x + 1)(x + 1)
2 2(x − 1) 2(x + 1) x + 1
40 2 Polynomials

(d) In this case numerator polynomial has higher order than the denominator. For that reason,
first we use long division then partial fraction decomposition.

x+1
(x 3 + x 2 − 16x + 16) ÷ (x 2 − 4x + 3) = x + 5 +
x2 − 4x + 3
(−) x 3 − 4x 2 + 3x
5x 2 − 19x + 16
(−) 5x 2 − 20x + 15
x+1

Denominator Q(x) has two unique real zeros of x = 1 and x = 3, thus we write

P (x) x+1 x+1 A B

= 2 = = +
Q(x) x − 4x + 3 (x − 1)(x − 3) x−1 x−3
A(x − 3) + B(x − 1) (A + B)x − 3A − B
= =
(x − 1)(x − 3) (x − 1)(x − 3)

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as

x + 1 ≡ (A + B)x − 3A − B
∴
term x 1 : 1 = A + B ∴ A = 1 − B
term x 0 : 1 = −3A − B ∴ 1 = −3(1 − B) − B
∴
B = 2 ∴ A = −1

Consequently,

P (x) x 3 + x 2 − 16x + 16 1 2
= =x+5− +
Q(x) x 2 − 4x + 3 x−1 x−3

(e) An example of Q(x) with one unique real zero x = 1 and two complex zeros.

P (x) 1   1
= 3 = see Example 2.2-1 =
Q(x) x −1 (x − 1)(x 2 + x + 1)
Solutions 41

A Bx + C A(x 2 + x + 1) + (Bx + C)(x − 1)

= + 2 =
x−1 x +x+1 (x − 1)(x 2 + x + 1)
(A + B)x 2 + (A − B + C)x + A − C
=
(x − 1)(x 2 + x + 1)

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as

1 ≡ (A + B)x 2 + (A − B + C)x + A − C
∴
term x 2 : 0 = A + B ∴ A = −B
term x 1 : 0 = A − B + C ∴ 0 = (−B) − B + C ∴ C = 2B
1
term x 0 : 1 = A − C ∴ 1 = (−B) − (2B) ∴ B = −
3
2 1
C=− ∴ A=
3 3

Consequently,

P (x) 1 1 1 x+2
= 3 = = −
Q(x) x −1 (x − 1)(x + x + 1)
2 3(x − 1) 3(x + x + 1)
2

(f) An example of Q(x) with two complex pairs of zeros. Complete the squares to enforce
bi-quadratic form, then use difference of square identities to rearrange given fourth order
polynomial.

P (x) 4 4 4
= 4 = 4 = √ 2
Q(x) x +1 x +2x + 1 −2x
2 2
(x 2 + 1)2 − 2x

4 Ax + B Cx + D
= √ √ = √ + √
+ 1 − 2 x)(x + 1 + 2 x)
(x 2 2 x − 2x + 1 x + 2x + 1
2 2
√ √
(Ax + B)(x 2 + 2 x + 1) + (Cx + D)(x 2 − 2 x + 1)
= √ √
(x 2 − 2 x + 1)(x 2 + 2 x + 1)
√ √ √ √
(A + C)x 3 + ( 2A + B − 2C + D)x 2 + (A + C + 2B − 2D)x + B + D
= √ √
(x 2 − 2 x + 1)(x 2 + 2 x + 1)
42 2 Polynomials

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as
√ √ √ √
4 ≡ (A + C)x 3 + ( 2A + B − 2C + D)x 2 + (A + C + 2B − 2D)x + B + D
∴
term x 3 : 0 = A + C ∴ A = −C
√ √ √
term x 2 : 0 = 2A + B − 2C + D ∴ 0 = −2 2C + B + D
√ √ √ √
term x 1 : 0 = A + C + 2B − 2D ∴ 0 = + 2B − 2D ∴ B = D
term x 0 : 4 = B + D ∴ 4 = D + D ∴ D = 2 ∴ B = 2
√ √ √
∴ 0 = −2 2C + 4 ∴ C = 2 ∴ A = − 2

Consequently,
√ √
P (x) 4 − 2x + 2 2x + 2
= 4 = √ + √
Q(x) x +1 x − 2x + 1 x + 2x + 1
2 2

(g) An example of Q(x) with multiple complex pair of zeros.

P (x) x3 + x − 1 Ax + B Cx + D (Ax + B)(x 2 + 2) + Cx + D

= = 2 + 2 =
Q(x) (x + 2)
2 2 x +2 (x + 2)2 (x 2 + 2)2
Ax 3 + Bx 2 + (C + 2A)x + 2B + D
=
(x 2 + 2)2

Therefore, by equalizing numerators P (x) before and after decomposition, we calculate the
unknown constants as

x 3 + x − 1 ≡ Ax 3 + Bx 2 + (C + 2A)x + 2B + D
∴
term x 3 : A = 1
term x 2 : B = 0
term x 1 : 1 = C + 2A ∴ C = −1
0
 + D ∴ D = −1
term x 0 : −1 = 2B

Consequently,

P (x) x3 + x − 1 x x+1
= = 2 − 2
Q(x) (x + 2)
2 2 x + 2 (x + 2)2
 Linear Equations and Inequalities
3

Important to Know

Given, for example, system of two linear equations,

a1 x + b1 y = c1
a2 x + b2 y = c2

one of the techniques for calculating the solutions is known as Cramer’s rule

Δx Δy
x= ; y= ; (Δ = 0)
Δ Δ

where
% %
%a1 b1 % def
Δ=% % % = a1 b2 − a2 b1
a2 b2 %
% %
%c1 b1 % def
%
Δx = % % = c1 b2 − c2 b1
c2 b2 %
% %
%a1 c1 % def
%
Δy = % % = a1 c2 − a2 c1
a2 c2 %

3.1 Exercises

3.1 * Linear Equations

1. Comment on the solutions of the following equations

x−1 3−x 1 + 3x 6x + 3 x
(a) 1− = (b) − =
3 3 4 12 4
x 2x + 3 x2
(c) 3x + 14 = 5x − 2 (x − 7) (d) − =
x−2 x+2 4 − x2
© Springer Nature Switzerland AG 2021 43
R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_3
44 3 Linear Equations and Inequalities

2. Solve the following equations

1 16
=
1 1 37
1− =2 2+
(a) 1 (b) 1
1− 3+
1−x 3
4+
x+1

3.2 * Systems of Linear Equations

1. Solve the following systems of equations

⎧ 14 24 ⎧ 4 1
⎪ ⎪
⎨ x + y = 10
⎪ ⎨x + y − 1 + x − y + 1
⎪ =1
(a) (b)
⎪
⎪ ⎪
⎪
⎩ 7 − 18 = −5 ⎩ 18
−
5
=1
x y x + y − 1 2(x − y + 1)

3.3 * Linear Inequalities

1. Solve the following inequalities.

% %
% 4x % (x + 2) |x − 2| 1 1 x2 − 2
(a) %% %<3 (b) >1 (c) +
2x + 4 %
< 2
x2 + 2 x x+1 x +x
2. Solve the following equation and inequalities:
(a) −3x 2 + 30x − 75 > 0 (b) x 4 − 2x 2 + 1 < 0
2x 2 − 5x − 3
(c) −2x 2 + 4x − 2 > 0 (d) ≤0
−2x + 1

Solutions

Exercise 3.1, page 43

1. x−1 3−x 3−x+1 3−x

(a) 1− = ∴ = ∴ 4 − x
= 3 − x
∴ 4 = 3
3 3 3 3
Conclusion: there are no valid solutions for this equation.

1 + 3x 6x + 3 x 3
+ 9x − 6x − 3
3x
(b) − = ∴ = ∴  =
3x  ∴ 1=1
3x
4 12 4 12 12
Conclusion: this equation is true for any x, thus there are infinitely many solutions.

(c) 3x + 14 = 5x − 2 (x − 7) ∴  +
3x −
14 −
3x =0 ∴ 0=0
14
Conclusion: this equation is true for any x, thus there are infinitely many solutions.
Solutions 45

(d) In this example, the equation is not defined for x = ±2, because for all three denominators
x = ±2 results in the polynomial division limiting to infinity.

x 2x + 3 x2 x(x + 2) − (2x + 3)(x − 2) x2

− = ∴ = ∴
x−2 x+2 4 − x2 (x − 2)(x + 2) 4 − x2
−x 2 + 3x + 6 x2 x 2 − 3x − 6 x2
= ∴ =
x −4
2 4−x 2 4−x 2 4 − x2
Which implies that x 2 − 3x − 6 = x 2 ∴ 3x + 6 = 0. However, this equation is valid only if
x = 2, which is already excluded. Thus we must conclude that the equation in this example does
not have valid solutions.

2. “Telescopic” forms of equations are common in iterative processes.

(a) Group numbers, then invert the fractions.
1 1 1 1 1  
1− = ∴ =1− = ∴ invert both sides
1 2 1 2 2
1− 1−
1−x 1−x
1 1
∴ 1− =2 ∴ = −1
1−x 1−x
∴ 1 − x = −1 ∴ x = 2

(b) Invert fractions, group the numbers and repeat.

⇒ ∴
1 16 1 5
= =
1 37 1 16
2+ 3+
1 3
3+ 4+
3 x+1
4+
x+1 ∴
∴ 1 16
3+ =
1 37 3 5
2+ = 4+
1 16 x+1
3+
3 ∴
4+
x+1
1 16 1
∴ = −3=
3 5 5
4+
1 37 x+1
= −2
1 16 ∴
3+
3
4+ 3
x+1 4+ =5
x+1
∴ ⇒

3
Finally, we find =1 ∴ x+1=3 ∴ x =2
x+1
46 3 Linear Equations and Inequalities

Exercise 3.2, page 44

1. Use the variable substitution technique

(a) Given system, the inverse of variables is substituted by new variables
⎧ 14 24
⎪ 
⎨ x + y = 10
⎪ 
1 1
 14t + 24k = 10
∴ t= , k= ∴
⎪
⎪ x y 7t − 18k = −5
⎩ 7 − 18 = −5
x y

14t + 24k = 10   1
⇒ ∴ (2) − (1) ∴ 60k = 20 ∴ k =
14t − 36k = −10 3

∴

1 1
y = 3 ∴ 14t + 24 = 10 ∴ 14t + 8 = 10 ∴ t = ∴ x=7
3 7

(b) Two denominators are substituted as

⎧ 4 1
⎪
⎨x + y − 1 + x − y + 1 = 1
⎪ 
1 1

∴ t= , k=
⎪
⎪ 16 2 x+y−1 x−y+1
⎩ − =1
x+y−1 x−y+1

4t + k =1 1−k 1−k 1
∴ t= ∴  4
16 − 2k = 1 ∴ k =
16t − 2k = 1 4 4
1 2

1 1
∴ 4t + =1 ∴ t =
2 8
Putting back the original variables leads into the second system of equations
 
x+y−1 =8 x+y =9
∴ ∴ x = 5, y = 4
x−y+1 =2 x−y =1

Exercise 3.3, page 44

1. Comparison of two quantities may be done by comparing, for example, if their ratio is greater
than, equal to, or less than one, or if their difference is positive, zero or negative, etc. In addition,
absolute values are converted into two equations (inequalities).
Solutions 47

% %
% 4x % 4x
% %
(a) % 2x + 4 % < 3 ∴ −3 < 2x + 4 < 3

Therefore, there are two inequalities to solve.

4x
< 3 ∴ 4x < 6x + 12 ∴ −2x < 12 ∴ x < −6
2x + 4
4x 6
> −3 ∴ 4x > −6x − 12 ∴ 10x > −12 ∴ x > −
2x + 4 5

Conclusion: x ∈ [−∞, −6[ and x ∈] − 6/5, +∞], where the equalizing points at the interval
boundaries, i.e. x = −6 and x = −6/5, are not included.

(b) There are two inequalities to solve,

 
(x + 2) |x − 2| 1: x − 2 ≥ 0 ∴ x ≥ 2 ∴ |x − 2| = x − 2
>1
x2 + 2 2: x − 2 < 0 ∴ x < 2 ∴ |x − 2| = −(x − 2)
 
(x + 2)(x − 2) x2 − 4 x 2 1 − x42
(x ≥ 2): >1 ∴ >1 ∴   ≯1
x2 + 2 x2 + 2 x 2 1 + x22
because numerator is less than one and denominator is greater than one, thus their ratio is always
less than one.

−(x + 2)(x − 2) 4 − x2
> 1 ∴ > 1 ∴ 4 − x2 > x2 + 2
(x < 2) : x2 + 2 x2 + 2
∴ 4 − 2 > 2x 2 ∴ 1 > x 2 ∴ x < ±1
Conclusion: x ∈]−1, 1[, where the boundary points are not included. Indeed, plot of this function
shows that its value exceeds one only within the calculated interval, Fig. 3.1

Fig. 3.1 Example 3.1-1(b)

f ( x)

0 x

−1

−2 0 2
48 3 Linear Equations and Inequalities

(c) One possible way to solve this inequality relative to one is to convert it into its equivalent
inequality relative to zero. We note that x = 0 and x = −1 because of division by zero cases.

1 1 x2 − 2 x+1+x x2 − 2
+ < 2 ∴ < 2 ∴
x x+1 x +x x(x + 1) x +x
(2x + 1)(x 2 + x) < x(x + 1)(x 2 − 2) ∴
x(x + 1)(x 2 − 2) − (2x + 1)(x 2 + x) > 0 ∴
x(x + 1)[(x 2 − 2) − (2x + 1)] > 0 ∴ x(x + 1)(x 2 − 2x − 3) > 0 ∴
x(x + 1)2 (x − 3) > 0 the equivalent inequality f (x)

Obviously, there are zeros of this polynomial are x = 0, x = −1, x = 3 that should be excluded,
where x = −1 is double zero. We look for the intervals where this inequality f (x) is satisfied. In
factored form we look for sign of each factor and the sign of total product, i.e.

x [−∞, −1[ ] − 1, 0[ ]0; 3[ ]3; +∞]

x − − + +
(x + 1)2 + + + +
(x − 3) − − − +
f (x) + + − +

Conclusion, this inequality is valid for x ∈ [−∞, −1[, x ∈] − 1, 0[, and x ∈]3; +∞], where
x = 3, x = 0, and x = −1.
Indeed, zoom-in plot of f (x) function shows that it is positive only within the calculated intervals,
Fig. 3.2

Fig. 3.2 Example 3.1-1(c)

f (x)

0 x

−1 0 3
Solutions 49

2. Solve the following inequalities:

(a) Knowing that if A B = 0, then it must be that either A = 0, or B = 0, or both A and B
equal to zero. More general, if A B > 0, then either both A, B > 0 or both A, B < 0. Similarly,
if A B < 0, then A and B must have the opposite signs, i.e. one must be positive while the other
is negative. With that idea, we use the factorized forms of polynomials, so that there are product
terms that help us reach conclusions.

−3x 2 + 30x − 75 > 0 ⇔ −3(x 2 − 10x + 25) > 0

⇔ −3(x 2 − 5x − 5x + 25) > 0
⇔ −3[x(x − 5) − 5(x − 5)] > 0
⇔ − 3 (x − 5)2 > 0 ∴ not possible
     
<0 >0

because, product of negative and positive terms is always negative. Here, we have a negative
number (i.e. “−3”) multiplied by square of something (the squared number is always positive),
therefore, there is no real solution to this inequality.
(b) x 4 − 2x 2 + 1 = (x 2 )2 − 2x 2 + 1 = (x 2 − 1)2 = (x + 1)2 (x − 1)2 ≥ 0 ∀x
Therefore, x 4 − 2x 2 + 1 < 0 is never satisfied.

(c) −2x 2 + 4x − 2 = −2(x 2 − 2x + 1) = −2(x − 1)2 ≤ 0 ∀x

Therefore, −2x 2 + 4x − 2 > 0 is never satisfied.

(d) After converting into factorized form, we must verify all combinations that factors of that
rational function can have,

2x 2 −5x − 3 2x 2 −6x + x − 3 (2x + 1)(x − 3)

H (x) = = ≤0⇔ ≤0
−2x + 1 −2x + 1 −2x + 1
We systematically review all combinations when H (x) ≤ 0, i.e. the numerator and denominator
must have opposite signs, as

(2x + 1)(x − 3)(−2x + 1)H (x)

(1) + + − −
(2) − − − −
(3) + − + −
(4) − + + −
50 3 Linear Equations and Inequalities

therefore,

(1) : (2x + 1) > 0 and (x − 3) > 0 and (−2x + 1) < 0

∴ x > −1/2 and x > 3 and x > 1/2 ∴ x > 3
(2) : (2x + 1) < and (x − 3) < 0 and (−2x + 1) < 0
∴ x < −1/2 and x < 3 and x > 1/2 ∴ no solution
(3) : (2x + 1) > 0 and (x − 3) < 0 and (−2x + 1) > 0
∴ x > −1/2 and x < 3 and x < 1/2 ∴ −1/2 < x < 1/2
(4) : (2x + 1) < 0 and (x − 3) > 0 and (−2x + 1) > 0
∴ x < −1/2 and x < 3 and x < 1/2 ∴ no solution
In conclusion:

(2x + 1)(x − 3) 1 1
≤0 ∴ − ≤x≤ or x ≥ 3
−2x + 1 2 2
 Exponential and Logarithmic
Functions 4

Important to Know

Basic identities for log and exp functions,

ln e = 1
if: a f (x) = a g(x) ⇒ f (x) = g(x)
for: a > 1 and a f (x) > a g(x) ⇒ f (x) > g(x)
for: 0 < a < 1 and a f (x) > a g(x) ⇒ f (x) < g(x)
if: x = loga b ⇒ a x = b, (a, b > 0), (a = 1)
log 1 = 0
loga a = 1
a loga b = b, (a, b > 0), (a = 1)
log xy = log x + log y
log x n = n log x
x
log = log(x y −1 ) log x − log y
y
ln b
loga b =
ln a

© Springer Nature Switzerland AG 2021 51

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_4
52 4 Exponential and Logarithmic Functions

4.1 Exercises

4.1 * Exponential Function

1. Sketch graphs of the following functions

 x
1
(a) f (x) = 2 x
(b) f (x) = 2 − 2 x
(c) f (x) =
2
√ % %
(d) f (x) = 2 x2 (e) f (x) = %5|x| − 2% (f) f (x) = 3x + 3−x

4.2 * Exponential Equations and Inequalities

1. Solve the following equations.

(a) 2x = 8 (b) 2x−1 = 16 (c) 2x−5 = 3
(d) e7−4x = 6 (e) e2x − 3ex + 2 = 0 (f) 5x+1 − 5x−1 = 24

2. Solve the following equations.

 2
(a) x 2 − x − 1 x −1 = 1
(b) 2(x + 1)(2x + 1)x − (x − 1)x = (2x + 1)x+1
 
√ x √ x
(c) 5 + 24 + 5 − 24 = 10

3. Solve the following inequalities

 1/x 
1
1 < 3|x −x|
2
(a) 2x+2 > (b) <9 (c) 9x − 3x+2 > 3x − 9
4

4.3 * Logarithmic Function

1. Sketch graphs of the following functions

(a) f (x) = log2 x (b) f (x) = | log2 x| (c) f (x) = eln x

2. Without using calculator, calculated values of the following logarithms.

1 1
(a) log2 (b) log3 (c) log√2 8
128 81
√
3
√5
(d) log0.1 1000 (e) log2 512 (f) loga a 2
Solutions 53

4.4 * Logarithmic Equations and Inequalities

1. Solve the following equations (without a calculator),

(a) x = log10 10 (b) x = log10 100 (c) x = log10 1000
1 1
(d) x = log2 16 (e) x = log25 (f) x = log8
625 4
  
(g) x = log8 log4 log2 (16)

2. Solve the following equations (without a calculator),

√ 243
(a) x = 25log5 3 (b) x = log5 5 (c) x = log2/3
32
(d) x= log21/2 (4) (e) x= log1/2 (4)

3. Solve the following equations.

(a) x log3 x = 9 (b) log2 x = 4 (c) ln(3x − 10) = 2
(d) ln(x 2 − 1) = 3 (e) ln x + ln(x − 1) = 1 (f) ln(ln x) = 1
(g) log4 (x + 1) = 1

4. Solve the following equations.

2
(a) 52(log5 2+x) − 2 = 5x+log5 2 (b) 3(log3 x) + x log3 x = 162
(c) 81x − 16x − 2 × 9x (9x − 4x ) + 36x = 0

5. Solve the following inequalities.

x−1
(a) log >0 (b) log2x+3 x 2 < 1 (c) log2x 2 −x (2x + 2) < 1
x+2

Solutions

Exercise 4.1, page 52

1. Regardless of its base, exponential functions alone cross vertical axis x = 0 at y = 1, then they
may be shifted horizontally and/or vertically.
(a) Basic exponential function f (x) = 2x whose base is strictly greater than one, Fig. 4.1.
54 4 Exponential and Logarithmic Functions

Fig. 4.1
Example 4.1-1(a)(b)(c) 2x f (x)
2x−2
0.5x

1
0 x
−1

−1 0 1

(b) Basic exponential function shifted along vertical axis f (x) = 2x − 2, Fig. 4.1.
(c) Basic exponential function f (x) = 0.5x whose base is positive and strictly inferior to one,
Fig. 4.1.

(d) Simple transformation shows, Fig. 4.2, that


√ f (x) = 2x ; x≥0
x2 |x|
f (x) = 2 =2 ∴
−x
f (x) = 2 ; x≤0

Fig. 4.2 Example 4.1-1(d)

2x f (x)
2−x

0 x
−1 0 1
Solutions 55

(e) In Example 1-1(d) we found the shape of an absolute exponent. From there we can deduce
the following series of transformations
% %
5|x| → 5|x| − 2 → %5|x| − 2%

and derive the function graph in Fig. 4.3, where in the last step the negative portion of the graph
is flipped up in the positive region so that the absolute function is satisfied.

Fig. 4.3 Example 4.1-1(e)

f (x) f (x)
5|x| |5|x|−2|
5|x|−2

1 1
0 x 0 x
−1 −1
−1 0 1 −1 0 1

(f) Sum of two exponential forms f (x) = 3x + 3−x produces graph in Fig. 4.4.

Fig. 4.4 Example 4.1-1(f)

f (x)

3x
0 3−x x
3x + 3−x

Exercise 4.2, page 52

1. By using the properties of log and exp functions we find

(a) 2x = 8 ∴ 2x = 23 ∴ x = 3

(b) 2x−1 = 16 ∴ 2x−1 = 24 ∴ x − 1 = 4 ∴ x = 5

56 4 Exponential and Logarithmic Functions

  ln 3
(c) 2x−5 = 3 ∴ log2 2x−5 = log2 3 ∴ x − 5 = log2 3 =
ln 2
5 ln 2 + ln 3
∴ x=
ln 2
 
(d) e7−4x = 6 ∴ ln e7−4x = ln 6 ∴ 7 − 4x = ln 6 ∴ 7 − ln 6 = 4x
7 − ln 6
∴ x=
4

(e) This is quadratic equation in disguise, we use the change of variables technique to convert
the original equation into
 x 2  
e2x − 3ex + 2 = 0 ∴ e − 3 ex + 2 = 0 ∴ t = ex
∴ t 2 − 3t + 2 = 0
t 2 − 2t − t + 2 = 0
t (t − 2) − (t − 2) = 0
(t − 2)(t − 1) = 0 ∴ t1 = 2, t2 = 1
 
Case t1 = 2: ∴ ex1 = 2 ∴ ln ex1 = ln 2 ∴ x1 = ln 2
 
Case t2 = 2: ∴ ex2 = 1 ∴ ln ex2 = ln 1 ∴ x1 = 0

(f) After factorization, we find

5x+1 − 5x−1 = 24
5x−1 52 − 5x−1 = 24
 
5x−1 52 − 1 = 24
5x−1 × 24 = 24 ∴ 5x−1 = 1 ∴ x − 1 = 0 ∴ x = 1

2. (a) Equation a b = 1 is possible in two cases: (1) if a > 0 and b = 0, or (2) a = 1. Thus for the
first case we write
 x 2 −1
x2 − x − 1 = 1 ∴ x2 − x − 1 > 0 and x 2 − 1 = 0 ∴ x = ±1

however, only x = −1 satisfies both conditions.

Solutions 57

The second case leads into the following solutions

x 2 − x − 1 = 1 ∴ x 2 − x − 2 = 0 ∴ (x + 1)(x − 2) = 0
∴
x = −1, x = 2

Thus, x1 = −1 (double root), and x2 = 2.

(b) The equation is converted into the following form.

2(x + 1)(2x + 1)x − (x − 1)x = (2x + 1)x+1

2(x + 1)(2x + 1)x − (2x + 1)x+1 = (x − 1)x
(2x + 1)x [2(x + 1) − (2x + 1)] = (x − 1)x
# $
(2x + 1)x  +2
2x
1− −1
2x
= (x − 1)
x

(2x + 1)x = (x − 1)x

 x
2x + 1
=1
x−1

which is possible if x = 0, or

2x + 1
= 1 ∴ 2x + 1 = x − 1 ∴ x = −2
x−1

(c) The use of change of variables technique helps to convert this equation into the form of
quadratic equation.
 
√ x √ x
5+ 24 + 5− 24 = 10
  √ x 
√ 5 − 24 √ x
5 + 24  √ + 5 − 24 = 10
5 − 24
⎛ √ √ ⎞x 
(5 + 24)(5 − 24) √ x
⎝  √ ⎠ + 5 − 24 = 10
5 − 24
√ x 
25 − 24 √ x
 √ + 5 − 24 = 10
5 − 24

1 √ x
 √ x + 5 − 24 = 10
5 − 24
58 4 Exponential and Logarithmic Functions

 √ x
This form is converted into quadratic equation with 5 − 24 = t, which gives

1 √
+ t = 10 ∴ t 2 − 10 t + 1 = 0 ∴ t1,2 = 5 ± 24
t

Return to the original variable results in

 
√ x √ x √
5 − 24 =t ∴ 5− 24 =5− 24 ∴ x = 2

√ x √ 1
∴ 5− 24 =5+ 24 =
√ ∴ x = −2
5 − 24
 
1
x −n = n ; (a − b)(a + b) = a 2 − b2
x

3. Inequalities compare two functions within one or multiple intervals.

(a) The idea is to compare powers of the same base, i.e.

1 1/x
 −1/x
2x+2 > ∴ 2x+2 > 4−1/x ∴ 2x+2 > 22 ∴ 2x+2 > 2−2/x
4

Due to the same base, this inequality is equivalent to

2 2
x+2>− ∴ x+2+ >0
x x

There is discontinuity at x = 0 for the inverse function, thus we examine two intervals:

1. x > 0 : Obviously, within interval x ∈]0, +∞] the inequality is always satisfied because
all terms on the left side are positive. Therefore, one solution is interval x > 0.
2. x < 0 : Within interval x ∈ [−∞, 0[, terms including x are obviously negative; however,
there is one term that is positive (i.e. “+2”). We find that

2
x+2+ = 0 ∴ x 2 + 2x + 2 = 0 ∴ x1,2 ∈
/R
x

in other words, there is no change of sign. In conclusion, for x ∈ [−∞, 0[⇒ x+2+2/x < 0.
Therefore, the inequality is not satisfied for x < 0.
Solutions 59

Relative relationship between x + 2 and −2/x is shown in Fig. 4.5.

Fig. 4.5 Example 4.2-3(a)

f (x)

x
0

x−2
− 2/x

−2 0

(b) Double inequality with the absolute expression |x 2 − x| is decomposed in intervals x 2 − x ≥

0, and x 2 − x ≤ 0. In this specific example, however, the equality points are not included (the
problem has “strict” inequalities).

1. x 2 − x < 0 : Within this interval, x(x − 1) < 0, that is to say: either x < 0 and x − 1 > 0
(not possible), or x > 0 and x − 1 < 0 which is possible if x ∈]0, 1[ (strict inequality).
(Hint: AB < 0 means that A and B must have opposite signs.)
2. x 2 − x > 0 : Within this interval, x(x − 1) > 0, that is to say: either x < 0 and x − 1 < 0
thus x < 0, or x > 0 and x − 1 > 0 thus x > 1. (Hint: AB > 0 means that A and B must
have same signs.)

Given these intervals, inequalities are solved after converting it into the form with same base, i.e.

1 < 3|x −x|

< 9 ∴ 30 < 3|x −x|
2 2
< 32 ∴ 0 < |x 2 − x| and |x 2 − x| < 2

Which leads into x 2 − x − 2 < 0 resulting in x > −1 and x < 2 interval. Putting everything
together, we conclude that x ∈] − 1, 2[ where, x = −1, 0, 2.
60 4 Exponential and Logarithmic Functions

(c) Following the idea of reducing all terms to the same base, we write

9x − 3x+2 > 3x − 9
 %2
%
32x − 3x 32 > 3x − 32 %
%

3

2x
3
− 3x 32 > 2x
− 2 × 3x 32 + 34
2 × 3x 32 − 3x 32 > 34

3x 32 > 34 2
∴
x>2

Exercise 4.3, page 52

1. Plots of these three functions are shown in Fig. 4.6. We note that all logarithmic functions have
zero at x = 1, and that exponent of logarithmic function (and vice versa) is a linear function.

Fig. 4.6
Example 4.3-1(a)(b)(c) f (x)

x
0

| log2x |
elnx
log2x
0 1

2. By definitions of exp and log functions, we write

1 1 1
∴ 2x = 2AHH2 128 = 7 = 2−7 ∴ x = −7
1
(a) x = log2 log
∴ 2x =
128 128 2
H 1 1
3 = 3A H3 81 ∴ 3 = −4
1
(b) x log x
= 4 =3 ∴ x = −4
81 3
√ x √ H √ 6
2 = ZZ
√
(c) x = log√2 8 ∴ 2 logH2 8 = 8 = 2 ∴ x = 6
(d) x = log0.1 1000 ∴ 0.1x = 1000 = 0.1−3 ∴ x = −3
√ √ 9
x = log2 512 ∴ 2x = 512 = 2 3 = 23 ∴ x = 3
3 3
(e)
√ √ 2
x = loga a 2 ∴ a x = a 2 ∴ x =
5 5
(f)
5
Solutions 61

Exercise 4.4, page 53

1. In examples (a) to (c) note the relationship between the numbers of zeros in 10, 100, 1 000, ... and
the respective equation solutions. In this special case of logarithmic functions with the base 10 and
decade numbers, the noted shortcut helps us to perform mental calculations of these logarithms.
(a) x = log10 10 ∴ 10x = 10 ∴ x = 1
(b) x = log10 100 ∴ 10x = 100 ∴ x = 2
(c) x = log10 1000 ∴ 10x = 1000 ∴ x = 3
In addition, writing “log” without specific index number assumes the base 10, while writing “ln”
assumes the base e. For example, x = log 1,000,000 is found by counting the zeros, therefore
x = 6.
Examples (d) to (g) illustrate some simple techniques for calculating various logarithms based on
the knowledge of relationships between numbers, i.e. squares, roots, the multiplication/division
table, and the basic log identities
(d) x = log2 16 ∴ 2x = 16 ∴ 2x = 24 ∴ x = 4

(e)
1
x = log25
1
= log25 2 = log25 25−2 = −2  log : 1 = −2
25 25
625 25
1 1 1 2 :
 1 2
(f) x = log8 = log8 √ = log8 √ = log8 8− /3 = −  log 8 =−
2
3 3 8
4 64 82 3 3
  
(g) Given x = log8 log4 log2 (16) we start from the right side, as
 
   4 
x = log8 log4 log2 2 = log8 log4 4  log :1
2 (2)

 
= log8 log :

4
1
⇒ x = log8 (1) ⇒ 8x = 8log8 (1) ⇒ 8x = 1
4

∴
x=0
√
2. Exercises with problems that include both log and n a functions, as well as absolute values of
numbers.
 log 3
(a) x = 25log5 3 = 52 5 = 5(2 log5 3) = 5log5 9 = 9
√ 1
(b) x = log5 5 = log5 51/2 =  log :1 = 1

55
2 2
 −5 >

1
(c) x = log 243 = log 3 = log
5
2 2
2/3 2/3 2/3 = −5 log 2/3 = −5
32 25 3  3
% %
%  −2 %% %% 1%%
% % % % >

1 % % 1 %
(d) x = log21/2 (4) = %log1/2 (4)% = %log1/2 % = %−2 log %=2
% % % 2 %
1/2

% 
2
%
+
* ,
 −2 , >

1
(e) 1 - 1 √ √
x = log1/2 (4) = log1/2 = −2 log 1/2 = −2 = j 2
2  2
62 4 Exponential and Logarithmic Functions

3. By using the properties of log and exp functions we find

   
(a) x log3 x = 9 ∴ log3 x log3 x = log3 9 ∴ log3 x log3 x = 2
 2 √
∴ log3 x log3 x = 2 ∴ log3 x = 2 ∴ log3 x = ± 2
√ √
∴ 3log3 x = 3± 2
∴ x = 3± 2

(b) log2 x = 4 ∴ 2AHH2 x = 24 ∴ x = 24

log

(c) Functions ln and exp are cancelling each other, similar as add/subtract, or multiply/divide,
thus we write

e2 + 10
ln(3x − 10) = 2 ∴ eln(3x−10) = e2 ∴ 3x − 10 = e2 ∴ x =
3


2
−1)
(d) ln(x 2 − 1) = 3 ∴ eln(x = e3 ∴ x 2 − 1 = e3 ∴ x = ± e3 + 1

(e) ln x + ln(x − 1) = 1 ∴ ln[x(x − 1)] = 1 ∴ eln[x(x−1)] = e

x2 − x − e = 1
∴ √
1 ± 1 + 4e
x1,2 =
2

(f) ln(ln x) = 1 ∴ eln(ln x) = e ∴ ln x = e ∴ eln x = ee ∴ x = ee

(g) log4 (x + 1) = 1 ∴ 4AHH4 (x+1) = 41 ∴ x + 1 = 4 ∴ x = 3

log

4. Convert all terms to the same base.

(a) A number can be written as exponent of an arbitrary base logarithm, i.e. 2 = 5log5 2 , thus

52(log5 2+x) − 2 = 5x+log5 2

.  b /
a b+c = a b a c ; a = nlogn a ; x ab = x a

52 log5 2 52x − 5log5 2 = 5x 5log5 2

 2
5log5 2 52x − 5log5 2 − 5x 5log5 2 = 0
 
5AHH5 2 5log5 2 52x − 1 − 5x = 0
log

 
2 5log5 2 52x − 1 − 5x = 0
∴
5AHH5 2 52x − 1 − 5x = 0
log
Solutions 63

Which is further converted into a quadratic equation.

 2  
2 · 5x − 5x − 1 = 0 5x = t
2t 2 − t − 1 = 0 ∴ 2t 2 − 2t + t − 1 = 0 ∴ 2t (t − 1) + (t − 1) = 0
∴
1
(t − 1)(2t + 1) = 0 ∴ t1 = 1 or t2 = −
2
∴
1
5x = − not possible because 5x > 0 ∀x
2
5x = 1 ∴ x = 0

2
(b) We recall that a b = a b×b = (a b )b , thus

2  log4 x
4(log4 x) + x log4 x = 512 ∴ 4log4 x + x log4 x = 512 ∴ 2 x log4 x = 512
∴
 
x log4 x = 256 ∴ log a b = b log a
 
log4 x log4 x = log4 256 ∴ log4 x log4 x = log4 44 = 4 ∴ log4 x = ±2
∴
1
x = 4±2 ∴ x1 = 16, x2 =
16

(c) After factorization, this equitation is transformed into quadratic equation.

81x − 16x − 2 9x (9x − 4x ) + 36x = 0

(34 )x − (24 )x − 2 92x + 2 9x 4x ) + (62 )x = 0
34x − 24x − 2 34x + 2 32x 22x + 62x = 0
34x − 2 34x − 24x + 2 62x + 62x = 0
−34x − 24x + 3 62x = 0
%
%
34x + 24x − 3 62x = 0 %% /24x

34x 2
4x
2
32x 2x
+ − 3 =0
24x 2
4x 2
4x 2x
! "2  2x
3 2x 3
−3 +1=0
2 2
64 4 Exponential and Logarithmic Functions

Thus, change of variable produces

 2x
√
3 3± 5
t= ∴ t − 3t + 1 = 0 ∴ t1,2
2
=
2 2

Return to the original variable gives

 2x  x
√
3 9 3± 5
= =
2 4 2
∴
√ √
9 3± 5 ln(3 ± 5) − ln 2
x1,2 ln = ln ∴ x1,2 =
4 2 ln 9 − ln 4

5. (a) Logarithmic function is positive when its argument is superior to one, i.e.

x−1 x−1 x−1

log >0 ∴ > 1 where, x = −2 and lim =1
x+2 x+2 ±∞ x + 2

We conclude that for x ∈ [−∞, −2[ it follows that always (x −1)/(x +2) > 1, thus the inequality
is satisfied. When x > −2 the inequality is not satisfied.
(b) We distinguish logarithms whose base is inferior to “1” from those whose base is superior
to one. In the case of inverse bases, the two log functions are symmetric and opposite sign.

if 0 < 2x + 3 < 1 ∴ x ∈] − 3/2, 1[, then

log2x+3 x 2 < 1 = log2x+3 (2x + 3) ∴ x 2 > 2x + 3 ∴ (x < −1) and (x > 3)
∴
x ∈] − 3/2, −1[
if 2x + 3 >1 ∴ x > −1, then
log2x+3 x 2 < 1 = log2x+3 (2x + 3) ∴ x 2 < 2x + 3 ∴ −1 < x < 3 and x = 0

Therefore, the solution is x ∈] − 3/2, 3[ (x = 0).

Solutions 65

(c) Logarithm function of an arbitrary base may be evaluated by converting given function into
ratio of two logarithmic functions as
 
ln b ln(2x + 2)
log2x 2 −x (2x + 2) < 1 loga b = ∴ <1
ln a ln(2x 2 − x)

In addition, we know that ln(1) = 0, that ln x < 0 if (0 < x < 1), and that ln x > 0 if (x > 1).
With that understanding we evaluate both logarithms in this function as follows.

1. Exclude points when denominator equals to zero, when the result of division becomes either
infinity or undetermined, thus

ln(2x 2 − x) = 0 ∴ 2x 2 − x = 1 ∴ 2x 2 − x − 1 = 0
1
∴ x = 1, x = −
2
1
ln(2x 2 − x) < 0 ∴ 0 < x(2x − 1) < 1 ∴ 0 < x <
2

and, this logarithm function is defined for positive argument, i.e.

1
2x 2 − x > 0 ∴ x(2x − 1) > 0 ∴ x < 0 or, x >
2

2. Similarly, logarithm in the numerator is defined for

2x + 2 > 0 ∴ x > −1 and,

1
ln(2x + 2) < 0 ∴ 0 < 2x + 2 < 1 ∴ x < −
2

1. Condition of inequality gives

ln(2x + 2)
f (x) = < 1 ∴ ln(2x + 2) < ln(2x 2 − x)
ln(2x 2 − x)
∴ 2x + 2 < 2x 2 − x
∴ 0 < 2x 2 − 3x − 2
∴ 0 < 2(x + 1)(x − 2)

Which is satisfied when x > 2, because values x < −1 are already excluded.
66 4 Exponential and Logarithmic Functions

(c) (cont.) With the above intervals, and knowing signs of numerator and denominator in each
interval, we summarize the intervals when inequality is satisfied, see Fig. 4.7, as

1
−1 < x < 0 and < x < 1 and x > 2
2

Fig. 4.7 Example 4.4-5(a)

f (x)

x
0

−1 0 1/2 1 2
 Trigonometry
5

Important to Know

Among many trigonometric identities, some are more often used than the others,

a 2 + b = c2 (Pythagoras’s theorem)
sin2 α + cos2 β = 1 (Pythagoras’s theorem in disguise, for c = 1 )

| sin α| = 1 − cos2 α

| cos α| = 1 − sin2 α
sin α
tan α =
cos α

Sum to product identities,

sin(α + β) = sin α cos β + cos α sin β

sin(α − β) = sin α cos β − cos α sin β
cos(α + β) = cos α cos β − sin α sin β
cos(α − β) = cos α cos β + sin α sin β

When α = β, special cases of the sum identities (note: double angle),

sin 2α = 2 sin α cos α

cos 2α = cos2 α − sinα

Product to sum identities,


1
sin α sin β = cos(α − β) − cos(α + β)
2

1
cos α cos β = cos(α + β) + cos(α − β)
2

© Springer Nature Switzerland AG 2021 67

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_5
68 5 Trigonometry


1
sin α cos β = sin(α + β) + sin(α − β)
2

Square to double angle identities,

1 − cos(2α)
sin2 α =
2
1 + cos(2α)
cos2 α =
2

5.1 Exercises

5.1 Basic Definitions

1. Explain the concept of “similar triangles” (thus, “similar figures” in general) and illustrate the
consequent properties that are used in trigonometry.
2. Elementary trigonometric definitions.
(a) Using right-angled triangles and Pythagoras’s theorem illustrate the definitions of sin x, cos,
and tan x
(b) Calculate sin x, cos, and tan x for each of the following angles in the first quadrant:
0◦ , 30◦ , 45◦ , 60◦ , and 90◦ .
(c) Sketch the unit circle and illustrate how angles in I I , I I I , and I V quadrant are related to
angles found in the first quadrant.

3. Convert degrees into radians

(a) 300◦ (b) 330◦ (c) −18◦

4. Convert radians into degrees

3π 5π
(a) (b) (c) 2
4 12
5. Calculate exact values

π 7π 3π
(a) tan (b) sin (c) cos −
3 6 4
5.1 Exercises 69

6. Given right-angled triangle in Fig. 5.1 express lengths of sides a and b as a function of angle θ.

Fig. 5.1 Example 5.1-6

10
b

7. Given α, is it possible to have

1 3
(a) sin α = − et cos α =
5 5
4 1
(b) sin α = √ et cos α = √
17 17
√ √
15 5
(c) sin α = − et cos α =
5 5
8. Calculate the exact values of cos α and tan α given that

1 3π
sin α = − et π < α <
3 2

9. Calculate the following:

(a) cos 150◦ (b) sin 120◦ (c) tan 300◦
(d) tan 225◦ (e) tan 480◦

10. Calculate the following:

(a) tan 840◦ (b) cos 1320◦
π π π π π π
(c) sin sin2 (d) cos cos + sin cos
6 3 4 6 4 6
π 1 π π π π
(e) sin2 − π (f) sin cos + cos tan
3 tan2 4 4 6 3
3
70 5 Trigonometry

5.2 * Trigonometry Identities

1. Calculate the following trigonometry functions (without a calculator),

π  π 
(a) cos (b) tan (c) cos(150◦ )
3 4
(d) sin(120◦ ) (e) tan(−45◦ ) (f) tan(−60◦ )
(g) tan(300◦ ) (h) tan(225◦ ) (i) tan(480◦ )
π  π 
(j) tan(840◦ ) (k) cos(1320◦ ) (l) sin sin2
6 3
2. Calculate the following expressions (without a calculator),
   π 
7π 13π 5π
(a) −4 sin sin (b) 4 sin cos
12 12 8 8
 
5π 3π
(c) 4 cos cos (d) tan 20◦ tan 40◦ tan 80◦
8 8

5.3 Sum Identities: sin(x ± y), cos(x ± y)

Identities for trigonometric functions of sum or difference of two angles are

sin (a ± b) = sin a cos b ± cos a sin b

cos (a ± b) = cos a cos b ∓ sin a sin b

1. Calculate sin α, cos α, and tan α of the following angles.

(a) α = 15◦ (b) α = 75◦ (c) α = 105◦
(d) α = π/12 (e) α = 5π/12 (f) α = 7π/12

2. Calculate the following expressions

(a) sin 20◦ cos 10◦ + cos 20◦ sin 10◦ (b) cos 43◦ cos 13◦ + sin 43◦ sin 13◦

3. Calculate sin(x + y) and sin(x − y) given that cos x = 4/5 and sin y = −3/5, where x is in IV
quadrant and y is in III quadrant.
4. Simplify the following expression
√ π 
2 cos α − 2 cos +α
π  √4
2 sin + α − 2 sin α
4

5. Derive the double-angle identities for cos 2a, sin 2a, and tan 2a relative to the single-angle
functions.
5.1 Exercises 71

5.4 Product to Sum Identities

1. Knowing the sum identities, derive the identities for the following products
(a) sin α sin β (b) cos α cos β (c) sin α cos β

2. Knowing the sum identities, derive the identities for the following sums
(a) sin x + sin y (b) cos x + cos y

3. Calculate the following products.

5π 7π 1  π  π
(a) −4 sin cos (b) + 2 sin 1 + cos 1 +
12 12 2 6 3
(c) 8 sin 20 sin 40 sin 80◦
◦ ◦
(d) ◦ ◦
8 cos 10 cos 50 cos 70 ◦

√ √ √
4. Solve for α in I quadrant if: tan α = 6 + 3 − 2 − 2.

5.5 Trigonometric Equations

1. Solve the following equations for x.

(a) sin x = 0 (b) cos x = 0

2. Solve the following equations for x.

(a) sin x = sin α (b) sin x = cos x (c) sin x = sin 2x
(d) 2 sin x + sin x = 0
2
(e) cos (x + π/6) = sin (x − π/3)

3. Solve the following equations for x.

π  π 
(a) cos − x = cos ; x ∈ [ 0; 2π [
3 6
 π 
(b) sin (3x) = sin x − ; x ∈ ] − π; π ]
2

5.6 Trigonometric Inequalities

1. Solve the following inequalities for x.

√
(a) 2 cos x + 3 < 0 (b) 2 sin x − 1 > 0
√
cos x − sin x < 1 3
(c) (d) sin 3x − ≥0
2
1
(e) | sin x| ≥ (f) | sin x| ≤ sin x + 2 cos x
2
72 5 Trigonometry

Solutions

Exercise 5.1, page 68

1. Two figures are said to be similar if one figure can be derived from the other by operations
of uniformly scaling, translation, rotation and/or reflection. For example, starting with triangle
OMN in Fig. 5.2, triangle OP Q is derived after multiplying each side of OMN by factor of
two, similarly any other triangle can be derived by using factor n, for example triangle OKL.
By inspection of Fig. 5.2 it is evident that except for the scaling factors the form of all three
triangles did not change, thus, they are similar. By using Pythagoras’s theorem it is not difficult
to verify that all three triangles are indeed right-angled whose corresponding sides are simply
proportional to each other.

Fig. 5.2 Example 5.1-1

K

5
n×
β

n×4
×5
M

2 ×4
2

β
5

α
O N Q L
3
2 ×3
n×3

We take a look at angles as well. Obviously, each triangle has one of its angles equal to 90◦ . In
addition, the corresponding sides of α in each triangle are proportional, thus α is shared by all
three triangles. Consequently, knowing two angles in a triangle the third angle β is calculated so
that their sum adds to 90◦ + α + β = 180◦ . In conclusion, direct consequence of similarity is that
not only the figure sides are proportional (including the factor n = 1) but also their corresponding
angles are equal. The similarity of right-angled triangles is among most important and exploited
properties in mathematics.
Solutions 73

2. By inspection of Fig. 5.2 we write ratios among triangle sides.

(a) There are three basic ratios,

1. The cathetus forming α (i.e. adjacent) and hypotenuse (the side opposite the right angle)
are related as

ON
=
OQ
=
OL 3
= =
6
3
= ··· =
× 3 = const.
n
OM OP OK 5 5

10
×5
n

2. The cathetus opposite to α and hypotenuse are related as

MN PQ KL 4 8
4 n×4
= = = = = ··· =
= const.
OM OP OK 5 

10 5
×5
n

3. The catheti opposite and adjacent to α are related as

MN
=
PQ
=
KL 4
= =
8
4
= ··· =
× 4 = const.
n
ON OQ OL 3 6
3
×3
n

The fact that these three ratios are constant for all corresponding similar triangles is very useful
because, given a right-angled triangle, by knowing the side ratios we can calculate various
unknown sides and angles.
Thus, relative to α we name these three ratios as

adjacent opposite opposite

= cos α, = sin α, = tan α
hypotenuse hypotenuse adjacent

Similarly, the equivalent ratios exist relative to β,

adjacent opposite opposite

= cos β, = sin β, = tan β
hypotenuse hypotenuse adjacent

We search these ratios related to α et β when triangles are rotated, scaled, translated and/or
reflected by looking for “adjacent” and “opposite” sides relative to the angle under consideration.
In addition, it is useful to notice that by knowing sin x and cos x, the third ratio, i.e. tan x is also
known because

opposite
sin x ( ((((
hypotenuse opposite
= = = tan x
cos x adjacent adjacent
( ( ( (
(
hypotenuse
74 5 Trigonometry

(b) Angles between 0◦ and 90◦ are said to be “in the first quadrant” of the unit circle. They are
reused to calculate trigonometric functions of angles in other three quadrants by rotating their
respective right-angled triangles into the first quadrant.
Case: x = 0◦ In this special case, right-angled triangle whose hypothenuse equals one degener-
ates into a horizontal line segment, as a consequence the hypotenuse is found “on top” of adjacent
side, thus the two are equal. At the same time, the opposite side of the triangle is reduced to zero.
By definitions we write

 
adjacent
cos 0◦ = = 1,
( ((((
hypotenuse
0
sin 0◦ = = 0,
hypotenuse
sin 0◦ 0
tan 0◦ = = =0
cos 0◦ 1

Case: x = 30◦ = π/6 We calculate trigonometric functions with the help of triangle in Fig. 5.3.
Here, hypotenuse is two times longer than a one of the cathetus, thus the two associated angles
must be in the same proportion. We conclude the following:

90◦ + α + β = 180◦
α = 2β
∴
90◦ + 3β = 180◦
β = 30◦ ∴ α = 60◦

In addition, Pythagoras’s theorem gives


2= 12 + b
∴
√
a= 3

By definitions we write
√
◦ adjacent 3
cos 30 = =
hypotenuse 2
opposite 1
sin 30◦ = = ,
hypotenuse 2
opposite 1
tan 30◦ = =√
adjacent 3
Solutions 75

(b) (cont.)
Case: x = 60◦ = π/3 We calculate trigonometric functions of this angle also with the help of
triangle in Fig. 5.3, it is also found in the same triangle. Again, knowing the three sides by
definitions we write

adjacent 1
cos 60◦ = =
hypotenuse 2
√
◦ opposite 3
sin 60 = = ,
hypotenuse 2
√
◦ opposite 3
tan 60 = =
adjacent 1

Fig. 5.3 Example 5.1-2

30 ◦

2
3

60 ◦
1

Case: x = 45◦ = π/4 Here, the two catheti are equal, thus the two associated angles must be in
the same proportion. By Pythagoras’s theorem, we calculate hypotenuse
 √
c= 12 + 12 = 2

Therefore, by definition we conclude the following (Fig. 5.4):

√
◦ adjacent 1 2
cos 45 = =√ =
hypotenuse 2 2
√
◦ opposite 1 2
sin 45 = =√ =
hypotenuse 2 2
opposite 1

tan 45◦ = = =1
adjacent 1

76 5 Trigonometry

(b) (cont.)

Fig. 5.4 Example 5.1-2

45 ◦

√
2
1

45 ◦

Case: x = 90◦ In this special case, right-angled triangle whose hypothenuse equals one degener-
ates into a vertical line segment, as a consequence the hypotenuse is found “on top” of opposite
side, thus the two are equal. At the same time, the adjacent side of the triangle is reduced to zero.
By definitions we write

0
cos 90◦ = = 0,
hypotenuse

sin 90◦ =  (( = 1,
opposite
( ((
hypotenuse
opposite 1
tan 90◦ = = =∞
adjacent 0

(c) Trigonometric functions of sin and cos are commonly summarized in a graph known as “unit
circle”, see Fig. 5.5. The circle radius, i.e hypotenuse, is normalized to one. Consequently, the
length of adjacent side becomes numerically equal cos α and the length of opposite side becomes
numerically equal to sin α, because

adjacent adjacent
cos α = = ∴ adjacent = cos α
hypotenuse 1
opposite opposite
sin α = = ∴ opposite = sin α
hypotenuse 1
Solutions 77

(c) (cont.)
In the first quadrant we find angles 0◦ , 30◦ , 45◦ , 60◦ , and 90◦ , see Fig. 5.5. Relative to the first
quadrant, angles in the second quadrant are shifted by 90◦ , thus we find angles:

30◦ + 90◦ = 120◦

45◦ + 90◦ = 135◦
60◦ + 90◦ = 150◦
90◦ + 90◦ = 180◦

Angles in the third quadrant are shifted again by another 90◦ , thus we find angles:

30◦ + 180◦ = 210◦

45◦ + 180◦ = 225◦
60◦ + 180◦ = 240◦
90◦ + 180◦ = 270◦

Angles in the third quadrant are shifted again by the total of 270◦ , thus we find angles:

30◦ + 270◦ = 300◦

45◦ + 270◦ = 315◦
60◦ + 270◦ = 330◦
90◦ + 270◦ = 360◦ = 0◦

The symmetry among these “key” angles is clearly visible in the graph, where we can visualize
the associated right-angled triangle for each angle and deduce its corresponding values of sin x
and cos x, as listed in Fig. 5.5.

(cos, sin) (cos, sin)

√  II I (0, 1) √ 
−1/2, 3/2 1 90◦ 1/2, 3/2 
√ √  120◦ 60◦ √ √
− √2/2, 2/2 135◦ 45◦ √2/2, 2/2 
− 3/2, 1/2 150◦ 30◦ 3/2, 1/2

180◦ 0◦
(−1, 0) (1, 0)
−1 1
√  √ 
− 3/2, √
√ −1/2 210◦ 330◦ √3/2, −1/2√ 
− 2/2, −√2/2 225◦ 315◦ 2/2, √
− 2/2

−1/2, − 3/2 240◦ 300◦ 1/2, − 3/2
270◦ −1
III IV (0, −1)

Fig. 5.5 Example 5.1-2

78 5 Trigonometry

(c) (cont.)
For example, an angle β = 90◦ + α in the second quadrant and its corresponding α angle in
the first quadrant create similar right-angled triangles, see Fig. 5.6. By definitions we read its
sin β value as the segment length at the vertical axis, and its cos β value as the segment length at
the horizontal axis. After comparing side lengths of similar triangles in second and first quadrant
(marked in solid red lines), while accounting for the positive/negative signs, we find the following
identities
 π
sin α + = sin β = cos α
2
 π
− cos α + = − cos β = sin α
2
where the negative sign of cos β is direct consequence of its position in the second quadrant. In
conclusion, by knowing sin α and cos α in the first quadrant, we easily conclude that cos α =
sin(α + 90◦ ) and sin α = − cos(α + 90◦ ).
For example,
⎫
1
cos 120◦ = − sin 30◦ = − ⎪⎪ √
2⎬ ◦ 3/2
√
√ ∴ tan 120 = =− 3
3 ⎪⎪
⎭
−1/2

sin 120◦ = cos 30◦ =
2

Fig. 5.6 Example 5.1-2 90◦ 1 90◦ 1 90◦ 1

II I II I II I

sin(α)
α
sin β

β
180◦ 0◦ 180◦ α 0◦ 180◦ 0◦
−1 1 −1 cos(α) 1 −1 cos β 1

III IV III IV III IV

270◦ −1 270◦ −1 270◦ −1

Similarly, for example, we can read angles in fourth quadrant as negative values of the equivalent
angles in the first quadrant. After rotation of 180◦ , see Fig. 5.7, we can deduce the following

cos(π ± α) = − cos ±α
sin(π − α) = − sin(−α) = sin α

Fig. 5.7 Example 5.1-2

90 ◦ 1 90 ◦ 1
sin( π − α)

II I II I
sin

−α cos( ±α) 0 ◦ −α 0◦
−α −α
sin( −α)

cos( π ± α) 1 −1 1

III IV III IV
270 ◦ −1 270 ◦ −1
Solutions 79

In conclusion, knowing “key” angles in the first quadrant, unit circle is used to visually calculate
their correspond angles in the other three quadrants.

3. Knowing that 180◦ = π rad, we use a simple proportion equation as follows.

300◦ α 300
◦
5 5π
(a) = ∴ α=π   =
180 ◦ π 180◦ 3 3
330◦ α 330
◦
11 11π
(b) = ∴ α=π 
 ◦ =
180 ◦ π 180 6 6
−18◦ α  
−18 ◦
−1 π
(c) = ∴ α=π 
 ◦ =−
180◦ π 180 10 10
4. Knowing that 180◦ = π rad, we use a simple proportion equation as follows.
α 3π/4 3
π /4A H ◦ ◦ ◦
(a) = ∴ α= H 45 = 135
180
180◦ π π
α 5π/12 π /Z
5 ZH ◦ ◦
12 ◦
(b) = ∴ α= H 15 = 75
180
180◦ π 
π
α 2 2 360◦
(c) ◦
= ∴ α = 180◦ =
180 π π π
5. With the help of unit circle, after accounting for the positive negative angle directions, we
calculate
√
π sin(π/3) 3/2
√
(a) tan = = = 3
3 cos(π/3) 1/2

 
7π 6
π π π π  1
(b) sin = sin + = sin 6π + = − sin =−
6 6
6 6 6 2
 π  √
3π π 2
(c) cos − = cos ± π = − cos = −
4 4 4 2
6. By definition of sin x and cos x functions, assuming a, b to be the two catheti and c hypotenuse,
we write,

adjacent a
cos θ = = ∴ a = 10 cos θ
hypotenuse 10
opposite b
sin α = = ∴ b = 10 sin θ
hypotenuse 10

7. Knowing that hypotenuse of a right-angled triangle equals the unit circle radius, i.e. equals one,
then by Pythagoras’s theorem, assuming a, b to be the two catheti and c hypotenuse, we write,

a = cos α, b = sin α
∴
a 2 + b2 = 12 ∴ cos2 α + sin2 α = 12

which is direct consequence of hypotenuse c = 1 and definitions for sin x and cos x functions.
Therefore we verify each case.
80 5 Trigonometry

  2
1 2 3 1 9 10
(a) sin2 α + cos2 α = − + = + = =1 ✗
5 5 25 25 25
 
4 2 1 2 16 1 17
(b) sin2 α + cos2 α = √ + √ = + = =1 
17 17 17 17 17
 √ 2  √ 2
15 5 15 5 20
(c) sin α + cos α = −
2 2
+ = + = =1 ✗
5 5 25 25 25

8. Knowing the relationship

cos2 α + sin2 α = 12

and given

1 3π
sin α = − et π < α <
3 2

we write
 2
1
cos2 α + sin2 α = 1 ∴ cos2 α = 1 − sin2 α = 1 − −
3
∴
√ √
1 8 2 2
cos α = 1− ∴ cos α = =
9 3 3
√
However, α is in the third quadrant, thus its cos x is negative, see Fig. 5.5, i.e. cos α = −2 2/3.
Therefore, we write
√
sin α −1/3
1 2
tan α = = √ = √ =
cos α −2 2/3
2 2 4

9. With the help of unit circle, we write,

√
◦ ◦ ◦ 3 ◦
(a) cos 150 = cos(180 − 30 ) = − cos 30 = −
2
√
◦ ◦ ◦ 3 ◦
(b) sin 120 = cos(90 + 30 ) = cos 30 =
2

sin(−60◦ ) − sin(60◦ )
tan 300◦ = tan(360◦ − 60◦ ) = tan(−60◦ ) = ◦
=
cos(−60 ) cos(60◦ )
(c) √
− 3/2
√
= =− 3
1/2

Solutions 81

(d) tan 225◦ = tan(180◦ + 45◦ ) = tan 45◦ = 1

tan 480◦ = tan(360◦ + 120◦ ) = tan 120◦ = tan(180◦ − 60◦ )

√
(e) sin(180◦ − 60◦ ) sin 60◦ 3/2
√
= = = = − 3
cos(180◦ − 60◦ ) − cos 60◦ −1/2

10. With the help of unit circle, we calculate
√
(a) tan 840◦ = tan(2 × 360◦ + 120◦ ) = tan 120◦ = − 3

cos 1320◦ = cos(3 × 360◦ + 240◦ ) = cos 240◦ = cos(180◦ + 60◦ )

(b) 1
= − cos 60◦ = −
2
 √ 2
π π 1 3 13 3
(c) sin sin2 = = =
6 3 2 2 24 8
√ √
π π π π 1 3 1 3 3
(d) cos cos + sin cos = √ +√ =
4 6 4 6 2 2 2 2 2
 √ 2
π 1 3 1 3 1 9−4 5
(e) sin2 − π = − √ 2 = − = =
3 tan2 2 4 3 12 12
3 3
√
π π π π 1 1 3√ 1 3
(f) sin cos + cos tan = √ √ + 3= + =2
4 4 6 3 2 2 2 2 2

Exercise 5.2, page 70

1. Main idea is to decompose given angles into the basic angles found in the unity circle, whose
sin and cos (therefore, also tan) values are already known. In addition, understanding of relations
among similar angles, for example, 180◦ ± α and 90◦ ± α is very important, (Fig. 5.8).
82 5 Trigonometry

Fig. 5.8 Example 5.2-1

90 ◦ 1 90 ◦ 1
II I II I

sin( α)
α
−α cos( ±α) 0 ◦ 180 ◦ α 0◦
cos( π ± α) −α 1 cos(90 ◦ + α) 1

III IV III IV
270 ◦ −1 270 ◦ −1

90 ◦ 1 90 ◦ 1

sin( π − α)
II I II I

sin(90 ◦ + α)
α
−α 0◦ α 0◦
−α
sin( −α)

−1 1 cos( α) 1

III IV III IV
270 ◦ −1 270 ◦ −1

(a) Important to know values of sin and cos are found on the unity circle,
π  1
cos =
3 2
(b) Function tan is found by sin and cos ratio
√
π  sin(π/4) 2/2
tan = = √ =1
4 cos(π/4) 2/2

(c) Difference of 180 and 30 gives

√
◦ ◦ ◦ ◦ 3
cos(150 ) = cos(180 − 30 ) = − cos(30 ) = −
2
(d) Sum of 90 and 30 gives
√
3
sin(120◦ ) = sin(90◦ + 30◦ ) = cos(30◦ ) =
2
(e) Functions sin and cos of negative angles result in
sin(−45◦ ) − sin(45◦ ) sin(45◦ )
tan(−45◦ ) = ◦
= ◦
=− = − tan(45◦ ) = −1
cos(−45 ) cos(45 ) cos(45◦ )

(f) Functions sin and cos of negative angles result in

√
◦ sin(−60◦ ) − sin(60◦ ) sin(60◦ ) 3/2 √
tan(−60 ) = = = − = − = − 3
cos(−60◦ ) cos(60◦ ) cos(60◦ ) 1/2
Solutions 83

(g) Difference of 360 and 60 gives

sin(−60◦ ) − sin(60◦ )
tan(300◦ ) = tan(360◦ − 60◦ ) = tan(−60◦ ) = =
cos(−60◦ ) cos(60◦ )
√
3/2 √
=− =− 3
1/2

(h) Sum of 180 and 45 gives

sin(180◦ + 45◦ ) − sin(+45◦ )
tan(225◦ ) = tan(180◦ + 45◦ ) = ◦ ◦
=
cos(180 + 45 ) − cos(+45◦ )
= tan 45◦ = 1

(i) Difference of 360, 120, and 60 gives

tan(480◦ ) = tan(360◦ + 120◦ ) = tan(180◦ − 60◦ ) = tan(−60◦ )
√
= − tan(60◦ ) = − 3
(j) Multiple revolutions around the 2π circle result in
tan(840◦ ) = tan(2 × 360◦ + 120◦ ) = tan(180◦ − 60◦ ) = tan(−60◦ )
√
= − tan(60◦ ) = − 3
(k) Multiple revolutions around the 2π circle result in
cos(1320◦ ) = cos(3 × 360◦ + 240◦ ) = cos(180◦ + 60◦ ) = − cos(60◦ ) = −1/2

(l) More complicated problems may be also reduced as, for example,
 
π    1 √3 2 1 3 3
2 π
sin sin = = =
6 3 2 2 2 4 8

2. The idea is to use trigonometric identities for products of sin and cos functions,
(a) Product of two sin functions is identical to
   
7π 13π 1 7π 13π 7π 13π
−4 sin sin = −4 cos − − cos +
12 12 2 12 12 12 12
 
6π 20π
= −2 cos − − cos
12 12
!  "
π  0
* 5π
= −2 cos  − cos
 2 3
  π 
= −2 − cos − =1
3
84 5 Trigonometry

(b) Inter-product of sin and cos functions is identical to

 π   
5π 1 5π π 5π π
4 sin cos =4 sin + + sin −
8 8 2 8 8 8 8
 !√ "
1 3π π  1 2
=4 sin + sin =4 +1
2 4 2 2 2
√
= 2+2

(c) Product of cos functions results in

   
5π 3π 1 5π 3π 5π 3π
4 cos cos =4 cos + + cos −
8 8 2 8 8 8 8
! √ "
1  π  1 2
=4 cos (π ) + cos =4 −1 +
2 4 2 2
√
= 2−2

(d) This product is found first, the products of numerators, then the product of denominators,
sin 20◦ sin 40◦ sin 80◦
tan 20◦ tan 40◦ tan 80◦ =
cos 20◦ cos 40◦ cos 80◦
∴
1
sin 20◦ sin 40◦ sin 80◦ = [cos(20◦ − 40◦ ) − cos(20◦ + 40◦ )] sin 80◦
2
◦
1 : cos 20

=  ◦ )
cos(−20 − cos(60◦ ) sin 80◦
2
1 1 
: 1/2
= cos60
(sin 100◦ + sin 60◦ ) −  ◦
sin 80◦
2 2
 √ 
11 
 :
 3/2 
= sin ◦
100 +  sin60 ◦
−sin80 ◦
22
√
3
=
8
here, we note that sin 100◦ = sin(90◦ + 10◦ ) and sin(80◦ ) sin(90◦ − 10◦ ), in other words,
these two values are equal thus cancel. Similarly, we find that cos 100◦ = − cos 80◦ , and the
product of denominators is calculated as
Solutions 85

∴
1
cos 20◦ cos 40◦ cos 80◦ = [cos(20◦ + 40◦ ) + cos(20◦ − 40◦ )] cos 80◦
2
◦
1 
: 1/2 : cos 20

= cos60 ◦
+ ◦ )
cos(−20 cos 80◦
2
1 1
= cos 80◦ + cos 20◦ cos 80◦
2 2
⎡ ⎤
1⎢
⎢1
 1  1 
: 1/2⎥
= ⎣ cos80◦ + cos100◦ + 
cos60 ◦ ⎥
⎦
2 
2    2   2
cos(90−10)◦ cos(90+10)◦

1
=
8
Therefore, we conclude that√
3/8 √
tan 20◦ tan 40◦ tan 80◦ = = 3
1/8

Exercise 5.3, page 70

1. With the help of trigonometric functions for angle sum identities, we find
(a) Angle of 15◦ can be seen as (45◦ − 30◦ ), so that

sin 15◦ = sin(45◦ − 30◦ ) = sin 45◦ cos 30◦ − cos 45◦ sin 30◦
√ √ √ √
2 3 2 1 2 √
= − = ( 3 − 1)
2 2 2 2 4
cos 15◦ = cos(45◦ − 30◦ ) = cos 45◦ cos 30◦ + sin 45◦ sin 30◦
√ √ √ √
2 3 2 1 2 √
= + = ( 3 + 1)
2 2 2 2 4
√
2 √
◦  ( 3 − 1) √
sin 15 3−1
tan 15◦ = ◦
= √4 √
cos 15 2 √ 3−1
 ( 3 + 1)
4
√ √ √ √
( 3 − 1)( 3 − 1) 3−2 3+1 2
(2 − 3) √
= √ √ = = =2− 3
( 3 + 1)( 3 − 1) 3−1 2

86 5 Trigonometry

(b) Angle of 75◦ can be seen as (45◦ + 30◦ ), so that

sin 75◦ = sin(45◦ + 30◦ ) = sin 45◦ cos 30◦ + cos 45◦ sin 30◦
√ √ √ √
2 3 2 1 2 √
= + = ( 3 + 1)
2 2 2 2 4
cos 75◦ = cos(45◦ + 30◦ ) = cos 45◦ cos 30◦ − sin 45◦ sin 30◦
√ √ √ √
2 3 2 1 2 √
= − = ( 3 − 1)
2 2 2 2 4
√
2 √
 ( 3 + 1) √
sin 75◦ 3+1
◦
tan 75 = = √4 √
cos 75◦ 2 √ 3+1
 ( 3 − 1)
4
√ √ √ √
( 3 + 1)( 3 + 1) 3+2 3+1 2
(2 + 3) √
= √ √ = = =2+ 3
( 3 + 1)( 3 − 1) 3−1 2

(c) Angle of 105◦ can be seen as (60◦ + 45◦ ), so that

sin 105◦ = sin(60◦ + 45◦ ) = sin 60◦ cos 45◦ + cos 60◦ sin 45◦
√ √ √ √
3 2 1 2 2 √
= + = ( 3 + 1)
2 2 2 2 4
cos 105◦ = cos(60◦ + 45◦ ) = cos 60◦ cos 45◦ − sin 60◦ sin 45◦
√ √ √ √
1 2 3 2 2 √
= − = (1 − 3)
2 2 2 2 4
√
2 √ √
◦  ( 3 + 1)
sin 105 1 + 3
◦
tan 105 = = √4 √
cos 105◦ 2 √ 1 + 3
 (1 − 3)
4
√ √ √ √
( 3 + 1)( 3 + 1) 3+2 3+1 2
(2 + 3) √
= √ √ = = = −(2 + 3)
(1 − 3)(1 + 3) 1−3 
−2 −1

Solutions 87

(d) Similarly, angle of π/12 can be seen as (π/3 − π/4), so that

π π π π π π π
sin = sin − = sin cos − cos sin
12 3 4 3 4 3 4
√ √ √ √
3 2 1 2 2 √
= − = ( 3 − 1)
2 2 2 2 4
π π π π π π π
cos = cos − = cos cos + sin sin
12 3 4 3 4 3 4
√ √ √ √
1 2 3 2 2 √
= + = ( 3 + 1)
2 2 2 2 4
√
π 2 √
sin  ( 3 − 1) √
π 12 = 4 3−1
tan = π √ √
12 cos 2 √ 3−1
12  ( 3 + 1)
4
√ √
( 3 − 1)( 3 − 1) √
= √ √ = ··· = 2 − 3
( 3 + 1)( 3 − 1)

π 
(e) Angle of 5π/12 can be seen as 4
+ π
6
, so that
 π √
5π π 2 √
sin = sin + = ··· = ( 3 + 1)
12 4 6 4
 π √
5π π 2 √
cos = cos + = ··· = ( 3 − 1)
12 4 6 4
 √

5π 2 √
sin  ( 3 + 1) √
5π 12
tan =  = √4 = ··· = 2 + 3
12 5π 2 √
cos  ( 3 − 1)
12 4

π 
(f) Angle of 7π/12 can be seen as 3
+ π
4
, so that
 π √
7π π 2 √
sin = sin + = ··· = ( 3 + 1)
12 3 4 4
 π √
7π π 2 √
cos = cos + = ··· = (1 − 3)
12 3 4 4

7π
 sin
7π 12 √
tan =  = · · · = −(2 + 3)
12 7π
cos
12
88 5 Trigonometry

2. Knowing the sum identities, we write

1
(a) sin 20◦ cos 10◦ + cos 20◦ sin 10◦ = sin(20◦ + 10◦ ) = sin 30◦ =
2
√
◦ ◦ ◦ ◦ ◦ ◦ 3 ◦
(b) cos 43 cos 13 + sin 43 sin 13 = cos(43 − 13 ) = cos 30 =
2
3. Knowing Pythagoras’s theorem, unit circle and the sum identities, we write
⎧
 ⎫ ⎪ 3
cos x = 4/5 ∴ sin x = 1 − (4/5) = 3/5
2 ⎬ ⎪
⎨sin x =− (IV)
5
 ∴
sin y = −3/5 ∴ cos y = 1 − (−3/5)2 = 4/5⎭ ⎪
⎪
⎩cos y 4
= − (III)
5

   
3 4 4 3 12 12
sin (x + y) = sin x cos y + cos x sin y = − − + − = − =0
5 5 5 5 25 25
   
3 4 4 3 12 12 24
sin (x − y) = sin x cos y − cos x sin y = − − − − = + =
5 5 5 5 25 25 25

4. With the help of sum identities we write

√ 
π 
√ π  2
2 cos α − cos +α
2 cos α − 2 cos +α 2 4
π  √ 4 =  
π  √2
2 sin + α − 2 sin α
4 2 sin +α − sin α
4 2
 π  π  π 
2A coscos α −coscos α + sin sin α
=  4 4 4
π π  π 
2A sin cos α + cossin α − cossin α
4  4  4
π
sinsin α
= 4 = tan α
π
sin cos α
4

5. After setting a = b in the sum identities we have

sin (a + a) = sin 2a = sin a cos a + cos a sin a = 2 sin a cos a

cos (a + a) = cos 2a = cos a cos a − sin a sin a = cos2 a − sin2 a
sin 2a sin a cos a + cos a sin a sin a cos a + cos a sin a
tan 2a = = = 
cos 2a cos a cos a − sin a sin a sin a sin a
cos a cos a 1 −
cos a cos a
sin a  
cosa  
cosa sin a
 + 
 
= cos a 
cos a cos a cos a
1 − tan2 a
Solutions 89

2 tan a
=
1 − tan2 a

Further, we can write

⎧
 ⎪ 1 − cos 2a
cos2 a + sin2 a =1 ⎪ 2
⎨2 sin a = 1 − cos 2a ∴ sin2 a =
2
∴
cos2 a − sin2 a = cos 2a ⎪
⎪ 1 + cos 2a
⎩2 cos2 a = 1 + cos 2a ∴ cos2 a =
2

In addition,

cos a sin a cos2 a cos2 a

sin 2a = 2 sin a cos a =2 = 2 tan a 2
cos a cos a 1 cos a + sin2 a
1
= 2 tan a
cos2 a sin2 a
+
cos2 a cos2 a
2 tan a
=
1 + tan2 a

  sin2 a

cos2
a 1 −
cos2 a − sin a
2
cos2 a − sin a
2
cos2 a
cos 2a = = = 
1 cos a + sin a
2 2
2 sin2 a
cos
 a 1+
cos2 a
1 − tan2 a
=
1 + tan2 a

These identities are used to integrate rational functions by trigonometric substitutions.

Exercise 5.4, page 71

1. Given the sum identities

sin(α + β) = sin α cos β + cos α sin β (5.1)

sin(α − β) = sin α cos β − cos α sin β (5.2)
cos(α + β) = cos α cos β − sin α sin β (5.3)
cos(α − β) = cos α cos β + sin α sin β (5.4)
90 5 Trigonometry

(a) Difference between (5.4) and (5.3) gives

cos(α(
cos(α − β) − cos(α + β) = ( ((
cos cos(α(
β + sin α sin β − ( ((
cos β + sin α sin β
= 2 sin α sin β
∴
1
sin α sin β = (cos(α − β) − cos(α + β)) (5.5)
2

(b) Sum of (5.3) and (5.4) gives

 
cos (α + β) + cos (α − β) = cos α cos β −  α
sin sinβ + cos α cos β +  α
sin sinβ
= 2 cos α cos β
∴
1
cos α cos β = (cos (α + β) + cos(α − β)) (5.6)
2

(c) Sum of (5.1) and (5.2) gives

 
cosα
sin (α + β) + sin (α − β) = sin α cos β +  sin cosα
β + sin α cos β −  sin β
= 2 sin α cos β
∴
1
sin α cos β = (sin(α + β) + sin(α − β)) (5.7)
2

Identities (5.5)–(5.7) are extensively used in RF electronics and communication theory.

Solutions 91

2. Given the sum identities (5.5)–(5.7)

(a) From (5.7) we write

1
sin α cos β = (sin(α + β) + sin(α − β))
2
∴
sin(α + β) + sin(α − β) = 2 sin α cos β
⎧ ⎧ ⎫
⎪  ⎪ x+y ⎪
⎪
⎨α + β = x ⎪
⎨α = ⎪
⎬
2
∴
⎪ ⎪ x−y ⎪
⎩α − β
⎪ =y ⎪
⎩β = ⎪
⎭
2
∴
 
x+y x−y
sin x + sin y = 2 sin cos
2 2

(b) Similarly, from (5.6) we find

 
x+y x−y
cos x + cos y = 2 cos cos
2 2

Using the same technique, other similar identities are easily derived.

3. Using the product to sum identities and the unit circle, we find
(a)
 
5π 7π 1 5π 7π 5π 7π
−4 sin cos = −4
2 sin + + sin −
12 12 2
12 12 12 12
 


12π 2
π
= −2 sin + sin −

12  6
12
  π  
1
= −2 0 − sin = −2 −
6 2
=1
92 5 Trigonometry

(b)

1  π  π 1   π π
+ 2 sin 1 + cos 1 + = + sin 1 + + 1 +
2 6 3 2 6 3
 π π 
+ sin 1 + − 1 −
6 3
1   π   π 
= + sin 2 + + sin −
2 2 6
1   π   π 
= + sin + 2 − sin
2 2 6
1 1
= + cos (2) −
2 2
= cos(2)

(c)

8 sin 20◦ sin 40◦ sin 80◦ = 4 [cos(20◦ − 40◦ ) − cos(20◦ + 40◦ )] sin 80◦
= 4 [sin 80◦ cos(−20◦ ) − sin 80◦ cos 60◦ ]
sin 80◦
= 4 sin 80◦ cos(20◦ ) −
2
sin 100◦ sin 60◦ sin 80◦
=4 + −
2 2 2
 
sin 100◦ = sin(90◦ + 10◦ ) = sin(90◦ − 10◦ ) = sin 80◦
! √ "
 ◦  ◦

sin 80 3 
sin 80
= 4A  + − 
2 4A 2
√
= 3
Solutions 93

(d)

8 cos 10◦ cos 50◦ cos 70◦ = 4 [cos 60◦ + cos 40◦ ] cos 70◦
= 4 [cos 60◦ cos 70◦ + cos 40◦ cos 70◦ ]
cos 70◦ cos 110◦ cos 30◦
=4 + +
2 2 2
 
cos 110◦ = cos(90◦ + 20◦ ) = − cos(90◦ − 20◦ ) = − cos 70◦
! √ "


cos 70 ◦ 

cos 70 ◦
3
= 4A  −  +
 2  2 4A
√
= 3

4. With the help of identities and unit circle, we write

√ √ √ √ √ √ √ √
tan α = 6 + 3 − 2 − 2 = 3 2 + 3 − 2 − ( 2)2
√ √  √ √  √2 − 1 √  √ √ 
= 3 2+1 − 2 2+1 = √ 2+1 3− 2
2−1
 
(a − b)(a + b) = a − b 2 2

√ √
√ √ 3 2
3− 2 2 − ◦ ◦
= √ = 2√ 2 = sin 60 − sin 45
2−1 2 2 1 sin 45◦ − sin 30◦
−
2 2
   
x+y x−y
sin x − sin y = 2 cos sin
2 2
 
 ◦  15
 ◦ (180−75)◦  ◦
2
cos 105  2
sin cos 2 cos 90◦ − 752
=  ◦
2
 ◦ =  ◦ =  ◦
2
cos 752  sin152
cos 752 cos 752
 
cos(90◦ − α) = sin α
 ◦  ◦
sin 752 75
=  75◦  = tan
cos 2 2

∴
75◦
α= = 37.5◦
2
94 5 Trigonometry

Exercise 5.5, page 71

1. Equations that include periodic functions have periodic solutions.

(a) sin x = 0 ∴ x = nπ, where n = 0, ±1, ±2, . . .
π
(b) cos x = 0 ∴ x = (2k + 1) , where k = 0, ±1, ±2, . . .
2
2. The factorized form of equations helps us to exploit the property if A × B × C = 0, then either
A = 0 or B = 0 or C = 0. Thus, we convert sums into products as
(a) We convert the equation into its equivalent factorized form as follows.

sin x = sin α ∴ sin x − sin α = 0

∴
   
α+β α−β
sin α − sin β = 2 cos sin
2 2
 
x+α x−α
2 cos sin =0
2 2

Therefore,

x+α x+α π
cos =0 ∴ = (2n + 1) ∴ x = −α + (2n + 1)π
2 2 2

x−α x−α
sin =0 ∴ = kπ ∴ x = α + 2kπ
2 2

where (k, n) = 0, ±1, ±2, . . .

(b) We could also exploit the equivalence relationship between sin x and cos x functions (i.e.
their argument difference of π/2) as follows.
 π π
sin x = cos x ∴ cos x − = cos x ∴ x − = ±x + 2nπ
2 2

because x takes both positive and negative signs. Therefore, there are two possible solutions that
must be reviewed as follows.
π
case ‘ + x’ : ∴ x
− = x
+ 2nπ ∴ no solution
2
π π
case ‘ − x’ : ∴ x − = −x + 2nπ ∴ 2x = 2nπ +
2 2
∴
π
x = nπ +
4
Solutions 95

where n = 0, ±1, ±2, . . . The solutions are shown in the graph of sin x and cos x, Fig. 5.9.

Fig. 5.9 Example 5.5-2(b)

sin(x), cos(x)
0

−2S −S 0 S 2S
x

(c) Factorized form of this equation is solved as follows.

 
3x x
sin x = sin 2x ∴ sin 2x − sin x = 0 ∴ 2 cos sin =0
2 2
∴

3x 3x π π
cos =0 ∴ = (2n + 1) ∴ x = + 2nπ
2 2
2
3
x  x
sin =0 ∴ = kπ ∴ x = 2kπ
2 2

where n, k = 0, ±1, ±2, . . . The solutions are shown in the graph of sin x and sin 2x, Fig. 5.10,
where the solutions (i.e. the intersect points) are generated by two streams: for n = 0, ±1, ±2, . . .
and for k = 0, ±1, ±2, . . . . In the factorized form of the equation (i.e. the equality to zero), roots
are found at the intersect with the horizontal axis (i.e y = 0).

Fig. 5.10 Example 5.5-2(c)

sin(x), sin(2x)

−2S −S 0 S 2S
x
96 5 Trigonometry

(d) Using the same idea, we factorize this equation and solve as follows.

2 sin2 x + sin x = 0 ∴ (2 sin x + 1) sin x = 0

∴ (2 sin x + 1) = 0 or sin x = 0

We solve the two equations separately.

Equation: sin x = 0

sin x = 0 ∴ x = nπ where n = 0, ±1, ±2, . . .

Equation: (2 sin x + 1) = 0

1 π 
(2 sin x + 1) = 0 ∴ sin x + = 0 ∴ sin x + sin =0
2 6
∴
x + π/6 x − π/6
2 sin cos =0
2 2

Therefore, we search solutions for two equations.

1. Roots of sin function are found at multiples of π , i.e. 0, ±π, ±2π, . . .

x + π/6 x + π/6
sin =0 ∴ = kπ
2 2
∴
π
x=− + 2kπ where k = 0, ±1, ±2, . . .
6

2. Roots of cos function are at odd multiples of π/2, i.e. ±(2m + 1)π/2

x − π/6 x − π/6 π
cos =0 ∴ = (2m + 1)
2 2
2

∴
7π
x= + 2mπ where m = 0, ±1, ±2, . . .
6

These solutions are clearly visible in graph of f (x) = 2 sin2 x + sin x function, Fig. 5.11. One
stream of the solutions (i.e. the intersect points) is generated for n = 0, ±1, ±2, . . . , second for
k = 0, ±1, ±2, . . . , and third for m = 0, ±1, ±2, . . .
Solutions 97

Fig. 5.11 Example 5.5-2(d)

f (x)
0

0 S 2S 3S
x

(e) Knowing periodicity of tan x = sin x/ cos x function (it is periodic for nπ ), we can solve
this equation as follows

cos (x + π/6) = sin (x − π/3)

∴
π  π  π  π 
cos (x) cos − sin (x) sin = sin (x) cos − cos (x) sin
6 6 3 3
∴
√ √
3 1 1 3
cos (x) − sin (x) = sin (x) − cos (x)
2
2
2
2

∴
√
√ √ 3/2 sin(π/3)
2
3 cos x = 2
sin x ∴ tan x = 3 = =
1/2 cos(π/3)
∴
π
x= + nπ where n = 0, ±1, ±2, . . .
2
3. Trigonometric equations are specific because each equality periodically repeats both in positive
and negative directions. For that reason, given the interval of interest it is necessary to search all
solutions within the interval. In addition, we exploit factorized form of the equations by using
sum to product trigonometric identities,
 
α+β α−β
cos α − cos β = −2 sin sin
2 2
 
α+β α−β
sin α − sin β = 2 cos sin
2 2

The factorized form of equations helps us to exploit the property if A × B × C = 0, then either
A = 0 or B = 0 or C = 0.
98 5 Trigonometry

(a) In the interval 0 ≤ x < 2π we search all solution as follows.

π  π 
cos − x = cos
3 6
∴
π  π 
cos − x − cos =0
3 6
∴
⎛ππ ⎞ ⎛ π π ⎞
−x+ −x−
⎜ 6 ⎟ × sin ⎜ 3 6⎟=0
− 2 × sin ⎝ 3 ⎠ ⎝ ⎠
2 2
        
A B C

Obviously, −2 = 0 leaving either B = 0 or C = 0 terms. Both A and B are in the form of sin θ,
a sine function equals to zero for all arguments

sin θ = 0 ∴ θ = 0, ±π, ±2π, . . . in general: θ = nπ, n = 0, ±1, ±2, . . .

Therefore, we systematically write the cases as

n=0:
π
−x+ π
π π
B: 3 6
=0 ∴ − x = 0 ∴ x = ∈ [0, 2π ] 
2 2 2
π
−x− π
π π
C: 3 6
=0 ∴ − x = 0 ∴ x = ∈ [0, 2π ] 
2 6 6

n=1:
π
−x+ π π
−x 3π
B: 3 6
=π ∴ 2
=π ∴ x=− <0 ✗
2 2 2
π
−x− π π
−x 11π
C: 3 6
=π ∴ 6
=π ∴ x=− <0 ✗
2 2 6

n = −1 :
π
−x+ π π
−x 5π
B: 3 6
= −π ∴ 2
= −π ∴ x = > 2π ✗
2 2 2
π
−x− π π
−x 13π
C: 3 6
= −π ∴ 6
= −π ∴ x = > 2π ✗
2 2 6

In conclusion, in interval 0 < x ≤ 2π , there are two solutions, see Fig. 5.12 (left).
.π π /
x= ,
6 2
Solutions 99

(b) Similarly, within the interval −π < x ≤ π , we write,

 π
sin(3x) = sin x −
2
∴
 π 
sin(3x) − sin x − =0
2
∴
 
3x + x − π
3x − x + π
2 × cos 2
× sin 2
=0
2 2
      
A B C

Obviously, 2 = 0 leaving either B = 0 or C = 0 terms. Term C is a sine function, and B is in the

form of cos θ. A cosine function equals to zero for all arguments

π 3π π
cos θ = 0 ∴ θ = ± , ± , . . . in general: θ = ± nπ, n = 0, 1, 2, . . .
2 2 2

Therefore, we systematically write the cases as

n=0:

3x + x − π2 π π 3π
B: = ∴ 4x − = π ∴ x = ∈ [−π, π ] 
2 2 2 8
3x − x + π2 π π
C: = 0 ∴ 2x + = 0 ∴ x = − ∈ [−π, π ] 
2 2 4

n=1:

3x + x − π2 π π 7π
B: = + π ∴ 4x − = 3π ∴ x = ∈ [−π, π ] 
2 2 2 8
3x − x + π2 π 3π
C: = π ∴ 2x + = 2π ∴ x = ∈ [−π, π ] 
2 2 4

n = −1 :

3x + x − π2 π π π
B: = − π ∴ 4x − = −π ∴ x = − ∈ [−π, π ] 
2 2 2 8
3x − x + 2π
π 5π
C: = −π ∴ 2x + = −2π ∴ x = − < −π ✗
2 2 4

n=2:

3x + x − π2 π π 11π
B: = + 2π ∴ 4x − = 5π ∴ x = >π ✗
2 2 2 8
3x − x + 2
π
π 7π
C: = 2π ∴ 2x + = 4π ∴ x = >π ✗
2 2 4
100 5 Trigonometry

(b) (cont.)
n = −2 :

3x + x − π2 π π 5π
B: = − 2π ∴ 4x − = −3π ∴ x = − ∈ [−π, π ] 
2 2 2 8
3x − x + 2
π
π 9π
C: = −2π ∴ 2x + = −4π ∴ x = − < −π ✗
2 2 2

In conclusion, within the interval −π < x ≤ π , there are six solutions (see Fig. 5.12 (right)),
 
5π 2π π 3π 6π 7π
x = − ,− ,− , , ,
8 8 8 8 8 8

Fig. 5.12
Example 5.5-3(a)(b) 1 1
f (x)

f (x)
0 0

cos( S6 )  sin(3x) 
−1 cos S3 − x −1 sin x − S2
0 S S 3S 2S −S − S2 0 S S
2 2 2
x x

Exercise 5.6, page 71

1. Inequalities are solved by calculating the limits of intervals.

(a) We convert given equation and use unit circle as follows
√
√ 3
2 cos x + 3 < 0 ∴ cos x < −
2

√ that satisfy equality x1 = 5π/6 and

In other words, there are two angles, see Fig. 5.13 (right),
x2 = −5π/6. Marked region is where −1 ≤ cos x < − 3/2 (note that inequality cos x < −1
does not have real solution, i.e. amplitude of −1 ≤ sin x, cos x ≤ 1). Therefore, we conclude that
periodic solution is the interval, see Fig. 5.13,

5π 5π
− + 2nπ < x < + 2nπ where n = 0, ±1, ±2, . . .
6 6
Solutions 101

Fig. 5.13 Example 5.6-1(a) 1

cos√x II I
1 − 23
x1 5S
6

f (x)
0
− 5S
6
1
x2
−1

S
III IV
−S 0 −1
x

(b) Using the same technique as in (a) we write,

1 π 5π
2 sin x − 1 > 0 ∴ sin x > ∴ + 2nπ < x < + 2nπ
2 6 6
∴
where n = 0, ±1, ±2, . . .

as illustrated in Fig. 5.14.

Fig. 5.14 Example 5.6-1(b) 1

1 II I
x2 x1
f (x)

0
−1 1
sin x
1
−1 2
S
III IV
−S 0 −1
x

(c) The idea is to convert given inequality that contains multiple trigonometric functions into its
equivalent form where all trigonometric functions are combined into a single either sin(f (x)) or
cos(f (x)). For example, we can enforce the sum of two arguments form as
102 5 Trigonometry

% √
%
% 2
cos x − sin x < 1 %×
% 2
√ √ √
2 2 2
cos x − sin x <
2 2 2
√
π π 2
sin cos x − cos sin x <
4 4 2
∴
π √ 
2
sin −x <
4 2
√
From unit circle we know that sin x = 2/2 ⇒ x = π/4 or x = −5π/4, see Fig. 5.15. Therefore
we conclude that
π π
−x < ∴ x > 0 + 2nπ, n = 0, ±1, ±2, . . .
4 4
and
π 5π 3π
−x >− ∴ x< + 2nπ, n = 0, ±1, ±2, . . .
4 4 2

Fig. 5.15 Example 5.6-1(c) 1

2 II x2 x1 I
1
f (x)

0
−1 1
−1
cos x − sin x
1
−2
S
III IV
−S 0 2S −1
x

√
(d) Knowing that periodicity of sin 3x is 2π/3, and from unit circle we find that if sin x = 3/2
then x = π/3 or x = 2π/3. Thus, we solve given inequality as follows.
√ √
3 3
sin 3x − ≥ 0 ∴ sin 3x ≥
2 2
∴
π π 2nπ
3x ≥ ∴ x≥ + , n = 0, ±1, ±2, . . .
3 9 3
Solutions 103

or,
2π 2π 2nπ
3x ≤ ∴ x≤ + , n = 0, ±1, ±2, . . .
3 9 3

(e) Inequalities that include absolute functions are equivalent to two inequalities, thus define
and interval.

1 1 1
| sin x| ≥ ∴ − ≥ sin x ≥
2 2 2

By inspection of Fig. 5.16 we write the solutions as

5π π
− + 2nπ ≤ x ≤ − + 2nπ
6 6

and,

π 5π
+ 2nπ ≤ x ≤ + 2nπ
6 6

where n = 0, ±1, ±2, . . .

Fig. 5.16 Example 5.6-1(e) 1

II I

x2 x1

−1 1
x4 x3
III IV
−1

(f) We work out absolute values by separating them into two cases.

Case 1: sin x > 0 ∴ | sin x| = sin x, thus we find the interval’s boundary as

  π
sin
 sin
x = x + 2 cos x ∴ 2 cos x = 0 ∴ x = + nπ
2
104 5 Trigonometry

However, the initial condition is that sin x > 0, which is valid only for

π
x= + 2nπ, where n = 0, ±1, ±2, . . .
2

Case 2: sin x < 0 ∴ | sin x| = − sin x, thus we find the interval’s boundary as

− sin x = sin x + 2 cos x ∴ 2 sin x + 2 cos x = 0 ∴ sin x + cos x = 0

We solve this equation using same technique as in (c),

% √
%
% 2
sin x + cos x = 0 %×
% 2
√ √
2 2
sin x + cos x = 0
2 2
π π
cos sin x + sin cos x = 0
4 4
∴
 π π π
sin x + = 0 ∴ x + = 0 ∴ x = − + nπ
4 4 4

However, the initial condition is that sin x < 0, which is valid only for

π
x=− + 2nπ, where n = 0, ±1, ±2, . . .
4

The two streams of periodic solutions are illustrated in Fig. 5.17. We conclude that x found in the
periodic interval

π π
− + 2nπ ≤ x ≤ + 2nπ
4 2

satisfy given inequality.

Fig. 5.17 Example 5.6-1(f)

2

1
f (x)

−1
| sin x|
−2 2 cos x + sin x
0 S 2S
x
 Complex Algebra
6

Important to Know

z = a + ib ∴ z∗ = a − ib ∴ |z|2 = z z∗ = a 2 + b2
b
z = |z|eiθ = |z| (cos θ + i sin θ ) ∴ θ = arctan
a
z1 = a + ib and z2 = c + id ∴ if z1 = z2 ⇒ a = c and b = d

Geometrical interpretation of the equivalence between Pythagorean triangle, complex numbers, and
vector addition, see Fig. 6.1.

Fig. 6.1 Pythagorean

y y j
triangle, vector addition, and c2 = a2 + b2 z = a + jb
complex plane

b b
c
φ b a + φ b Im(z) |z| φ
.

a x a x a x
Re(z)

6.1 Exercises

6.1 * Complex Numbers

1. Calculate |z| and (z),

(a) z=j (b) z = j2 (c) z = j3
(d) z = j4 (e) z = j5 (f) z = j 2020

© Springer Nature Switzerland AG 2021 105

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_6
106 6 Complex Algebra

2. Calculate |z| and (z) in each case and show the following complex numbers z as vectors in the
complex plane.
√ √ √
(a) z = 1 (b) z = 3 1 2 2
+j (c) z = +j
2 2 2 2
√ √ √
1 3 z=j 2 2
(d) z= +j (e) (f) z=− +j
2 2 2 2
√ √ √
3 1 z = −1 2 2
(g) z=− +j (h) (i) z=− −j
2 2 2 2
√ √ √
z = −j 2 2 3 1
(j) (k) z= −j (l) z= −j
2 2 2 2
3. Show the following complex numbers as vectors in the complex plane. Then calculate |z| and
(z) in each case.
π π π
(a) z = ej 0 (b) z = ej 6 (c) z = ej 4 (d) z = ej 3

π 3π 5π
(e) z = ej 2 (f) z = ej 4 (g) z = ej 6 (h) z = ej π
5π 3π 7π 11π
(i) z = ej 4 (j) z = ej 2 (k) z = ej 4 (l) z = ej 6

4. Calculate (z) and |z| of the following complex numbers.

π 1 jπ
z = −2 e−j
5π
(a) z = 2 ej 4 (b) z= e 3 (c) 6
2
1 7π √ π
2  π 3
(d) z = − ej 4 (e) z= 2 ej 4 (f) z = 3 ej 3
2
 5  4  π 3
z = 4 e−j 6 z = −3 e−j 6
5π 7π
(g) (h) z = 0.2 ej 4 (i)

6.2 * Basic Calculations

1. Given that z = i, calculate the following complex numbers and show each result in the complex
plane:
(a) z1 = z 2 (b) z2 = z 3 (c) z3 = z 4 (d) z4 = z 5
√ √ √ √
(e) z5 = z (f) z6 = 3 z (g) z7 = 4 z (h) z8 = 5 z

2. Given that z = 2i, calculate the following complex numbers and show each result in the complex
plane:
(a) z1 = z 2 (b) z2 = z 3 (c) z3 = z 4 (d) z4 = z 5
6.1 Exercises 107

3. Calculate the following:

(a) i2 + i3 + i4 (b) i −5 + i −17 + i 36
(c) (1 + i)4 + (1 − i)4 (d) i 5 + i −4 + i 121

4. Calculate the following:

(a) (3 + 4i)(3 − 4i) (b) i 125 + (−i)60 + i 83

1 1+i 4
(c) (2i)2 + (−2i)−4 (d) + √
i 2
   √ 3000  √ 3000
1000 1000
1+i 1−i 1+i 3 1−i 3
(e) √ + √ (f) +
2 2 2 2

5. Calculate the following:

(a) (1 − i)100 (b) (1 + i)50 (c) (2 − i)6
√ √ 
1+i 31−i 3 (1 + i)1000 4 12
(d) (e) (f) √
2 2 (1 − i)500 i 3−1
√
i 102 + i 101 (i 3 − 1)2 −41 + 63i 6i + 1
(g) (h) (i) −
i 100 − i 99 (2i)2 50 1 − 7i
6. Express z in polar form:
(a) z=i (b) z=2−i
(1 + i)200 (6 + 2i) − (1 − i)198 (3 − i)
(c) z=
(1 + i)196 (23 − 7i) + (1 − i)194 (10 + 2i)

6.3 * Complex Equations

1. Solve the following equations in C:

1 1 1
(a) (2 + 3i)x + (3 + 2i)y = 1 (b) = +
x + iy 2+i −2 + 4i
2. Solve the following equations for z in C:
(a) z2 = i (b) z2 = −i (c) z2 = 5 + 12i
√
3. Solve the equation z4 + 8 + 8 3 i = 0 for z in C.
108 6 Complex Algebra

Solutions

Exercise 6.1, page 105

1. We note that powers of a complex number in effect force z to rotate around the circle, therefore
there is periodicity in the results. Calculate |z| and (z),
(a) z = j therefore, |z| = 1, and z = π/2
(b) z = j = −1 therefore, |z| = 1, and
2
z=π
(c) z = j = j × j = −j therefore, |z| = 1, and
3 2
z = 3π/2
(d) z = j = j × j = 1 therefore, |z| = 1, and
4 2 2
z = 2π = 0
(e) z = j = j × j = j therefore, |z| = 1, and
5 4
z = π/2
 505
z = j 2020 = j 4×505 = j 4 = 1505 = 1
(f)
therefore, |z| = 1, and z = 2π = 0

Here, complex number modules are found simply by Pythagoras’s theorem, i.e. |z| =
2. 
(z)2 + (z)2 .
√
(a) Given: z = 1, we write (b) Given: z = 3/2 + j 1/2, we write
 √ 
(z) = 1 > 0 (z) = 3/2 > 0 > 0
(I) (I)
(z) = 0 (z) = 1/2 > 0
then (Fig. 6.2), module is then (Fig. 6.3), module is

|z| = 12 + 02 = 1 √ 2
|z| = 3/2 + (1/2)2 = 1

and phase is
(z) 1/2
(z) 0 z = arctan = arctan √ 
z = arctan = arctan (z) 3/2

(z) 1
◦ π
= arctan(0) = 0◦ = 30 =
6

j j
II I II I
π
0◦ 6
−1 1 −1 1

III −j IV III −j IV

Fig. 6.2 Example 6.1-2(a) Fig. 6.3 Example 6.1-2(b)

Solutions 109

√ √ √
(c) Given: z = 2/2 + j 2/2, we write (d) Given: z = 1/2 + j 3/2, we write
√  
(z) = 2/2 > 0 (z) = 1/2 > 0
√ (I) √ (I)
(z) = 2/2 > 0 (z) = 3/2 > 0
then (Fig. 6.4), module is then (Fig. 6.5), module is

√ 2 √ 2 √ 2
|z| = 2/2 + 2/2 =1 |z| = (1/2)2 + 3/2 =1

√ √
(z) 2 (z)
3
z = arctan = arctan 
√
2
z = arctan = arctan 21
(z) 2 (z)
2 2
π √ π
= arctan(1) = 45◦ = = arctan( 3) = 60◦ =
4 3

j j
II I II I
π π
4 3

−1 1 −1 1

III −j IV III −j IV

Fig. 6.4 Example 6.1-2(c) Fig. 6.5 Example 6.1-2(d)

110 6 Complex Algebra

√ √
(e) Given: z = j , we write (f) Given: z = − 2/2 + j 2/2, we write
 √ 
(z) = 0 (z) = − 2/2 < 0
(I) √ (II)
(z) = 1 > 0 (z) = 2/2 > 0
then (Fig. 6.6), module is then (Fig. 6.7), module is

 √ 2 √ 2
|z| = (0)2 + (1)2 = 1 |x| = − 2/2 + 2/2 =1
√
(z) 2
z = arctan
(z)
= arctan
1 z = arctan = arctan 2√
(z) 0 (z) −22
π 
= arctan(∞) = 90◦ = = arctan
1
= 135◦ =
3π
2 −1 4

j j
II I II I
π 3π
2 4
−1 1 −1 1

III −j IV III −j IV

Fig. 6.6 Example 6.1-2(e) Fig. 6.7 Example 6.1-2(f)

Solutions 111

√
(g) Given: z = − 3/2 + j 1/2, we write (h) Given: z = −1, we write
√  
(z) = − 3/2 < 0 (z) = −1 < 0
(II) (II)
(z) = 1/2 > 0 (z) = 0
then (Fig. 6.8), module is then (Fig. 6.9), module is

 √ 2
|x| = (1/2)2 + − 3/2 = 1 |x| = (−1)2 + (0)2 = 1

(z)
1 (z) 0
z = arctan = arctan 2√ z = arctan = arctan
(z) (z) −1
− 3
2
= 180◦ = π
√ 5π
= arctan (1/− 3) = 150◦ =
6

j
j II I
II I
5π π
6
−1 −π 1
−1 1

III −j IV
III −j IV
Fig. 6.9 Example 6.1-2(h)
Fig. 6.8 Example 6.1-2(g)
112 6 Complex Algebra

√ √
(i) Given: z = − 2/2 − j 2/2, we write (j) Given: z = −j , we write
√  
(z) = − 2/2 < 0 (z) = 0
√ (III) (III)
(z) = − 2/2 < 0 (z) = −1 < 0
then (Fig. 6.10), module is then (Fig. 6.11), module is

 √ 2  √ 2
|x| = − 2/2 + − 2/2 = 1 |x| = (0)2 + (−1)2 = 1
√
(z) − 2
/2 (z) −1
z = arctan = arctan 
√
z = arctan
(z)
= arctan
0
(z) 
−/2
2
 3π π
−1 = 270◦ = −90◦ = =−
= arctan = 225◦ = −135◦ 2 2
−1
5π 3π
= =−
4 4
j
II I
j 3π
II I 2
5π
4
−1 − π2 1

−1 1
− 3π
4
III −j IV

III −j IV
Fig. 6.11 Example 6.1-2(j)

Fig. 6.10 Example 6.1-2(i)

Solutions 113

√ √ √
(k) Given: z = 2/2 − j 2/2, we write (l) Given: z = 3/2 − j 1/2, we write
√  √ 
(z) = 2/2 > 0 (z) = 3/2 > 0
√ (IV) (IV)
(z) = − 2/2 < 0 (z) = −1/2 < 0
then (Fig. 6.12), module is then (Fig. 6.13), module is

√ 2  √ 2 √ 2
|x| = 2/2 + − 2/2 = 1 |x| = 3/2 + (−1/2)2 = 1

√
(z) − 2 (z) −1/2
z = arctan = arctan √
/2 z = arctan = arctan √ 
(z) 2 (z) 3/2

/2
 −1
−1 = arctan √ = 330◦ = −30◦
= arctan = 315◦ = −45◦ 3
1
11π π
=
7π
=−
π = =−
4 4 6 6

j j
II I II I
11π
7π 6
4
−1 −1 − π6
− π4
III IV III −j IV
−j

Fig. 6.13 Example 6.1-2(l)

Fig. 6.12 Example 6.1-2(k)

3. Complex numbers in this example given in exponential form are identical to those in Example 6.2-
2. Thus, by exploiting Euler’s equation, phase and module of a complex number are found by
inspection.
Aside from Pythagorean theorem used to calculate |z| it is also important to know how to calculate
|z| in the exponential form,

|z|2 = z z∗ = ej 6 e−j −j
π π π π
6 = ej 6 6

= e0 = 1 ∴ |z| = 1

which illustrates that module of any complex number found on the unity circle consequently
equals one. That is to say, complex exponent by itself keeps only the phase, while numbers whose
module |z| = A = 1 are written in the form
% %
z = A ej φ ∴ |z| = %A ej φ %
114 6 Complex Algebra

% 1
% j φ
%e *
= |A|  %

where A is any real number.

π
(a) z = ej 0 , therefore, φ = 0rad, its vector (b) z = ej 6 therefore, φ = π/6, its vector
graph is in Example 6.2-2(a): graph is in Example 6.2-2(b):
√
(z) = cos 0 = 1 (z) = cos π/6 = 3/2

(z) = sin 0 = 0 (z) = sin π/6 = 1/2

 
|z| = (z)2 + (z)2 = 1 |z| = (z)2 + (z)2 = 1

π π
(c) z = ej 4 therefore, φ = π/4, its vector (d) z = ej 3 therefore, φ = π/3, its vector
graph is in Example 6.2-2(a): graph is in Example 6.2-2(d):
√
(z) = cos π/4 = 2/2 (z) = cos π/3 = 1/2
√ √
(z) = sin π/4 = 2/2 (z) = sin π/3 = 3/2

|z|2 = z z∗ = ej 4 e−j |z|2 = z z∗ = ej ( 3 − 3 ) = e0 = 1

π π π π
4 = e0 = 1

π 3π
(e) z = ej 2 therefore, φ = π/2, its vector (f) z = ej 4 therefore, φ = 3π/4, its vector
graph is in Example 6.2-2(e): graph is in Example 6.2-2(f):
√
(z) = cos π/2 = 0 (z) = cos 3π/4 = − 2/2
√
(z) = sin π/3 = 1 (z) = sin 3π/4 = 2/2

|z|2 = z z∗ = ej ( 2 − 2 ) = e0 = 1
π π
|z|2 = z z∗ = ej ( 4 − 4 ) = e0 = 1
3π 3π

5π
(g) z = ej 6 therefore, φ = 5π/6, its vector (h) z = ej π therefore, φ = π , its vector
graph is in Example 6.2-2(g): graph is in Example 6.2-2(h):
√
(z) = cos 5π/6 = − 3/2 (z) = cos π = −1
(z) = sin 5π/6 = 1/2 (z) = sin π = 0

|z|2 = z z∗ = ej ( 6 − 6 ) = e0 = 1 |z|2 = z z∗ = ej (π −π ) = e0 = 1
5π 5π
Solutions 115

5π 3π
(i) z = ej 4 therefore, φ = 5π/4 = −3π/4, its (j) z = ej 2 therefore, φ = 3π/2 = −π/2, its
vector graph is in Example 6.2-2(i): vector graph is in Example 6.2-2(j):
√
(z) = cos 5π/4 = − 2/2 (z) = cos 3π/2 = 0
√
(z) = sin 5π/4 = − 2/2 (z) = sin 3π/2 = −1

|z|2 = z z∗ = ej ( 4 − 4 ) = e0 = 1 |z|2 = z z∗ = ej ( 2 − 2 ) = e0 = 1
5π 5π 3π 3π

7π 11π
(k) z = ej 4 therefore, φ = 7π/4 = −π/4, its (l) z = ej 6 therefore, φ = 11π/6 = −π/6, its
vector graph is in Example 6.2-2(k): vector graph is in Example 6.2-2(l):
√ √
(z) = cos 7π/4 = 2/2 (z) = cos 11π/6 = 3/2
√
(z) = sin 7π/4 = − 2/2 (z) = sin 11π/6 = −1/2

|z|2 = z z∗ = ej ( 4 − 4 ) = e0 = 1 |z|2 = z z∗ = ej ( − 11π

7π 7π 11π
6 6 ) = e0 = 1

4. By exploiting Euler’s exponential form of a complex number, phase and module are found as:
π π
(a) z = 2 ej 4 therefore, φ = π/4 and: (b) z = 1/2 ej 3 therefore,
φ = π/3 and |z| = 1/2
|z|2 = z z∗ = 2ej 4 2e−j
π π
4 = 4e0
|z| = 2

(c) z = −2 e−j 6 therefore,

5π 7π
(d) z = −1/2 ej 4 therefore,
φ = −5π/6 and |z| = 2 φ = 7π/4 = −π/4 and |z| = 1/2

In the following examples, we note that powers of a complex number, in effect, aside from
increasing its module also rotate the associated vector by multiples of the original phase.
(e) First, we must calculate the exponential (f) First, we must calculate the exponential
form of z as, form of z as,
√ 2  π 3
z=
π
2 ej 4 z = 3 ej 3
 √  2  π 2 −1
= 2 ej 4 ej
= 9 >
π
= 9 (−1) = −9

= 2 ej 4 ×2 = 2 ej 2
π π
therefore, φ = π and |z| = 9. We note that this
complex number rotated horizontal position,
therefore, φ = π/2 and |z| = 2. thus into a real number.
116 6 Complex Algebra

(g) First, we must calculate the exponential (h) First, we must calculate the exponential
form of z as, form of z as,
 5π
5 25π
 7π
4
z = 4 ej 6 = 45 e j 6 z = 0.2 ej 4 = 0.24 ej 7π

= 5−4 ej π = 1.6 × 10−3 (−1)

π π
= 210 ej 6 = 1024 ej 6
= −1.6 × 10−3
therefore, after making four turns around the
unity circle φ = π/6, and |z| = 1024.
therefore, after making two turns around the
unity circle φ = π the complex number rotated
into its position, thus into a real number, and
|z| = 1.6 × 10−3 (in engineering units).

(i) First, we must calculate the exponential

form of z as,
 π 3
z = −3 e−j 6 = (−3)3 e−j 2
π

= −27 (−j ) = 27j

therefore, φ = −π/2 and |z| = 27.

Exercise 6.2, page 106

1. We use the equivalent forms or roots and fractional powers in the exponential form of a complex
number. Therefore, z = i = ei π/2 = cos π/2 + i sin π/2.

 π 2
(a) z2 = i 2 = ei 2 = eiπ = cos π + i  sinπ
: 0 = −1

⎛ ⎞
0
 π 3 3π ⎜ 
>

3π 3π ⎟
(b) z3 = i 3 = e i 2 = e i 2 = ⎝ cos + i sin ⎠ = −i
 2 2

 π 4  2
(c) z4 = i 4 = ei 2 = ei 2π = i 2 = (−1)2 = 1
−1 1
(d)
z5 = i 5 = i × i 4 = i ∴ i 6 =  i
7
2
i
×7
4
= −1, etc.
 √ √
√ √  π 1 π π π 2 2
(e) z = i = ei 2 2 = ei 4 = cos + i sin = +i
4 4 2 2
 √
√ √  i π  13 i π6 π π 3 1
(f) z= i= e = e = cos + i sin = +i
3 3
2
6 6 2 2
√ √  π  1 π
 π π 
z = i = ei 2 4 = ei 8 = cos + i sin
4 4
(g)
8 8
Solutions 117

√ √  π 1 π
 π π
z = i = ei 2 5 = ei 10 = cos + i sin
5 5
(h)
10 10

Fig. 6.14 Example 6.2-1

j z
j z
z
z2 z4 3
z
4
−1 1 z
5
z
1
− j z3

We note that powers of a complex number whose |z| = 1 effectively rotate its vector, Fig. 6.14.
2. Radical and fractional power forms are equivalent, thus the exponential form of a complex number
results in z = 2i = 2ei π/2 = 2 (cos π/2 + i sin π/2).

 π 2
(a) z2 = (2i)2 = 2ei 2 = 4eiπ = 4 cos π + i  π
sin: 0 = −4
⎛ ⎞
0
 π 3 3π ⎜ 
3π>
 3π ⎟
(b) z3 = (2i)3 = 2ei 2 = 8ei 2 = 8 ⎝ cos + i sin ⎠ = −8i
 2 2
 π 4  2
(c) z4 = (2i)4 = 2ei 2 = 16ei 2π = 16 i 2 = 16(−1)2 = 16
(d) z5 = (2i)5 = 25 × i × i 4 = 32 i

We note that powers of a complex number whose |z| = 1 effectively rotate its vector and increase
its module, thus create spiral function, Fig. 6.15.

Fig. 6.15 Example 6.2-2(d)

‚(z)
32
z5

z2 z z4 ƒ(z)
0
z3

−4 0 16
118 6 Complex Algebra

3. By definition of the imaginary number i 2 = −1 we find

 2
(a) i 2 + i 3 + i 4 = i 2 + i × i 2 + i 2 = i 2 + i × (−1) + (−1)2
= −1 − i + 1 = −i
−5 −17 1 1 1 1  4 9
(b) i + i + i 36 = 5 + 17 + i 36 = +  4 4 +  i
i i i×i 4
i× i
2 i 2
= +1= + 1 = 1 − 2i
i # i i$ # $2
2
(c) (1 + i)4 + (1 − i)4 = (1 + i)2 + (1 − i)2
= (1
+ 2i − 1
)2 + (1
− 2i − 1
)2 = 4i 2 + 4i 2
= −4 − 4 = −8
−4 1  4 30
(d) i +i
5
+i 121
=i×i4 + 4 + i × 
i = i + 1 + i = 1 + 2i

i
4. By definition we find
(a) (3 + 4i)(3 − 4i) = 32 − (4i)2 = 9 + 16 = 25
 31  15  20
(b) i 125 + (−i)60 + i 83 = i × i4 + (−1)60 i4 + i × i 2 × i4
= i + 1 − i = 1
−4 1 1 63
(c) (2i) + (−2i) = −4 +
2
= −4 + =−
(2i)4 16 i 4 16
 # $2
1 1+i 4 i 1 (1 + i)2 (2i)2 4
(d) + √ = + 2
= −i + = −i − = −1 − i
i 2 i i 2 4 4

(e) Large power in complex number indicate large number of its rotations
   
1 + i 1000 1 − i 1000 1 + i 2×500 1 − i 2×500
√ + √ = √ + √
2 2 2 2
 500  500
2
i 2
i  125
= + − = 2 i4 =2
2
2

(f) Often, symmetrical forms of complex numbers may be simplified

 √ 3000  √ 3000
1+i 3 1−i 3
+
2 2
 √ √ 1000  √ √ 1000
(1 + i 3)2 (1 + i 3) (1 − i 3)2 (1 − i 3)
= +
23 8
 √ √ 1000  √ √ 1000
(−2 + 2i 3)(1 + i 3) (−2 − 2i 3)(1 − i 3)
= +
8 8
 √ √ 1000  √ √ 1000
−2
(1 − i 3)(1 + i 3) −2
(1 + i 3)(1 − i 3)
= +
8
8

 
1 − (−3) 1000 1 − (−3) 1000
= − + −
4 4
= (−1)1000 + (−1)1000 = 1 + 1 = 2
Solutions 119

5. By exploiting square and fourth power of i we find

(a) (1 − i)100 = (1 − i)2×50 = (−2i)50 = (2i)50 = 250 × i 2 ×   = −250
i 4×12
(b) (1 + i)50 = (1 + i)2×25 = (2i)25 = 225 × i ×   = 225 × i
i 4×6
(c) (2 − i)6 = (2 − i)2×3 = (4 − 4i − 1)3 = (3 − 4i)2 × (3 − 4i)
= (9 − 24i − 16)(3 − 4i) = (−7 − 24i)(3 − 4i)
= −21 + 28i − 72i − 96 = −117 − 44i
√ √ √
1+i 31−i 3 1 − (i 3)2 1+3
(d) = = =1
2 2 4 4

(1 + i)1000 (1 + i)2×500 (2i)500 2500 ×  

i 4×125
(e) = = = 
(1 − i)500 (1 − i)2×250 (−2i)250 2250 × i 2 ×  
i 4×62
500−250
2
= = −2250
i2
 12  12 412
4 4 412 212
(f) √ = − √ = √   = √
i 3−1 1−i 3 (1 − i 3)12 2 2 12 (1−i 3)12
 2 12

412
22×6 412−6
=  √ 3×4 = = 4 = 4096
6
1−i 3 (−1)4
2

i +i
102 101
i ×i
2 4×25
+ i × i 4×25 i2 + i
(g) = =
i 100 − i 99  i
4×25 − i 3 × i 4×24 1 − i3 
= z z∗ = (a + ib)(a − ib) = a 2 + b2 = |z|2
−1 + i 1 − i +i+i+1
−1
2i
= = = =i
1+i 1−i 1+1 2
Alternative solution:
−i −1
# 3 $
i 102 + i 101 i
99
i
7 + i
7
2
−1 − i −1 − i 2i
= = = =i
i −i
100 99
i [i − 1]
 99 −1 + i −1 − i 2
√ √ √ √
(i 3 − 1)2 1 − 2i 3 − 3 −2 − 2i 3 1 3
(h) = = = +i
(2i) 2 −4 −4 2 2
−41 + 63i 6i + 1 41 63 6i + 1 1 + 7i
(i) − =− + i+
50 1 − 7i 50 50 1 − 7i 1 + 7i

41 63 6i − 42 + 1 + 7i
=− + i−
50 50 1 + 49
  − 13i
−41 + 63i + 
41 63 − 13
= = i=i
50 50
120 6 Complex Algebra

6. Express z in polar form:

π π
(a) z = i ∴ |z| = 1, arg(z) = ⇒ z = ei 2
2
(b) By definition,
 √
z = 2 − i ∴ |z| = (22 + 1) = 5
(z) 1
tan(arg(z)) = =−
(z) 2

1 ◦
∴ arg(z) = arctan − ≈ −26.56
2
∴
√ −j ×26.56◦
z= 5e
(1 + i)200 (6 + 2i) − (1 − i)198 (3 − i)
(c) z=
(1 + i)196 (23 − 7i) + (1 − i)194 (10 + 2i)
(1 + i)2×100 (6 + 2i) − (1 − i)2×99 (3 − i)
=
(1 + i)2×98 (23 − 7i) + (1 − i)2×97 (10 + 2i)
(2i)100 (6 + 2i) − (−2i)99 (3 − i) 299 × 4(3 + i) − 299 i(3 − i)
= =
(2i)98 (23 − 7i) + (−2i)97 (10 + 2i) −298 (23 − 7i) − 298 i(5 + i)
2
99
(12 + 4i − 3i − 1) 2
(11 + i) 11 + i (11 + i)2
= = =
2
−98
(23 − 7i + 5i − 1) 2
(−11 + i) 11 + i (i − 11)(11 + i)
121 + 22i − 1 120 + 22i 60 + 11i 60 11
= = =− =− − i
−1 − 121 −122 61 61 61
(c) (cont.)
√
|11 + i| 122
|z| = =√ = 1,
| − 1||11 − i| 122
(z) − 11
tan(arg(z)) = =  61
(z) − 60

61
−11
arg(z) = arctan ≈ −170◦ (in III quadrant)
−60
∴
◦
z = e−170

Exercise 6.3, page 107

1. The necessary condition that two numbers are equal is that, respectively, their real parts are equal
and their imaginary parts are equal.
Solutions 121

(a) We explicitly separate real and complex parts of complex equations, then recall that
imaginary part of a real number, by definition, equals zero:

(2 + 3i)x + (3 + 2i)y = 1
2x + 3ix + 3y + 2iy = 1
(2x + 3y) + (3x + 2y)i = 1 + 0i
∴
2x + 3y = 1
3x + 2y = 0

This system of equation is solved, for example, by Cramer’s rule as

% % ⎫
%2 3%
Δ = %% %% = 4 − 9 = −5 = 0⎪
⎪
⎪
32 ⎪
⎪
⎪
⎪
% % ⎪
⎬
%1 3% Δx 2 Δy 3
Δx % %
=% %=2−0=2 ∴ x= = − and y = =
02 ⎪
⎪ Δ 5 Δ 5
% % ⎪
⎪
%2 1% ⎪
⎪
% % ⎪
⎪
Δy = % % = 0 − 3 = −3 ⎭
30

(b) This equation may be solved by enforcing same form on both sides of the equation, as

1 1 1 −2
+ 4i + 2
+ i 5i
= + = =
x + iy 2+i −2 + 4i (2 + i)(−2 + 4i) −4 + 8i − 2i − 4
5i 1 1
= = =
−8 + 6i 6i − 5i i
−8 6 8 i
5i
+ 5
5i

∴
1 1
= (6.1)
x+iy 6 8
+i
5 5

By inspection of (6.1) we write

6 8
x= and y =
5 5

2. Solutions of complex equations are symmetrically distributed in the complex plane.

(a) We exploit the equivalence between radicals and fractional powers in the exponential form
of a complex number. We note that the two solutions of quadratic equation, z1 and z2 , are located
symmetrically in the complex plane, Fig. 6.16.
122 6 Complex Algebra

 √
z2 = i ∴ z2 = i
1  π 1 π
|z| = i 2 = ei 2 2 = ei 4
∴
π
z1,2 = ±ei 4
 π π
= ± cos + i sin
4 4
√ √ 
2 2
=± +i
2 2

Fig. 6.16 Example 6.2-2(a)

90◦ 1
z1
p
4
−1 1
z2
270◦ −1

(b) Similarly, using the same technique, we find two solutions of quadratic equation z1 and z2
that are located symmetrically in the complex plane, Fig. 6.17.
 √
z2 = −i ∴ z2 = −i
∴
 π 1
|z| = (−i) 2 = e−i 2 2 = e−i 4
1 π

∴
z1,2 = ±e−i 4
π

 π π
= ± cos − i sin
4 4
√ √ 
2 2
=± −i
2 2
Solutions 123

Fig. 6.17 Example 6.2-2(b)

90◦ 1
z2

0◦
−1 − p4
z1
270◦ −1

(c) Two solutions to

z2 = 5 + 12i (6.2)

are complex numbers, thus their general algebraic form must be as

z = a + bi, (a, b) ∈ R
∴
z2 = a 2 + 2abi − b2 = (a 2 − b2 ) + (2ab) i (6.3)

By the equivalence of real and complex parts in (6.2) and (6.3), we write two identities

6
2ab = 12 ∴ ab = 6 ⇒ a =
b
 2
6
a 2 − b2 = 5 ∴ − b2 = 5
b

Where the last bi-quadratic equation per b is solved as follows, Fig. 6.18,

36
− b2 = 5 ∴ 36 − b4 = 5b2
b2
b4 +5b2 − 36 = 0
b4 +9b2 − 4b2 − 36 = 0
b2 (b2 + 9) − 4(b2 + 9) = 0
(b2 − 4)(b2 + 9) = 0
(b − 2)(b + 2)(b2 + 9) = 0
(6.4)
124 6 Complex Algebra

∴
b1,2 = ±2, ∈ R,  
b3,4 ∈
C (6.5)
6
b1,2 = ±2 ∴ a = = ±3
±2
∴
z1,2 = ±(3 + 2i) (6.6)

Fig. 6.18 Example 6.2-2(c) √

j 13
II I
z1
33.7◦
√ √
− 13 13
z2
III √ IV
− j 13

We find the module and arguments of z1,2 in (6.6) as

 √ 2
|z| = 32 + 22 = 13 and arg(z) = arctan ≈ 33.69◦ (6.7)
3

√
3. One possible way of solving z4 + 8 + 8 3 i = 0 may be as follows.
√
z4 + 8 + 8 3 i = 0 (6.8)
∴
4 √
z= −8 − 8 3 i

Main idea in this kind of problems is that powers and roots of a complex number are much easier
to find if we convert its algebraic into exponential form.
Thus, first we calculate module of a temporary variable |t| and argument θ of the 4th root’s
argument t, then we write its exponential form as follows.
√
t = −8 − 8 3 i
∴
 √ √
|t| = 2 (x) + 2 (x) = (−8)2 + (−8 3)2 = 64 + 64 × 3 = 16
Solutions 125

√
(x) −8 3 √ √ π
tan θ = = = 3 ∴ θ = arctan 3 =
(x) −8 3

We note that θ = π/3 is found in the first quadrant; however, both real and imaginary parts
of t are negative, thus t is found in the third quadrant, see Fig. 6.19. Consequently, the correct
argument of t is actually

π 4π
θ =π+ =
3 3
Therefore, we write

√ 4π
t = −8 − 8 3 i = 16 exp i
3

Fig. 6.19 Example 6.2-3 j16

II I

p + p/3
60◦
−16 60◦ 16

III x IV
− j16

With this conversion, the original problem takes the following form.

*   1
√ 4π 4π  
z= 4t= 4
16 exp i = 2 exp i 4 = 2 exp i π (6.9)
3 3 3

The original equation (6.8) is 4th order polynomial, therefore there must be, in total, four values
of z that are valid solutions. Obviously, (6.9) is only the first of the four values. All solutions of a
complex equations are found on circle whose radius equals to z.
Arguments of the solutions are symmetrically distributed within 2π . That is to say, for example,
in the case of second order equation, the two solutions are separated by 2π/2 = π angle. In the
case of third order equation, the three solutions are separated by 2π/3 angle. Therefore, the four
solutions of fourth order equation are separated by 2π/4 = π/2 and are found as
π π
i +k
zk = 2 e 3 2
 π π π π 
= 2 cos +k + i sin +k ; k = 0, 1, 2, 3
3 2 3 2
126 6 Complex Algebra

In summary,
 π   π  √
k = 0 : z0 = 2 cos + i sin =1+i 3
3 3
 π π π π  √
k = 1 : z1 = 2 cos + + i sin + =− 3+i
3 2 3 2
 π  π  √
k = 2 : z2 = 2 cos + π + i sin + π = −1 − i 3
3 3
 
π 3π π 3π √
k = 3 : z3 = 2 cos + + i sin + = 3−i
3 2 3 2

We note that, indeed, all four modules are equal

√ 2
|zk | = 12 + 3 =2

and all arguments are uniformly distributed on the circle,1 see Fig. 6.20.

Fig. 6.20 Example 6.2-3 2j

z0
z1
5p
6 p
3
−2 4p 2
3 11p
6 z3
z2
− 2j

1 Most calculators do not have atan2 function that correctly takes into account signs of real and imaginary parts.
Solutions 127

Proof Each of the four solutions z0,1,2,3 should be verified against (6.8), for example,
√
z14 + 8 + 8 3 i = 0
⎛  ⎞4
5π
⎜
i √
⎝2 e 6 ⎟
⎠ + 8 + 8 3i = 0


10π
i √
16 e 3+ 8 + 8 3i = 0

6π 4π
i  0+ √
16 e  3 3 + 8 + 8 3i = 0
 
4π 4π √
16 cos + i sin + 8 + 8 3i = 0
3 3
 √ 
1 3 √
16 − − i + 8 + 8 3i = 0
2 2
√ √
8
−8
−  8
3 i + 8
+  3i = 0
0=0
 Linear Algebra
7

Important to Know

Column matrix:

⎡ ⎤
a1
⎢ .. ⎥
⎣ . ⎦
am

Raw matrix:

# $
a1 · · · am

The transpose of a row vector is a column vector

⎤ ⎡
x1
# $T ⎢ ⎥
⎢ x2 ⎥
x1 x2 . . . xm = ⎢ . ⎥
⎣ .. ⎦
xm

The transpose of a column vector is a row vector

⎡ ⎤T
x1
⎢ x2 ⎥ # $
⎢ ⎥
⎢ .. ⎥ = x1 x2 . . . xm
⎣ . ⎦
xm

Matrix Amn is written as

⎡ ⎤
a11 . . . a1,n
⎢ ⎥
Amn = ⎣ ... . . . ... ⎦
am1 . . . amn

© Springer Nature Switzerland AG 2021 129

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_7
130 7 Linear Algebra

The identity matrix In is written as

⎡ ⎤
1 0 0 ··· 0
⎢0 1 0 · · · 0⎥
⎢ ⎥
⎢ ⎥
In = ⎢0 0 1 · · · 0⎥
⎢. . . . .⎥
⎣ .. .. .. . . .. ⎦
0 0 0 ··· 1

Determinant of matrix A2 is calculated as

% %
%a b %%
|A2 | = %% = ad − bc
c d%

Cramer’s rule: Given a system of m linear equations

a11 x1 + a12 x2 + · · · + a1n xn = b1

a21 x1 + a22 x2 + · · · + a2n xn = b2
..
.
am1 x1 + am2 x2 + · · · + amn xn = bm

where x1 , x2 , . . . , xn are the n unknown variables, a11 , a12 , . . . , amn are the coefficients of the system,
and b1 , b2 , . . . , bm are the constant terms of the equations. Then,
% %
% a11 a12 . . . a1n %%
%
% . %
Δ = |Amn | = % ... ..
. . . . .. %% (Δ = 0)
%
%am1 am2 . . . amn %

After replacing the i-th column of |Amn | by the column vector bm , we write
% % % % % %
% b1 a12 . . . a1n %% % a11 b1 . . . a1n %% % a11 a12 . . . b1 %%
% % %
% .. % Δ = % .. .. % · · · Δ = % .. . %
Δx1 = % ... ..
. ... . % % x % .
..
. ... . % % x % .
..
. . . . .. %%
% 2
% n
%
%bm am2 . . . amn % %am1 bm . . . amn % %am1 am2 . . . bm %

and,

Δx1 Δx 2 Δxn
x1 = ; x2 = ; · · · ; xn =
Δ Δ Δ
7.1 Exercises 131

7.1 Exercises

7.1 Vector Definitions

1. In a simple graph, illustrate the geometrical interpretation of dimensions in 3D, 2D, 1D, and 0D
spaces, where “xD” refers to number of dimensions.
2. Show the “preferred” vector representations in physics, mathematics, and informatics.
#» #» #»
3. Given three basis vectors i , j , k , sketch these three vectors in a single graph. Comment on the
space defined by these basis vectors?
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
i 1 j 0 k 0
#» ⎣ x ⎦ ⎣ ⎦ #» ⎣ x ⎦ ⎣ ⎦ #» ⎣ x ⎦ ⎣ ⎦
i = iy = 0 j = jy = 1 k = ky = 0
iz 0 jz 0 kz 1

4. Given four points in 2D space: A = (−2, 2), B = (1, 4), C = (5, 6), D = (3, 1),
# » # »
(a) Sketch a graph with AB and CD vectors.
# » # »
(b) Write the matrix form of AB and CD vectors

5. Sketch a graph that shows the following vectors

#» 3 #» −4 #» −3
a = b = c =
3 3 −4

#», that is to say | p

6. Calculate magnitude of p #»| if

(a) p#» = (0, 1) #» = (3, 4)

(b) p
(c) #» = (−3, 1)
p #» = (2, 3, 6)
(d) p
# » # »
7. Given coordinates of points in space, calculate coordinates of AB and BA in matrix form, then
calculate its magnitude (i.e. module).
(a) A(4, 1), B(1, −3) (b) A(2, 3), B(−1, 4)
(c) A(−1, −3), B(4, 2) (d) A(1, −2), B(3, 2)
# » # »
8. Given coordinates of points in space, calculate coordinates of AB and BA in matrix form, then
calculate its magnitude (i.e. module).
(a) A(4, 1, 6), B(2, 4, −2)
# » # »
(b) OA = (2, 3, 4), OB = (3, 0, −1) (Note: O = (0, 0, 0)).

9. Given data, calculate coordinates of the following vectors in matrix form.

# » # »
#» = AB
(a) A(2, 3, 1), B(3, −1, 0), C = (−1, −2, 1), D = (−3, 3, −2), calculate m + CD,
#» # » # »
n = AB − CD.
#» #» #» #»
(b) #»
a = (2, −3), b = (3, −1), calculate #»
c = #»a + b , d = #»
a − b.
132 7 Linear Algebra

#» #» #» #»
(c) #»
a = (−1, 2), b = (3, 2), #» c = (−2, −3), calculate p#» = #»a + b, m = 2 #»
a − b , #»
n =
#» #» #» #» #» #»
3 a − b − c , r = 2 b − /2 c1

#» √ #»
10. Vectors #»
a and b create angle π/6, given that | #»
a | = 3 and | b | = 1, calculate angle θ between
#» = #» #» #»
p a + b et #»q = #»
a − b.

7.2 ** Analytical Geometry

1. Solve the following vector problems.

(a) Given three points in space A = (3, 5), B = (−1, 3), C = (1, −7), calculate the
# » # » # »
coordinates of point D, given that AD = 3AB − 3/2BC.
(b) Derive equation of a line that intersects point A = (2, 5) and is normal to vector #»
n = (1, 3).
(c) Given two points in space A = (5, 4), B = (8, 6) derive equation of a line that intersects
# »
the origin and is normal to AB.
(d) Given line equation 2x − 3y + 5 = 0 and point in space A = (7, 6), derive equation of a
line that intersects point A and is parallel to the given line.

2. Solve the following problems related to circle.

(a) Derive the equation of a circle whose centre is found at C = (4, 0) and its radius equals
r = 3.
(b) Given equation of a circle x 2 + y 2 − 10x + 4y + 7 = 0, calculate its radius.
(c) Given two points in space A = (1, 2), B = (−3, 4) derive equation of circle whose diameter
equals AB.
(d) Given coordinates of circle centre C = (2, 0) and its radius r = 5, derive equation of this
circle. Then, determine coordinates of the intersect points between this circle and line x −y −3 =
0.

3. Solve the following problems related to triangle.

(a) Given triangle whose sides are AB = 4, AC = 6, BC = 8 calculate cos α, where angle α is
associated with the point A.
(b) Given triangle as AB = 3, AC = 4, α = π/3 calculate side BC.

7.3 Vector Operations

1. Sketch a graph to illustrate the operation of addition in 1D and 2D spaces. As an example, show
the following vector additions:
(a) 2 + 3 = 5 and 2 − 3 = −1
#» #» #»
c = #»
a + b if #»
1 3
(b) a = and b =
2 −1

2. Given vector
7.1 Exercises 133

#» x
x = x =
3
(7.1)
xy 2

sketch diagram of #»
x multiplied by a constant.
#» #» 2 #»
a = 2 #» #» 3
(a) x (b) b = x (c) c = − #»
x
3 4

7.4 Linear Transformation

1. By a simple sketch show similarities between a function f (x) and a linear transformation L( #» v)
operation, then give a brief explanation of the relation between linear transformations and space.
#» #»
2. Given vector #»v = −2 i + j ,
(a) By sketching a diagram, show that the numerical form of #» v stays the same after an arbitrary
linear transformation L( #»
v ).
(b) With the help of the diagram, calculate the matrix form of L( #» v ) used in your example.

3. Show the geometrical interpretation of linear transformation L( #»

x ) given that
#» 31 #» −1
L( x ) = and x =
12 2

4. Find linear transformations to:

(a) rotate 2D space 90◦ counterclockwise,
(b) shear 2D space 45◦ clockwise,
(c) first rotate 90◦ counterclockwise then shear 45◦ clockwise 2D space.

5. Give geometrical interpretation of determinant for the following linear transformations:

L( #» L( #» L( #»
30 1 2 42
(a) x) = (b) x) = (c) x) =
02 1 −1 21

6. Given vectors,
#» #»
(a) #» #» = (9, 1) as linear combination of #»
a = (2, 1), b = (1, 0), write m a and b .
(b) #»
u = (3, −1), #» #» = (−1, 7), write m
v = (1, −2), w #» = #»
u + #»
v +w #» as linear combination of
#» #»
u and v .
#»
7. First show that #»
a , b , and #»
c linearly independent, then write #»
c as linear combination of #»
a and
#»
b if
#»
(a) #»a = (3, −2), b = (−2, 1), #» c = (7, −4)
#» #» #»
(b) a = (1, 2), b = (3, 4), c = (0, −2)

8. (a) Calculate parameter λ so that given vectors are linearly independent:

#»
(b) #»
a = (3, λ), b = (2, 6)
#»
(c) #»
a = (6, 8, 4), b = (3, 4, 2), #» c = (λ, 0, 1)
134 7 Linear Algebra

7.5 Determinants

1. Calculate determinants of the following second and third order matrix.

(a) (b)

30 1 2
A2 = A2 =
02 1 −1

(c) (d)
⎡ ⎤
42 2 34
A2 =
21 A3 = ⎣5 −2 1⎦
1 23

2. Calculate the volume of cuboid created by vectors

#»
(a) #»
a = (1, −3, 1), b = (2, 1, −3), #»
c = (1, 2, 1)
#»
(b) #»
a = (1, 0, 3), b = (0, 1, 2), #»
c = (3, 4, 0)

3. Calculate determinants
% % %√ √ % % % % %
%2 −1% % 3 −3 2% % 1 i% % sin α cos α %
(a) %% % (b) %√ √ % (c) %% % (d) %% %
1 0% % 3 2 2% −i 1% − cos α sin α %

4. Calculate determinants
% % % % % % % %
%3 2 1% %2 3 1% %5 0 4% % 1 3 5%
% % % % % % % %
(a) %%4 5 6%% (b) %%3 −1 2%% (c) %%8 0 −7%% (d) %% 7 9 11%%
%8 9 7% %1 1 −3% %3 2 1% %13 15 17%

7.6 Systems of Linear Equations

1. Solve the following system of linear equations for {x, y, z}:

5x − 5y − 15z = 40
4x − 2y − 6z = 19
3x − 6y − 17z = 41

2. Given coordinates of three points

A = (−2, 20); B = (1, 5); C = (3, 25)

determine quadratic function

7.1 Exercises 135

f (x) = ax 2 + bx + c

so that it can be used for curve fitting of these three data points, Fig. 7.1.

Fig. 7.1 Example 7.6-2

25
20

f (x)
5
f (x)
0
−2 0 1 3
x

3. Solve the following systems using the Cramer’s rule.

(a) 4x1 − 3x2 = 0 (b) 5x − 5y − 15z = 40
2x1 + 3x2 = 18 4x − 2y − 6z = 19
3x − 6y − 17z = 41

4. Solve the following system by using Cramer’s rule.

x+y =7
y+z=8
−x + 2z = 7

5. Solve the following systems of linear equations

(a) x + 2y = −5 (b) x + 2y = −8
3x − y = 13 2x − y = −1
(c) x + y + z = 36 (d) 2x + 3y − z = 5
2x − z = −17 x + y + 2z = 7
6x − 5z = 7 2x − y + z = 1

7.7 Matrix-Vector Product

1. Calculate the product and show its geometric interpretation

1 1 3 2 −1 −5 2 −3 3 1 2 −8
(a) (b) (c) (d)
2 −1 0 3 2 3 1 2 −2 2 −1 −1

2. Calculate the product and show its geometric interpretation

1 1 2 −1 2 −1 2 −3 2 −3 1 1
(a) (b) (c)
2 −1 3 2 3 2 1 2 1 2 2 −1
136 7 Linear Algebra

7.8 * Matrix Product

1. Calculate the product of AB if

⎡ ⎤
1 2
3 05
A= B = ⎣5 −1⎦
−2 −1 4
0 −6

7.9 * Eigenvalue and Eigenvector

1. Very important special case in linear algebra is when matrix-vector product equals zero, even
when neither matrix nor vectors equal zero, i.e.
#»
A · #»
x = 0 (7.2)

Verify if (7.2) is satisfied given

(a) Matrix A and x#», 1 (b) The same matrix A and x#»2 ,
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
−2 −9 −1 1 −2 −9 −1 13
A = ⎣ 2 −6 −5⎦ x#»1 = ⎣1⎦ A = ⎣ 2 −6 −5⎦ x#»2 = ⎣−4⎦
−4 −3 4 2 −4 −3 4 10

2. Verify if #»
v is eigenvector of matrix A, and determine eigenvalue λ.
(a) Matrix A and v#», 1 (b) The same matrix A and v#», 2

3 −3 3 −3
v#»1 = v#»2 =
3 2
A= A=
2 −4 1 2 −4 1

3. Calculate eigenvector(s) #»
v and eigenvalue(s) λ of matrix A.
⎡ ⎤
4 6 10
A = ⎣ 3 10 13⎦
−2 −6 −8

7.10 * Inverse Matrix

1. Calculate inverse A−1 of matrix A.

4 −2
A=
2 3
Solutions 137

7.11 * Matrix Powers

1. Calculate the following powers

20 20 20
(a) D 2 if D = (b) D 5 if D = (c) A5 if A =
02 05 −1 3

Solutions

Exercise 7.1, page 131

1. Geometrical interpretation of space is based on definition of point, which is assumed to have

size of zero. A loose visual interoperation would be a sphere whose radius equals zero, Fig. 7.2.
Appropriately, assuming nothing else exists, a point would represent “0-order space” annotated
as R 0 .

1. Line: Multiple points aligned next to each other (albeit, the distance between any two points
being zero) form line. Therefore, it takes infinitely points to create a line of any length L. If
nothing else exists, line represents one-dimensional space, annotated as R 1 . As an analogy, a
little ant walking along very long wire can move only along forward–reverse direction of the
wire, thus it has only one freedom of movement. Assuming one arbitrary point as the reference
for measuring relative distances, it takes only one number to describe distance from any other
point to the reference. In addition, sign “±” shows which side of the reference is measured,
forward (i.e. “+”) or reverse (i.e. “−”)).
2. Surface: Multiple lines put next to each other create surface, annotated as R 2 . Analogy would
be a sheet of paper whose thickness equals zero. In this case the little ant walking on the surface
has two freedoms of movement: left–right and forward–reverse. Assuming one arbitrary point
as the reference for measuring relative distances, it takes two numbers to describe position
of any other point on the surface. This pair of numbers is referred to as “coordinates”, by
convention coordinates are written as (x, y) pair, where each number represents distance from
the “origin” in one of two directions. In analogy, assuming lower left corner of a page to be the
reference point, we can describe position of each letter on the page by measuring horizontal
and vertical distances.
3. Volume: Multiple surfaces put on top of each other create volume, annotated as R 3 . Analogy
would be a book, where the stacked pages form the third dimension. In this R 3 space position
of any point is described by three numbers, (x, y, z), each measuring distance from the origin
in the three possible directions. In this space a bee has freedom to fly, that is to say to move in
up–down direction. Our physical universe is R 3 space, and our perceptions are limited to 3D
space dimensions. In mathematics, however, there is no limit on order of space.

An important observation is that each lower order space can be seen as one of possible projections
of the higher order space. That is to say, looked from “straight ahead” a line looks like a point.
Looked from “a side” a surface looks like a line. Looked from one direction, a cube (3D) looks
like a square (2D), or a sphere (3D) looks like a circle (2D). Mathematically, this reduction is
space is found when one or more equations in a system is not independent, i.e. they may be
written as the product of a simple factor and another equation. In linear algebra, a matrix may be
seen a compact form of writing a system of equations, where the zero value of its determinant
138 7 Linear Algebra

indicates that one or more equations is dependent, thus the real order of space is lower than the
number of equations in system.

Fig. 7.2 Example 7.1-1 R0 space R1 space

point line

d=0 L
R2 space R3 space
surface
L L D

W W

2. A vector signifies a variable that has two properties: magnitude and direction. That is to say,
a simple number is not vector because it has only magnitude but not direction. For example,
“speed” is not vector because it shows only the time rate at which an object is moving along the
path. A car can be speeding 100 km/h down the highway in both directions. On the other hand,
“velocity” shows both the rate of change and direction of an object’s movement.
Three commonly used vector representations are as follows. In physics, “vector” takes geomet-
rical interpretation in form of a “directed arrow” that indicates both magnitude (i.e. the arrow
length) and direction (i.e. from its tail to the head). In mathematics, assuming 3D space a vector
variable is annotated in algebraic form as

#» #» #» #»
a = ax i + ay j + az k

#» #» #»
where (ax , ay , az ) are vector magnitudes as measured along the three directions ( i , j , k ),
respectively (also referred to as the “unity vectors”).
In informatics, the preferred way of representing a vector is in one-column matrix form, as for
example
⎡ ⎤
3.14
#»
a = ⎣ −5⎦
2

where the column list contains coordinates of each direction relative to origin (0, 0, 0), in this
case 3D space. Of course, this matrix version is abstract and the list can be associated with any
physical or nonphysical variable.
3. As defined, the three unity vectors are used to define 3D space and are written in a form of 3 × 3
matrix as
⎡ ⎤
100
#» #» #»
( i , j , k ) = ⎣0 1 0⎦
001

#» #» #»
where the first column is for i , the second column is for j , and the third column is for k unity
vector, Fig. 7.3. This resulting matrix where all elements along the main diagonal equal one and
Solutions 139

all other matrix elements equal zero is known as “unity matrix” with properties similar to number
one.

z
Fig. 7.3 Example 7.1-3
Basis vectors
⎡ ⎤ 2
1
i = ⎣0⎦
0
⎡ ⎤ 1
0
j = ⎣1⎦
k
0 0
j y
⎡ ⎤
0 1 1 2
k = ⎣0⎦
1
x
2
i

4. By convention, vectors are oriented in direction from first to second coordinate.

(a) Given four points in 2D space: A = (−2, 2), B = (1, 4), C = (5, 6), D = (3, 1), we draw
two vectors as in Fig. 7.4.

Fig. 7.4 Example 7.1-4 y C(5, 6)

6

B(1, 4)
4

A(−2, 2)
2
1 D(3, 1)
x
0
-2 0 1 3 5

(b) The matrix form assumes that all vectors originate at (0, 0) point. Geometrically, that means
# »
vectors are translated so that their origin is at (0, 0) point, when their final position is as for RQ
in Fig. 7.5 (top), we write

# » 3
RQ =
2
140 7 Linear Algebra

Fig. 7.5 Example 7.1-4

y Q (xQ , yQ )
 
Q 3
2 +

R 2
(xR , yR )
R x
0
the origin 0 3

y C(5, 6)
6

B(1, 4)
4
A(−2, 2) 2 -5
3
1 D(3, 1) -2

0
x
-2 0 1 3 5

# » # »
Similarly, after being translated to the origin (see Fig. 7.5 (bottom)), for vectors AB and CD we
write

# » 3 # » −2
AB = and CD =
2 −5

It is important to note that, by definition, translated vector are still identical: they have same
magnitude and direction.
5. By convention, the vector matrix form lists coordinates in order, therefore the three vectors are as
in Fig. 7.6.

Fig. 7.6 Example 7.1-5    

−4 y 3
3 3
b a
3 3
x
-4 3

  c
−3 -4
−4 -3

6. Given px , py , and pz vector projections, we write,

#» = (0, 1) ∴ | p
(a) p #»| = √02 + 12 = 1
Solutions 141

#»| = √32 + 42 = 5
#» = (3, 4) ∴ | p
(b) p
#» = (−3, 1) ∴ | p
(c) p #»| = √32 + 12 = √10
#»| = √22 + 32 + 62 = 7
#» = (2, 3, 6) ∴ | p
(d) p
# » # »
7. Geometrical and matrix interoperation of AB and BA vectors are as follows.
# »
(a) Coordinates of AB we calculate as the end point minus the starting point,

# » 1 4 (1 − 4) −3
AB = − = =
−3 1 (−3 − 1) −4

# »
while coordinates of BA are

# » 4 1 (4 − 1) 3
BA = − = =
1 −3 (1 − (−3)) 4

# » # » # »
in both cases, by inspection of graph, as for AB in Fig. 7.7, we calculate |AB| and |BA| with the
help of Pythagoras’s theorem as
# » # »  
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 = 32 + 42 = 5

# »
(b) Similarly for AB we calculate

# » −1 2 (−1 − 2) −3
AB = − = =
4 3 (4 − 3) 1

# »
while coordinates of BA are

# » 2 −1 (2 − (−1)) 3
BA = − = =
3 4 (3 − 4) −1

# » # »
in both cases we calculate |AB| and |BA| with the help of Pythagoras’s theorem as
# » # »  
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 = 32 + 12 = 2
142 7 Linear Algebra

# »
(c) For AB we calculate

# » 4 −1 (4 − (−1)) 5
AB = − = =
2 −3 (2 − (−3)) 5

# »
while coordinates of BA are

# » −1 4 (−1 − 4) −5
BA = − = =
−3 2 (−3 − 2) −5

# » # »
in both cases we calculate |AB| and |BA| with the help of Pythagoras’s theorem as

# » # »   √
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 = 52 + 52 = 5 2

# »
(d) For AB we calculate

# » 3 1 (3 − 1) 2
AB = − = =
2 −2 (2 − (−2) 4

# »
while coordinates of BA are

# » 1 3 (1 − 3) −2
BA = − = =
−2 2 (−2 − 2) −4

# » # »
in both cases we calculate |AB| and |BA| with the help of Pythagoras’s theorem as
# » # »  
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 = 22 + 22 = 2

Fig. 7.7 Example 7.1-7(a) R2 espace

y A: (4,1)
1
j
0
0 i 1 4 x
4

B: (1,-3)

-3
Solutions 143

8. These vectors are placed in 3D space because there are three coordinates in each vector ax , ay , az .
# »
(a) Coordinates of AB we calculate as the end point minus the starting point,
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
2 4 (2 − 4) −2
# » ⎣ ⎦ ⎣ ⎦ ⎣
AB = 4 − 1 = (4 − 1) ⎦ = ⎣ 3⎦
−2 6 (−2 − 6) −8

# »
while coordinates of BA are
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
4 2 (4 − 2) 2
# » ⎣ ⎦ ⎣ ⎦ ⎣
BA = 1 − 4 = (1 − 4) ⎦ = ⎣−3⎦
6 −2 (6 − (−2)) 8

# »
in both cases, by inspection of graph, as for AB in Fig. 7.8, we find that vector is placed in
# » # »
diagonal of a cuboid, thus we calculate |AB| and |BA| with the help of Pythagoras’s theorem as
# » # »   √
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 + |zB − zA |2 = 22 + 32 + 82 = 77

(b) Relative to the point of origin, we write

⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
3 2 (3 − 2) 1
# » ⎣ ⎦ ⎣ ⎦ ⎣
AB = 0 − 3 = (0 − 3) ⎦ = ⎣−3⎦
−1 4 (−1 − 4) −5

# »
while coordinates of BA are
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
2 3 (2 − 3) −1
# »
BA = ⎣3⎦ − ⎣ 0⎦ = ⎣ (3 − 0) ⎦ = ⎣ 3⎦
4 −1 (4 − (−1)) 5

# »
in both cases, by inspection of graph, as for AB in Fig. 7.8, we find that vector is placed in
# » # »
diagonal of a cuboid, thus we calculate |AB| and |BA| with the help of Pythagoras’s theorem as
# » # »   √
|AB| = |BA| = |xB − xA |2 + |yB − yA |2 + |zB − zA |2 = 12 + 32 + 52 = 35
144 7 Linear Algebra

Fig. 7.8 Example 7.1-8(a)

R3 espace
z
k −j
x i
3 -2
y

-8

9. Given points and vectors, first we calculate the required vectors relative to the origin.
(a) A(2, 3, 1), B(3, −1, 0), C = (−1, −2, 1), D = (−3, 3, −2), therefore:
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
3 2 (3 − 2) 1
# » ⎣ ⎦ ⎣ ⎦ ⎣
AB = −1 − 3 = (−1 − 3)⎦ = ⎣−4⎦
0 1 (0 − 1) −1
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
−3 −1 (−3 − (−1)) −2
# »
CD = ⎣ 3⎦ − ⎣−2⎦ = ⎣ (3 − (−2)) ⎦ = ⎣ 5⎦
−2 1 (−2 − 1) −3

Then, we calculate
⎡
⎤ ⎡ ⎤ ⎡ ⎤
1 −2 −1
# » # » ⎣ ⎦ ⎣ ⎦ ⎣ ⎦
#» = AB
m + CD = −4 + 5 = 1
−1 −3 −4
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
1 −2 3
#» # » # »
n = AB − CD = ⎣−4⎦ − ⎣ 5⎦ = ⎣−9⎦
−1 −3 2
Solutions 145

#»
(b) In this case #»
a = (2, −3) and b = (3, −1) are already placed at the origin, thus

#» #»
c = #»
2 3 5
a +b = + =
−3 −1 −4

#» #» −1
c = #»
2 3
a −b = − =
−3 −1 −2

(c) Similarly,

#» = #» #» −1 3 2
p a +b = + =
2 2 4

#» = 2 #» #» −1 3 (2 × (−1) − 3) −5
m a − b =2 − = =
2 2 (2 × 2 − 2) 2

#» #» −1 −2 (3 × (−1) − 3 − (−2)) −4
n = 3 #»
a − b − #»
3
c =3 − − = =
2 2 −3 (3 × 2 − 2 − (−3)) 7

#» #» 1 1 −2 (2 × 3 − 1/2 × (−2))
r = 2 b − #»
3 7
c =2 − = =
2 2 2 −3 (2 × 2 − 1/2 × (−3)) 11/2

10. By inspection of sketch in Fig. 7.9 (left) we write

#»
a = a#»x + a#»y
#» #» #»
b = bx + by

where the horizontal and vertical projections are calculated as

√
#» #» π √ 3 3
|ax | = | a | cos = 3 =
6 2 2
√
π √ 1 3
|a#»y | = | #»
a | sin = 3 =
6 2 2

Solution 1: in matrix form we write

#» 3/2 #» 1
a = √ b =
3/2 0

then, we calculate

#» = #» #» 3/2 1 5/2
p a +b √ + = √
3/2 0 3/2

#» #»
q = #»
3/2 1 1/2
a −b √ − = √
3/2 0 3/2
146 7 Linear Algebra

#» and #»
Angle θ between p q is found by inspection of Fig. 7.9 (right). We write,

θ =α−β
√
3/2 √ √ π
tan α = = 3 ∴ α = arctan 3 = = 60◦
1/2 3
√ √ √
3/2 3 3
tan β = = ∴ α = arctan = 19.1◦
5/2 5 5
∴
θ = 40.89◦

Fig. 7.9 Example 7.1-10

y y = -
3
2
|a|= 3
q p
ay
6
|b|=1 x
0 0
0 1 x 0 1 5
ax 2 2

Solution 2: Alternatively, we can use for example Pythagoras’s theorem and write

#» = 5/2
#»| √ √
p √
3/2
∴ |p = (5/2)2 + ( 3/2)2 = 7

#» √
∴ | #»
1/2
q = √
3/2
q| = (1/2)2 + ( 3/2)2 = 1

Then we can use the law of cosines for triangle formed by p #» and #»
q , see Fig. 7.10, and find that
its third side is a = 2. Now all data is in place to calculate

a 2 = b2 + c2 − 2bc cos θ
∴
√ √ √ 2
22 = ( 7)2 + 12 − 2 7 cos θ ∴ 2 7θ = 4 ∴ cos θ = √
7
∴
2
θ = arccos √ = 40.89◦
7
Solutions 147

Fig. 7.10 Example 7.1-10

y
3 a=2
2
|q| = 1
|p| = 7

x
0
0 1 5
2 2

Exercise 7.2, page 132

1. Vector equations may be solved using the graphical representation.

# » # » # »
(a) We resolve vector equation AD = 3AB − 3/2BC by using, for example, the graphical
method. One possible order of calculations is as follows.
Each pair of points in cartesian space A = (3, 5), B = (−1, 3), C = (1, −7) defines one vectors.
# » # »
Vectors AB and BC, see Fig. 7.11, are translated from their initial positions to the origin 0, 0,
Fig. 7.11.

Fig. 7.11 Example 7.2-1(a)

We derive vector coordinates at the origin by separately calculating x and y vector projections.
That is
# »
AB : (xB − xA , yB − yA ) = (−1 − 3, 3 − 5) = (−4, −2)
# »
BC : (xB − xA , yB − yA ) = (1 − (−1), −7 − 3) = (2, −10)
148 7 Linear Algebra

Fig. 7.12 Example 7.2-1(a)

Multiplication of each vector by a constant gives

# » # » # »
3AB : (3AB x , 3AB y ) = (3 × (−4), 3 × (−2)) = (−12, −6)
 
3# » 3# » 3# » 3 3
− BC : − BC x , − BC y = − × 2, − × (−10) = (−3, 15)
2 2 2 2 2
# » # » # » # » # »
Vector AD = 3AB − 3/2BC as the sum of these two vectors, i.e. 3AB : (−12, −6) and −3/2BC :
(−3, 15), Fig. 7.12 (left), as
# »
AD : (−3 + (−12), 15 + (−6)) = (−15, 9)

# »
In order to find coordinates of point D, the origin of vector AD is translated to point A : (3, 5),
Fig. 7.12 (right), and therefore we calculate

D : (−15 + 3, 9 + 5) = (−12, 14)

(b) Two lines

y 1 = m1 x + b 1
y 2 = m2 x + b 2

are orthogonal (y1 ⊥ y2 ) if

1
m2 = −
m1
Solutions 149

Slope of vector #»
n = (1, 3) is found by definition (hint: tan x in a right angled triangle)

Δy 3
m1 = = =3
Δx 1
∴
1 1
m2 = − =−
m1 3

Thus, y2 = −1/3x + b2 , where we calculate b2 so that y2 crosses point A : (x, y) = (2, 5).
Therefore we write

1
y2 = − x + b2
3
∴
1 17
5 = − × 2 + b2 ∴ b2 =
3 3
∴
1 17
y2 = − x +
3 3

as illustrated in Fig. 7.13.

Fig. 7.13 Example 7.2-1(b)

# »
(c) Given two points in space A = (5, 4), B = (8, 6) slope m1 of vector AB is found as

Δy 2 1 3
m1 = = ∴ m2 = − =−
Δx 3 m1 2

We write line equation through point A : (0, 0), Fig. 7.14, where b2 = 0 as

3
y2 = − x
2
150 7 Linear Algebra

Fig. 7.14 Example 7.2-1(c)

(d) Two lines are parallel if their slopes are equal, i.e. m1 = m2 . Given line equation 2x − 3y +
5 = 0 and point in space A : (x, y) = (7, 6), Fig. 7.15, we derive equation of the parallel line as

2 5
2x − 3y + 5 = 0 ∴ y = x+
3 3
∴
2
y2 = x + b2
3
∴
2 4
6= × 7 + b2 ∴ b2 =
3 3
∴
2 4
y2 = x+
3 3

Fig. 7.15 Example 7.2-1(d)

2. (a) General equation of a circle is

(x − x0 )2 + (y − y0 )2 = r 2

where (x0 , y0 ) are coordinates of the centre. We recall that this equation is yet another
interpretation of Pythagoras’s theorem Fig. 7.16.
Solutions 151

Fig. 7.16 Example 7.2-2(a) y

(x, y)
y
r

y0
x
0 x0 x

Given C = (4, 0) and its radius equals r = 3, we derive the equation of a circle as follows

(x − 4)2 + (y − 0)2 = 32 ∴ (x − 4)2 + y 2 − 9 = 0

∴

y = ± 9 − (x − 4)2

(b) Given equation is extended to fit into “binomial square” form as

x 2 + y 2 − 10x + 4y + 7 = 0
x 2 − 2 × 5x +25 −25 + y 2 + 2 × 2y +4 −4 + 7 = 0
     
∴
(x − 5)2 + (y + 2)2 = 22
√
In conclusion, centre coordinates are (x0 , y0 ) = (5, −2) and radius equals r = 22.
152 7 Linear Algebra

(c) Given conditions, either point A or point B may serve as the circle centre, Fig. 7.17.
Choosing A : (x0 , y0 ) = (1, 2) while B = (−3, 4) we write
# »  √
r = |AB| = (1 − (−3))2 + (4 − 2)2 = 20
∴
(x − 1)2 + (y − 2)2 = 20

Fig. 7.17 Example 7.2-2(c) y

2
x
0

−3 0 1

(d) Given coordinates of circle centre C = (2, 0) and its radius r = 5 we write

(x − 2)2 + y 2 = 25 ∴ x 2 − 4x + 4 + y 2 = 25
∴

y1 = ± 21 + 4x − x 2

At the same time, the line equation is

x − y − 3 = 0 ∴ y2 = x − 3
Solutions 153

The intersect point are found at the coordinates when the two curves are equal, i.e. y2 = y1 ,
therefore we write the equality and square both sides as

 %2
%
x − 3 = ± 21 + 4x − x 2 %
%

∴
x 2 − 6x + 9 = 21 + 4x − x 2 ∴ 2x 2 − 10x − 12 = 0
∴
2x 2 + 2x − 12x − 12 = 0
2x(x + 1) − 12(x + 1) = 0
(x + 1)(2x − 12) = 0 ∴ 2(x + 1)(x − 6) = 0
∴
x1 = −1, x2 = 6

Therefore, y = x − 3 ∴ y1 = −4, y2 = 3, Fig. 7.18.

Fig. 7.18 Example 7.2-2(d)

y

3
x
0

−3

y2 = x − 3
0 23 6

3. Non-right angle triangles are solved by the sin/cos laws.

154 7 Linear Algebra

(a) Knowing “the cosine law” and “the sine law”, i.e.
⎫
c2 = a 2 + b2 − 2ab cos Ĉ ⎪⎪
⎪
⎬
a2 = b + c − 2bc cos Â
2 2 the cosine laws
⎪
⎪
⎪
⎭
b2 = a 2 + c2 − 2ac cos B̂

a b c
= = sine law
sin  sin B̂ sin Ĉ

In general,

1. knowing all three sides of a triangle we use the cosine law;

2. knowing two sides and one angle we use the sine law if the angle is not in between the two
sides, otherwise we use the cosine law;
3. knowing one side and two angles we use the sine law.

Given triangle whose sides are AB = 4, AC = 6, BC = 8, here c = 4, b = 6, a = 8, we write

b2 + c2 − a 2 62 + 42 − 82 1
cos  = = =−
2bc 2×6×4 4

(b) Given triangle as AB = 3, AC = 4, α = π/3, i.e. c = 3, b = 4, B ÂC = 60◦ , we write

1 √
a 2 = b2 + c2 − 2bc cos 60◦ = 25 − 24 ∴ a = 13
2

Exercise 7.3, page 132

1. Vector addition is done by a simple chaining head of one vector to tail of the next one. Vector
representing the total sum starts at tail of the first and ends at the head of last vector in the chain.
(a) All vectors in 1D space are collinear (i.e. parallel), therefore their direction is already known.
As a consequence, knowing only their magnitude is sufficient. In most general sense, numbers are
viewed as the special case of vectors. As illustrated in Fig. 7.19 (left), sum of two numbers equals
to sum of their magnitudes after accounting for their sign.

2 + 3 = 5 and 2 + (−3) = −1
Solutions 155

(b) in 2D space not all vectors are collinear, which is to say that in general case, the sum of two
vectors is not simply equal to the sum of their magnitudes. Instead, we must apply the rules of
triangles. see Fig. 7.19 (right). In order to illustrate the practicality of matrix notification, we note
that modules of vector projections add in the same manner as “regular” numbers. Formally, given
that

#» #» #»
c = #»
a + b if #»
a 1 b 3
a = x = and b = x =
ay 2 by −1

we write

#» #» 1+3
c = #»
a b 1 3 4 c
a +b = x + x = + = = = x
ay by 2 −1 2−1 1 cy

Fig. 7.19 Example 7.3-1 y

0
x
+ =

+ (- ) = -1 x
0 x x
x x
-

2. Multiplying a vector by a constant does not change the vector direction, only its magnitude, see
Fig. 7.20. The constant multiplies all vector components, in this case both xx and xy .

#» 2×3
a = 2 #»
x 3 6
(a) x =2 x =2 = =
xy 2 2×2 4
#» 2 #» 2 xx 2 3 2/3 ×3 2
(b) b = x = = = =
3 3 xy 3 2 2/3 ×2 4/3

#» 3 3 3 3 −3/4 × 3 −9/4
c = − #»
xx
(c) x =− =− = =
4 4 xy 4 2 −3/4 × 2 −3/2
156 7 Linear Algebra

Fig. 7.20 Example 7.3-2

y y    
3 6
2· =
  2 4
x 3 a
2
x  
x 3
y y 2
   
2 3 2
· =
3 2 4/3
c x
b    
x 3 3 −2.25
− · =
4 2 −1.50

Exercise 7.4, page 133

1. In its basic form, a function is equivalent to a machine that for any given number input x performs
an operation f (x) and produces the corresponding output y. Similarly, for any given vector input
#»
v linear operation L( #» #»
v ) performs an operation and produces the corresponding vector output w,
as illustrated in Fig. 7.21.

Fig. 7.21 Example 7.2-1 function linear transformation

x y v w

       
1 2 −2 −3/2 −3 3
... − 3 2 5 f (x) . . . 9 4 25 L(v ) −6 −4 2
−3 2 1
··· ···

input output vector input vector output

Another way to visualize operation of a linear transformation is to imagine that the input vector
is displaced in space until it overlaps the output vector, see Fig. 7.22.

Fig. 7.22 Example 7.2-1

R2 space
y  
  3
−2 2 L(v )
2
1
vector input
1 w

vector output
v
0
-2 -1 0 1 2 3 x
Solutions 157

However, linear operation L( #» v ) displaces all vectors in R 2 vector space, that is to say it displaces
all the points on the surface. For example, after linear transformation L( #» v ) displaces basis vector
#» #» #»
i = (1, 0) into the new vector L( i ) = (1, 3), while at the same time basis vector j = (0, 1) is
#»
displaced into the new vector L( j ) = (2, −1) relative to the initial position, see Fig. 7.23. Since
all points on the surface are also simultaneously displaced, we can imagine that the surface is
#» #»
“elastic” and it “stretched”. In this analogy, now L( i ) and L( j ) represent basis vectors in the
transformed space.
In summary, for a transformation to be linear it must obey the following rules:

1. after the transformation all grid lines must stay straight.

2. all the grid lines must stay parallel and equidistant because the distance between them must
always be one basis vector.
3. the origin point must not move.
#» #»
Due to these rules, the numerical form of a vector as written relative to the basis vectors ( i , j )
#» #»
before, and (L( i ), L( j )) after the transformation stays the same.

Fig. 7.23 Example 7.2-1 R2 space L(v )

j L(i )
L(j )
(0, 0) i (0, 0)

2. For example, we can create a graph as in Fig. 7.24, so that #»v is transformed into L( #»
v ). We note
the relationship between the original and transformed greed lines. In order to better visualize
space transformation, the original grid is kept in the background. In addition, by following the
#» #»
transformation of basis vectors ( i , j ) we can visualize how the original R 2 space is rotated,
#» #»
flipped, and stretched into the new basis vectors L( i ), L( j ).
By inspection of grid in transformed space, we write the expression for transformed L( #» v ) within
#» #»
new coordinating system based on basis vectors L( i ), L( j ), and we compare it with the original
vector #»
v as
#» #»
L( #»
v ) = −2L( i ) + L( j ) (7.3)
and
#» #» #»
v = −2 i + j (7.4)

In order to be truly linear transformation, the numerical forms of (7.3) and (7.4) are same.
#» #»
Consequently, coordinates of new basis vectors are, by definition, L( i ) = (1, 0) and L( j ) =
(0, 1).
158 7 Linear Algebra

Fig. 7.24 Example 7.4-2

R2 space
v = −2i + j L(v ) L(i )
L(j )
(0, 0)

L(v )
−2L(i )
j
v
i  
(0, 0) −2
L(j ) 1

#» #»
By inspection of graph, if expressed relative to the original basis vectors i , j , in this example
we find that coordinates of
#» #»
i = (1, 0) ∴ L( i ) = (1, 3)
#» #»
j = (0, 1) ∴ L( j ) = (2, −1)
∴ L( #»
v ) = (0, −7)

We can deduce matrix form of L( #»

v ) as follows.

#» 1 #» 2
L( i ) = L( j ) =
3 −1
∴
#» #» −2
L( #»
1 2 2 0
v ) = −2L( i ) + L( j ) = −2 + = + =
3 −1 −6 −1 −7

In compact matrix form, for an arbitrary #»

x we prefer to write

#» # #» #» $
∴ L( #»
x ) = L #»
x = L( i ) L( j ) #»
x 1 2 xx
x = x x =
xy 3 −1 xy

In this interpretation, all transformations of space can be formalized as the transformation matrix-
vector product.
3. Given data, we write

#» 3 #» 1
L( i ) = and L( j ) =
1 2

#» −1 #» #»
x = = (−1) i + 2 j
2
Solutions 159

Therefore,
#» #»
L( #»
x ) = (−1)L( i ) + 2L( j )
or, if measured in the original space:
3 1 (−1) × 3 + 2 × 1
= (−1) +2 =
1 2 (−1) × 1 + 2 × 2
−1
=
3

The relation between the original and transformed space is illustrated in Fig. 7.25.

Fig. 7.25 Example 7.4-3

   
3 1 −1
1 2 2 L(x) 
1  
2L(j) 2 3
1
(0, 0)
−1L(i)

#» #»
4. We follow movement of basis vectors ( i , j ) on graphs and read their new coordinates relative
to
(a) Rotation of space is shown in Fig. 7.26. At the final position we find

#» #» −1 0 −1
∴ L( #»
0
L( i ) = and L( j ) = x) =
1 0 1 0

which is matrix that performs 90◦ rotation of all points in R 2 space.

Fig. 7.26 Example 7.4-4(a) 90o rotation

 
0
  
0 j   1
−1
1 90◦  
1 0 L(i )
i 0 L(j )
160 7 Linear Algebra

(b) Operation of shear bends the space so that horizontal vectors are unchanged while vertical
vectors bend −45◦ , as shown in Fig. 7.27. At the final position we find

#» #»
∴ L( #»
1 1 11
L( i ) = and L( j ) = x) =
0 1 01

which is matrix that performs shear of all points in R 2 space.

Fig. 7.27 Example 7.4-4(b) shear

 
1
    1  
0 1 1
1 j 0 L(j ) 0
i L(i )

(c) Linear transformations may be performed one after another. In the first step we do rotation
L1 ( #»
x ) then sheer L2 ( #»
x ), as shown in Fig. 7.28. At the final position we find

0 −1 0 −1 1 −1
L1 ( #» and L2 ( #» ∴ L( #»
11 11
x) = x )= x) = =
1 0 01 01 1 0 1 0

which is matrix that performs rotation plus shear of all points in R 2 space. We note that in this
interpretation, L( #»
x ) is matrix that does both transformations at the same time. Alternatively,
multiple transformations are viewed as being equivalent to product of their respective matrix
where transformation that is done first is written on the right side.
Fig. 7.28 Example 7.4-4(c) 90o rotation then shear

 
  1
0  
  1
−1 1 −1
0 L1 (i ) 0 L(i )
L1 (j ) L(j )
Solutions 161

#» #»
5. We follow movement of basis vectors ( i , j ) on graphs and read their final coordinates.
#»
(a) This transformation simply enlarges the space by factor three in i direction and by factor
#»
two in j direction, see Fig. 7.29.

#» #»
L( #»
30 3 0
x) = ∴ L( i ) = and L( j ) =
02 0 2

#»
We note that relative position between basis vectors did not change, i.e. j is still on left relative
#»
to i . In addition, the unity area increased by 3 × 2 = 6 factor. This area multiplication factor is
geometrical interpretation of determinant, we use the syntax
% %
%3 0%
|L( #»
x )| = %% %% = 3 × 2 − 0 × 0 = 6
02

Fig. 7.29 Example 7.4-5(a)

L(x)
 
L(j )
  +6
0 3 0
1   0 2
1 1
(0, 0) 0 (0, 0) L(i )

(b) This transformation performs multiple operations at the same time. By following basis
vectors, we could imagine that 2D surface was rotated and then flipped over, see Fig. 7.30. In
analogy to a sheet of a transparent paper, if the basis vectors are drawn on one side of the paper
#» #»
so that j is on the left side of i , then if we flip the paper and look again at basis vectors, then
#» #»
from that perspective j is found on the right side of i .

#» #»
L( #»
1 2 1 2
x) = ∴ L( i ) = and L( j ) =
1 −1 1 −1
162 7 Linear Algebra

where determinant is calculated as

% %
%1 2%
|L( #»
x )| = %% % = 1 × (−1) − 1 × 2 = −3
1 −1%

Geometrical interpretation is that the unity area is increased by factor of three; however, 2D
surface is also “flipped over”, which is indicated by the negative sign of “-3” multiplication factor.

Fig. 7.30 Example 7.4-5(b)

L(x)
    L(i )
0 1 2
1   1 −1 -3
1 (0, 0)
1
(0, 0) 0 L(j )

(c) Linear transformations may be performed one after another. In the first step we do rotation
L1 ( #»
x ) then sheer L2 ( #»
x ), as shown in Fig. 7.31. At the final position we find

#» #»
L( #»
42 4 2
x) = ∴ L( i ) = and L( j ) =
21 2 1

where determinant is calculated as

% %
%4 2%
|L( #»
x )| = %% %% = 4 × 1 − 2 × 2 = 0
21

#» #»
We note that L( i ) and L( j ) are collinear. After the transformation, 2D surface collapsed into
1D line. In the paper analogy, we can imagine that 2D surface is rotated so that we see its sideway
projection, i.e. line. This reduction of space is geometrical interpretation of determinant equal
zero. Alternative interpretation is that at least one of equation within system of equations in L( #»
x)
is not independent.

Fig. 7.31 Example 7.4-5(c)

L(x)
L(i )
   
0 4 2 L(j )
1   2 1
1 1 (0, 0) 0
(0, 0) 0
Solutions 163

6. This type of problems we solve by both geometric and numerical methods.

#»
(a) Given three vectors, see Fig. 7.32, we find that 5 b to #»
a closes the vector triangle, thus

#» = #» #»
m a + 7b

General form of linear combination takes the form of a sum, thus

#» = n #» #»
m a +kb
∴
9 2 1
=n +k
1 1 0

which is matrix form of the following system of equations

9 = 2n + 1 k
1 = 1n + 0k

#» = #» #»
therefore, n = 1 and k = 7, which is to say m a + 7b.

(b) Similarly,

#» = #» −1
u + #» #» = 3 1 3
m v +w + + =
−1 −2 7 4

Then,

#» = n #»
m u + k #»
v
∴
3 3 1
=n +k
4 −1 −2

which is matrix form of the following system of equations

3 = 3n + 1k
4 = −1n − 2k
164 7 Linear Algebra

Cramer’s rule gives,

% % ⎫
% 3 1%
Δ = %% % = −5 = 0⎪
⎪
⎪
−1 −2% ⎪
⎪ Δn −10
⎪
⎪ n= = =2
% % ⎪
⎬ −5
%3 1% Δ
Δn = %% % = −10 ∴
4 −2% ⎪
⎪
% % ⎪
⎪
Δk 15
⎪
⎪ k= = = −3
% 3 3% ⎪
⎪ Δ −5
Δk = %% % = 15 ⎭
−1 4%

#» = 2 #»
which is to say, m u − 3 #»
v

Fig. 7.32 Example 7.4-6(a) R2 espace

y a: (2,1) 7xb
1

m: (9,1)
0
0 2 4 6 8 10 x
b: (1,0)

7. One method to prove that vectors are linearly independent is to confirm that the system’s
determinant is Δ = 0. Alternatively, graphically the vectors should not be collinear (i.e. parallel)
(a) Linear combination is

#» #»
c = k #»
a + nb
∴
7 3 −2
=k +n
−4 −2 1

which is matrix form of the following system of equations

7 = 3k − 2n
−4 = −2k + n

therefore, determinant is calculated as

% %
% 3 −2%
%
Δ=% % = −1 = 0
−2 1%
Solutions 165

which is to say that equations (i.e. vectors) are indeed independent. Cramer’s rule gives,
% % ⎫
% 7 −2% Δk −1
Δk = %% % = −1⎪
⎪
⎪ k= = =1
−4 1% ⎬ Δ −1
% % ∴
% 3 7% ⎪
⎪
= %% %=2 ⎭ ⎪ Δn 2
Δn
−2 −4% n= = = −2
Δ −1
#»
which is to say, #»
c = #»
a − 2b

(b) Linear combination is

#» #»
c = k #»
a + nb
∴
0 1 3
=k +n
−2 2 4

which is matrix form of the following system of equations

0 = k + 3n
−2 = 2k + 4n

therefore, determinant is calculated as

% %
% 1 3%
Δ = %% %% = −2 = 0
24

which is to say that equations (i.e. vectors) are indeed independent. Cramer’s rule gives,
% % ⎫
% 0 3% Δk 6
Δk = %% %=6 ⎪ ⎪
⎪ k= = = −3
−2 4% ⎬ Δ −2
% % ∴
%1 0% ⎪
⎪ −2
= %% ⎪
% = −2⎭ Δn
Δn
2 −2% n= = =1
Δ −2
#»
which is to say, #»
c = −3 #»
a +b

8. Main idea is that if vectors are linearly independent they are not collinear. Possible ways to
formalize that statement is to write

if #»
x = k #»
y ∴ ( #»
x , #»
y ) are collinear

where k is the multiplying constant. Alternatively, determinant of matrix that includes collinear
vectors equals zero.
166 7 Linear Algebra

(a) Stipulating that vectors are collinear, numerical method gives

#»
b = k #»
a
∴
2 3
=k
6 λ

which is matrix form of the following system of equations

 k = 3/2
2 = 3k
∴
6 = λk 6 = λk ∴ λ = 9

#» #»
which is to say, if λ = 9, then ( #»
a , b ) are collinear, i.e. dependent and can be written as b = 3/2 #»
a,
which is easily verified by performing the multiplication.
Alternatively, we calculate determinant as

#»
b = k #»
2 3
a ∴ =k
6 λ

Determinant is then calculated as

% %
%2 3%
Δ = %% %% = 3 × 6 − 2λ = 2λ − 18 = 0 ∴ λ = 9
6λ

which leads into the same conclusion as before.

(b) Given vectors
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
6 3 λ
#» #» ⎣ ⎦ #»
a = ⎣8⎦ b = 4 c = ⎣ 0⎦
4 2 1

#» #»
we see that #»
a = 2 b , which is to say that ( #»
a , b ) are collinear (i.e. dependent) regardless of λ.
Alternatively, determinant is
% %
%6 3 λ% 6 3
% %
Δ = %%8 4 0%% 8 4 = +6 × 4 × 1 + 3 × 0 × 4 + λ × 8 × 2
%4 2 1% 4 2

−4×4×λ−2×0×6−1×8×3

=
24 
+16λ 
−16λ 
−24
=0

that is to say, the determinant equals zero regardless of λ, same conclusion as already reached.
Solutions 167

Exercise 7.5, page 134

1. There are multiple methods for calculating determinants, here we illustrate two of the most
common ones where either “|A|” or “Δ” annotates “determinant of matrix A”.
(a) Second order determinants are calculated as the difference between products along their two
diagonals.
% %
%3 0%%
|A2 | = Δ2 = %% =3×2−0×0=6=0
0 2%

(b) Similarly,
% %
%1 2%%
|A2 | = %% = 1 × (−1) − 1 × (−2) = −3 = 0
1 −1%

(c) Here, we note that the determinant equals zero, which is to say that two rows (i.e. equations)
are not independent. Indeed, the two rows (i.e. equations) are related by a simple factor of two.
% %
%4 2%%
|A2 | = %% =4×1−2×2=0
2 1%

(d) Following the same idea of third and higher order determinants are calculated by calculating
cross-products

+ + + − − −
2 3 4 2 3
D3 =
5 −22 1 5 −2
1 2 3 1 2

= +(2 × (−2) × 3)+(3 × 1 × 1)+(4 × 5 × 2)

−(1 × (−2) × 4)−(2 × 1 × 2)−(3 × 5 × 3)
= −12 + 3 + 40 +8 − 4 − 45
= −10 = 0
168 7 Linear Algebra

Alternatively, we can use the method of cofactor expansions to calculate the same results as

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
2 3 4 2 3 4 2 3 4
D3 = ⎣ 5 -2 1 ⎦ + ⎣ 5 -2 1 ⎦ + ⎣ 5 -2 1 ⎦
1 2 3 1 2 3 1 2 3
     
 −2 1 
1+2 5 1 5 −2
= (−1)1+1 2   + (−1) 3 1 + (−1)1+3 4 
2 3 3 1 2
= 2 × (−2 × 3 − 2 × 1) − 3 × (5 × 3 − 1 × 1) + 4 × (5 × 2 − 1 × (−2))
= −10

2. Area of a surface formed by two vectors in R 2 space equals to the associated determinant. The
same is valid in higher order spaces, thus volume of a cuboid formed by three vectors in R 3 space
equals to the associated determinant, etc.
(a) Given three vectors in R 3 , we calculate the volume as
% %
% 1 2 1% 1 2
% %
Δ = %%−3 1 2%% −3 1 = 1 + 4 + 9 − 1 + 6 + 6 = 25
% 1 −3 1% 1 −3

(b) Similarly,
% %
%1 0 3% 1 0
% %
Δ = %%0 1 4%% 0 1 = −9 − 8 = −17
%3 2 0% 3 2

We conclude that V = 17, albeit due to negative sign it is “inverted” in R 3 space.

3. Second order determinants are calculated by simple cross-product

% %
%2 −1%
(a) %% % = 2 × 0 − 1 × (−1) = 1
1 0%
%√ √ %
% 3 −3 2% √ √ √ √ √ √ √
(b) %%√ √ % = 3 × 2 2 − 3 × (−3 2) = 2 6 + 3 6 = 5 6
%
3 2 2
% %
% 1 i%
(c) %% % = 1 − (−i) × i = 1 − 1 = 0
−i 1%
% %
% sin α cos α %
(d) %% % = sin α × sin α − (− cos α) × cos α = sin2 α + cos2 α = 1
− cos α sin α %

4. Third order determinants may be calculated by cross-products.

% % % %
%3 2 1% 3 2 %2 3 1% 2 3
% % % %
(a) %%4 5 6%% 4 5 = −21 (b) %%3 −1 2%% 3 −1 = 39
%8 9 7% 8 9 %1 1 −3% 1 1
Solutions 169

% % % %
%5 0 4% 5 0 % 1 3 5% 1 3
% % % %
(c) %%8 0 −7%% 8 0 = 134 (d) %% 7 9 11%% 7 9 = 0
%3 2 1% 3 2 %13 15 17% 13 15

Exercise 7.6, page 134

1. In order to keep track of matrix transformations, the three equations are numbered as (1), (2), (3)
and their coefficients are placed in first, second, and third matrix row. In consequence, the vertical
columns are where the coefficients of x, y, z variables are placed. In addition, the solutions of each
equation are placed in the fourth column of expanded matrix.

(x) (y) (z)

5x − 5y − 15z = 40 (1)
4x − 2y − 6z = 19 (2)
3x − 6y − 17z = 41 (3)

The objective of this technique is to transform the expanded matrix into its equivalent diagonal-
ized form (i.e. all coefficient found at diagonal equal “1”, while at the same time all the other
coefficients equal to “0”.
⎡ ⎤ ⎡ ⎤
5 −5 −15 40 100a 1 ×x+ 0 ×y+ 0 ×z = a x=a
⎣ 4 −2 −6 19 ⎦ ⇒ ⎣ 0 1 0 b ⎦ ∴ 0 ×x+ 1 ×y+ 0 ×z = b ∴ y =b
3 −6 −17 41 001 c 0 ×x+ 0 ×y+ 1 ×z = c z =c

We start by transforming coefficient found at first-row–first-column position into “1”. In order to

do so, obviously, it must be divided by five. Which is to say that all coefficients in the first row are
also divided by five. Then systematically transform the other coefficients at and below diagonal.
In order to annotate transformations, for example, the operation of division of each term in the
first row equation (1) by five is written as “← (1) ÷ [5]”. Of course, there are multiple ways to
transform the matrix that do not influence the final solution, it is matter of practice and preference
which particular intermediate steps are taken. However, good strategy is to transform the diagonal
and lower triangle coefficients first. Then, transform the upper triangle coefficients.
⎡ ⎤
5 −5 −15 40 ← (1) ÷ [5]
⎣4 −2 −6 19⎦
3 −6 −17 41
⎡ ⎤
1 −1 −3 8
⎣4 −2 −6 19⎦ ← (2) − 4 × (1)
3 −6 −17 41 ← (3) − 3 × (1)
⎡ ⎤
1 −1 −3 8
⎣0 2 6 −13⎦ ← (2) ÷ [2]
0 −3 −8 17
⎡ ⎤
1 −1 −3 8
⎣0 1 3 −13/2⎦
0 −3 −8 17 ← (3) + 3 × (2)
170 7 Linear Algebra

⎡ ⎤
1 −1 −3 8 ← (1) + (2)
⎣0 1 3 −13/2⎦
0 0 1 −5/2
⎡ ⎤
1 0 0 3/2
⎣0 1 3 −13/2⎦ ← (2) − 3 × (3)
0 0 1 −5/2
⎡ ⎤
1 0 0 3/2  
⎣0 1 0 ⎦ 3 5
1 ∴ {x, y, z} = , 1, −
2 2
0 0 1 −5/2

2. Since all three points A, B, C must satisfy the quadratic equation f (x) at the same time, we write
the following system of equations

∴ ax 2 + bx + c = f (x)
(x, y) = (−2, 20) ∴ a(−2)2 + b(−2) + c = 20

(x, y) = (1, 5) ∴ a(1)2 + b(1) + c = 5

(x, y) = (3, 25) ∴ a(3)2 + b(3) + c = 25

This problem can be formalized in matrix form as

(a) (b) (c)

4a − 2b + c = 20 (1)
a + b + c = 5 (2)
9a + 3b + c = 25 (3)

We start by taking advantage of the fact that the first coefficient of (2) already equals one, thus
we exchange positions of (1) and (2). In any case, the initial order of equations is arbitrary.
⎡ ⎤
4 −2 1 20 (1) ↔ (2)
⎣1 1 1 5⎦
9 3 1 25
⎡ ⎤
1 11 5
⎣4 −2 1 20⎦ ← (2) − 4 × (1)
9 3 1 25
⎡ ⎤
1 1 1 5
⎣0 −6 −3 0⎦ ← (2) ÷ [−6]
9 3 1 25
⎡ ⎤
11 1 5
⎣0 1 1/2 0⎦
9 3 1 25 ← (3) − 9 × (1)
Solutions 171

⎡ ⎤
1 1 1 5
⎣0 1 1/2 0⎦
0 −6 −8 −20 ← (3) + 6 × (2)
⎡ ⎤
11 1 5
⎣0 1 1/2 0⎦
0 0 −5 −20 ← (3) ÷ [−5]

⎡ ⎤
1 1 1 5 ← (1) − (2)
⎣0 1 1/2 0⎦
00 14
⎡ ⎤
1 0 1/2 5
⎣0 1 1/2 0⎦ ← (2) − 1/2(3)
00 14
⎡ ⎤
1 0 1/2 5 ← (1) − 1/2(3)
⎣0 1 0 −2⎦
00 1 4
⎡ ⎤
100 3
⎣0 1 0 −2⎦ ∴ {a, b, c} = {3, −2, 4}
001 4

Which is to say that f (x) = 3x 2 − 2x + 4.

3. Cramer’s rule is based on determinants of the extended matrix.
(a) Principal and two sub-determinants are
% %
4 −3 0 %4 −3%
A= ∴ Δ=% % % = 4 × 3 − 2 × (−3) = 18 = 0
2 3 18 2 3%
% %
% 0 −3%
%
Δx1 = % % = 0 × 3 − 18 × (−3) = 3 × 18
18 3%
% %
%4 0%
%
Δx2 = % % = 4 × 18 − 2 × (0) = 4 × 18
2 18%

By definition, we find

Δx 1 3×
18
x1 = = =3
Δ 

18
Δx 1 4×
18
x2 = = =4
Δ 

18
172 7 Linear Algebra

(b) Principal and three sub-determinants are

⎡ ⎤ ⎡ ⎤
5 −5 −15 40 5 −5 −15 5 −5
A = ⎣4 −2 −6 19⎦ ∴ Δ = ⎣4 −2 −6⎦ 4 −2 = 10 = 0
3 −6 −17 41 3 −6 −17 3 −6
⎡ ⎤
40 −5 −15 40 −5
Δx = ⎣19 −2 −6⎦ 19 −2 = 15
41 −6 −17 41 −6
⎡ ⎤
5 40 −15 5 40
Δy = ⎣4 19 −6⎦ 4 19 = 10
3 41 −17 3 41
⎡ ⎤
5 −5 40 5 −5
Δz = ⎣4 −2 19⎦ 4 −2 = −25
3 −6 41 3 −6

By definition, we find

Δx 15 3
x= = =
Δ 10 2
Δy 10
y= = =1
Δ 10
Δz −25 5
z= = =−
Δ 10 2

4. Matrix form of this system is

#»
A #»
v = b
where
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
110 x 7
#» #»
A = ⎣ 0 1 1⎦ v = ⎣y ⎦ b = ⎣8⎦
−1 0 2 z 7

That is to say,
⎡ ⎤⎡ ⎤ ⎡ ⎤
110 x 7
⎣ 0 1 1⎦ ⎣y ⎦ = ⎣8⎦
−1 0 2 z 7

Determinants are therefore,

% %
% 1 1 0% 1 1
% %
Δ = %% 0 1 1%% 0 1 = 2 + (−1) + 0 − 0 − 0 − 0 = 1 = 0
%−1 0 2% −1 0
Solutions 173

% %
%7 1 0% 7 1
% %
Δx = %%8 1 1%% 8 1 = 14 + 7 + 0 − 0 − 0 − 16 = 5
%7 0 2% 7 0
% %
% 1 7 0% 1 7
% %
Δy = %% 0 8 1%% 0 8 = 16 + (−7) + 0 − 0 − 7 − 0 = 2
%−1 7 2% −1 7
% %
% 1 1 7% 1 1
% %
Δz = %% 0 1 8%% 0 1 = 7 + (−8) + 0 − (−7) − 0 − 0 = 6
%−1 0 7% −1 0

Determinant Δ = 1(= 0) therefore we calculate

Δx 5 Δy 2 Δz 6
x= = =5 y= = =2 z= = =6
Δ 1 Δ 1 Δ 1
⎡ ⎤
5
that is to say, #»
v = ⎣2⎦
6
or, (x, y, z) = (5, 2, 6)

Geometric interpretation is that transformation A stretches the space as necessary until #»

v is equal
#»
to b .
5. It is important to master multiple techniques for solving the system of linear equations.
(a) For example, we can use matrix form of the given system of equations and Cramer’s rule to
write
% % ⎫
%1 2% ⎪
Δ=% % % = −7 ⎪
⎪
3 −1% ⎪
⎪ Δ −21
⎪
⎪x= x = =3
% % ⎪
⎬ −7
1 2 x −5 %−5 2% Δ
= ∴ Δx = % % % = −21
3 −1 y 13 13 −1% ⎪
⎪
% % ⎪
⎪
Δy 28
%1 −5% ⎪y =
⎪ = = −4
% % ⎪
⎪ Δ −7
Δy = % = 28 ⎭
3 13%
174 7 Linear Algebra

(b) For example we can use the elimination method

x + 2y = −8
2x − y = −1 ×2

x + 2y = −8
4x − 2y = −2 ∴

5x = −10 ∴ x = −2
−2 + 2y = −8 ∴ y = −3

(c) For example we can use the matrix transformation method and given system

x + y + z = 36
2x − z = −17
6x − 5z = 7

we write,

⎡ ⎤
1 1 1 36
⎣2 0 −1 −17⎦ ← (2) ÷ [2]
6 0 −5 7
⎡ ⎤
1 1 1 36
⎣1 0 −1/2 −17/2⎦ ← (2) − (1)
6 0 −5 7 ← (3) − 6 × (1)
⎡ ⎤
1 1 1 36
⎣0 −1 −3/2 −89/2⎦
0 −6 −11 −209 ← (3) − 6 × (2)
⎡ ⎤
1 1 1 36
⎣0 1 3/2 89/2⎦
0 0 −2 58 ← (3) ÷ [−2]
⎡ ⎤
1 1 1 36 ← (1) − (2)
⎣0 1 3/2 89/2⎦
0 0 1 −29
Solutions 175

⎡ ⎤
1 0 −1/2 −17/2 ← (1) + 1/2(3)
⎣0 1 3/2 89/2⎦ ← (2) − 3/2(3)
0 0 1 −29
⎡ ⎤
1 0 0 −23
⎣0 1 0 88⎦
0 0 1 −29
∴ {x, y, z} = {−23, 88, −29}

(d) For example, using Cramer’s rule

% %
%2 3 −1% 2 3
% %
Δ = %%1 1 2%% 1 1 = 2 + 12 + 1 + 2 + 4 − 3 = 18
%2 −1 1% 2 −1
% %
%5 3 −1% 5 3
% %
Δx = %%7 1 2%% 7 1 = 5 + 6 + 7 + 1 + 10 − 21 = 8
%1 −1 1% 1 −1
% %
%2 5 −1% 2 5
% %
Δy = %%1 7 2%% 1 7 = 14 + 20 − 1 + 14 − 4 − 5 = 38
%2 1 1% 2 1
% %
%2 3 5% 2 3
% %
Δz = %%1 1 7%% 1 1 = 2 + 42 − 5 − 10 + 14 − 3 = 40
%2 −1 1% 2 −1

Then,

Δx 4 Δy 19 Δz 20
x= = y= = z= =
Δ 9 Δ 9 Δ 9

Exercise 7.7, page 135

1. Matrix-vector product is calculated as linear combination.

1 1 3 1 1 3
(a) =3 +0 =
2 −1 0 2 −1 6
2 −1 −5 2 −1 −13
(b) = −5 +3 =
3 2 3 3 2 −9
2 −3 3 2 −3 12
(c) =3 + (−2) =
1 2 −2 1 2 −1
1 2 −8 1 2 −10
(d) = (−8) + (−1) =
2 −1 −1 2 −1 −15
176 7 Linear Algebra

In geometric interpretation, we follow the basis vectors. Columns of matrix hold the coordinates
of transformed basis vectors as measured in the original space.
For example, in (a) we look at matrix as linear transformation of vector (3, 0). The vector
coordinates stay numerically same in both original and transformed spaces, see Fig. 7.33. Note
that the product solution (3, 6) gives coordinates of #»
v as measured in the original space. In order
to create graphs for (b) to (d) we take the same approach.

Fig. 7.33 Example 7.7-1 R2 space L(i ) L(j) L(v ) = 3L(i )

 
1 1
v = 3i + 0j 2 −1  
3
0

j v L(
v)
  6
(0, 0) i 3
0 3
(0, 0)

2. Geometrically, matrix–matrix product can be interpreted as “linear transformation of linear

transformation”. Thus, we can reason it out in two steps.
#» #»
(a) First matrix on the left side of product holds coordinates of the original basis vectors ( i , j )
whose magnitudes equal one.

1 1 2 −1
2 −1 3 2

Result of this first transformation are transformed basis vectors whose coordinates (in the
original space) are (1, 2) and (1, −1), see Fig. 7.34 (narrow line arrows). Second matrix
contains coordinates of these two new basis vectors, as measured in the transformed space, see
Fig. 7.34(wide line arrows).

1 1 2 −1
2 −1 3 2

Therefore, we read the coordinates of two final vectors in the original units (grid in the
background) and conclude that

1 1 2 −1 5 1
= (7.5)
2 −1 3 2 1 −4

Alternatively, we use numerical method to multiply two matrix formally and to search the product
result in matrix form as follows

1 1 2 −1 i j
= x x
2 −1 3 2 iy jx
Solutions 177

where matrix–matrix multiplication is reduced to two matrix-vector multiplications and each

column of the resulting matrix (7.5) are calculated as follows

#» 1 1 2 1 1 (2 + 3) 5
i : =2 +3 = =
2 −1 3 2 −1 (4 − 3) 1
#» 1 1 −1 1 1 (−1 + 2) 1
j : = (−1) +2 = =
2 −1 2 2 −1 (−2 − 2) −4

Both methods, naturally, produce the same result.

(b) Similarly, we multiply two matrix as follows

2 −1 2 −3 i j
= x x
3 2 1 2 iy jx

and, matrix–matrix multiplication is reduced to two matrix-vector multiplications as follows

#» 2 −1 2 2 −1 (4 − 1) 3
i : =2 + = =
3 2 1 3 2 (6 + 2) 8
#» 2 −1 −3 2 −1 (−6 − 2) −8
j : = (−3) +2 = =
3 2 2 3 2 (−9 + 4) −5

Therefore,

2 −1 2 −3 3 −8
=
3 2 1 2 8 −5

(c) We multiply two matrix as follows

2 −3 1 1 i j
= x x
1 2 2 −1 iy jx

and, matrix–matrix multiplication is reduced to two matrix-vector multiplications as follows

#» 2 −3 1 2 −3 (2 − 6) −4
i : = +2 = =
1 2 2 1 2 (1 + 4) 5
#» 2 −3 1 2 −3 (2 + 3) 5
j : = + (−1) = =
1 2 −1 1 2 (1 − 2) −1

Therefore,

2 −3 1 1 −4 5
=
1 2 2 −1 5 −1
178 7 Linear Algebra

Fig. 7.34 Example 7.7-2

1
(0, 0) 5

−4

Exercise 7.8, page 136

1. Matrices with different sizes are multiplied as


n
(A · B)ij = aik bkj (7.6)
k=1

where size of product matrix (A · B)ij depends on the two matrices being multiplied. For example,
given A2×3 (two rows and three columns) and B3×2 (three rows and two columns) the resulting
matrix must have size [2×2] (two rows and two columns), which is to say that not all products are
possible. That is because the number of columns in the first matrix must be equal to the number
of rows in the second matrix, otherwise the product is not possible.
The sum of row–column products in (7.6) is calculated as follows.
⎡ ⎤
1 2
3 05 ⎣
(i, j ) = (1, 1) : 5 −1⎦ = 3 × 1 + 0 × 5 + 5 × 0 = 3
−2 −1 4
0 −6
(ab)11 = 3
⎡ ⎤
1 2
3 05 ⎣
(i, j ) = (1, 2) : 5 −1⎦ = 3 × 2 + 0 × (−1) + 5 × (−6) = −24
−2 −1 4
0 6
(ab)12 = −24
⎡ ⎤
1 2
3 05 ⎣
(i, j ) = (2, 1) : 5 −1⎦ = (−2) × 1 + (−1) × 5 + 4 × 0 = −7
−2 −1 4
0 −6
(ab)21 = −7
⎡ ⎤
1 2
3 05 ⎣
(i, j ) = (2, 2) : 5 −1⎦ = (−2) × 2 + (−1) × (−1) + 4 × (−6) = −27
−2 −1 4
0 6
(ab)22 = −27
Solutions 179

Therefore:

3 −24
A·B =
−7 −27

Exercise 7.9, page 136

1. The matrix-vector products are calculated as follows.

(a) In this case row–vector products are,
⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
−2 −9 −1 1 (−2 − 9 − 2) −13
A x#»1 = ⎣ 2 −6 −5⎦ ⎣1⎦ = ⎣ (2 − 6 − 10) ⎦ = ⎣−14⎦ = 0
−4 −3 4 2 (−4 − 3 + 8) 1

(b) However, the same matrix A and x#»2 product is,

⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
−2 −9 −1 13 (−26 + 36 − 10) 0
A · x#»2 = ⎣ 2 −6 −5⎦ ⎣−4⎦ = ⎣ (26 + 24 − 50) ⎦ = ⎣0⎦
−4 −3 4 10 (−52 + 12 + 40) 0

2. By definition, eigenvector #»
v has the property that when multiplied by a matrix A gives a scalar
multiple of itself. The multiplying scalar is called eigenvalue,

A · #»
v = λ #»
v (7.7)

(a) Matrix A and v#»1 ,

3 −3 (3 × 3 − 1 × 3)
A · v#»1 = = λ v#»1
3 6 3
= = =2
2 −4 1 (2 × 3 − 4 × 1) 2 1

We conclude that, indeed v#»1 is eigenvector of matrix A and eigenvalue is λ = 2.

(b) However, the same matrix A multiplied by v#»2 gives

3 −3 (2 × 3 − 1 × 3)
A · v#»2 =
2 3 2
= = =λ
2 −4 1 (2 × 2 − 4 × 1) 0 1

We conclude that v#»2 is not eigenvector of matrix A because there is no eigenvalue that is λ
possible.

3. Calculating eigenvectors and eigenvalues involves calculation of the characteristic polynomial

defined as
180 7 Linear Algebra

A · #»
v = λ #»
v ∴ A · #»
v − λ #»
v =0
∴
(A − λ I ) #»
v =0 (7.8)

where I is the identity matrix (the square matrix of the same size as A with ones on the main
diagonal and zeros elsewhere). Note that we must use the identity matrix I (which is equivalent
to number one for number operations), so that we can actually subtract a scalar from a matrix. In
order to satisfy (7.8), since #»
v = 0 it must be that determinant of (A − λ I ) equals zero, thus

det (A − λ I ) = 0 (7.9)

is the formal definition of the characteristic polynomial.

Therefore, we write,
%⎡ ⎤ ⎡ ⎤% %⎡ ⎤ ⎡ ⎤%
% 4 6 10 1 0 0 %% %% 4 6 10 λ 0 0 %%
%
det (A − λ I ) = %%⎣ 3 10 13⎦ − λ ⎣0 1 0⎦%% = %%⎣ 3 10 13⎦ − ⎣ 0 λ 0⎦%%
% −2 −6 −8 0 0 1 % % −2 −6 −8 00λ %
% %
%4 − λ 6 10%%
%
= %% 3 10 − λ 13%% (use any method to calculate det)
% −2 −6 −8 − λ%
% % % % % %
%10 − λ 13%% % 3 13%% % 3 10 − λ%
= (4 − λ) %% − 6 % + 10 % %
−6 −8 − λ% %−2 −8 − λ% %−2 −6%
= (4 − λ) [(10 − λ)(−8 − λ) − (−6) × 13]
− 6 [3(−8 − λ) − (−2) × 13]
+ 10 [3 × 6 − (−2)(10 − λ)]
= (−λ3 − 6λ2 − 6λ − 8) + (−12 + 18λ) + (20 − 20λ)
= −λ3 + 6λ2 − 8λ = 0 (7.10)

Roots of the characteristic polynomial (7.10) are found as,

−λ3 + 6λ2 − 8λ = 0
λ3 − 6λ2 + 8λ = 0
λ(λ2 − 6λ + 8) = 0
λ(λ2 − 4λ − 2λ + 8) = 0
λ[λ(λ − 4) − 2(λ − 4)] = 0
λ(λ − 2)(λ − 4) = 0
∴
λ1 = 0, λ2 = 2, λ3 = 4
Solutions 181

For each of eigenvalues λ1,2,3 we calculate eigenvectors #»

v = (x, y, z) by solving (7.8) so that
⎡⎡ ⎤ ⎡ ⎤⎤ ⎡ ⎤
4 6 10 100 x
(A − λ I ) #»
v = ⎣⎣ 3 10 13⎦ − λ ⎣0 1 0⎦⎦ ⎣y ⎦
−2 −6 −8 001 z
⎡ ⎤⎡ ⎤ ⎡ ⎤
4−λ 6 10 x 0
=⎣ 3 10 − λ 13⎦ ⎣y ⎦ = ⎣0⎦
−2 −6 −8 − λ z 0

Case λ = 0: By using the matrix transformations method we write,

% %
% 4 6 10 0% ← (1) ÷ [4]
% %
% 3 10 13 0%
% %
%−2 −6 −8 0%
% %
% 1 3/2 5/2 0%
% %
% 3 10 13 0% ← (2) − 3 × (1)
% %
%−2 −6 −8 0% ← (3) ÷ [−2]
% %
%1 3/2 5/2 0%
% %
%0 11/2 11/2 0% ← (2) × 2/11
% %
%0 3/2 3/2 0% ← (3) × 2/3
% %
%1 3/2 5/2 0%
% %
%0 1 1 0%
% %
%0 1 1 0% ← (3) − (2)
% %
%1 3/2 5/2 0% ∴ 1 x + 3/2 y + 3/2 z = 0
% %
%0 1 1 0% ∴ 0 x + 1 y + 1 z = 0
% %
%0 0 0 0% ∴ z: free variable !

In the z-column and the third row we reached the identity 0 × z = 0, which is to say that any
value of z is valid choice, thus we choose a dummy value z = t.
With this choice, from the second and first equations we calculate

y + z = 0 ∴ y = −z ∴ y = −t
3 5
x+ y+ z=0
2 2
∴
3 5
x − t + t = 0 ∴ x + t = 0 ∴ x = −t
2 2

Which is to say that t has an arbitrary value except t = 0. The reason is that any multiple of
eigenvector is still vector in the same direction, thus we choose a convenient scale.
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
x −t −1 −1
v#»1 = ⎣y ⎦ = ⎣−t ⎦ = t ⎣−1⎦ = ⎣−1⎦ (since t = 0, we choose t = 1)
z t 1 1
182 7 Linear Algebra

Case λ = 2 : thus, we write,

% %
% 2 6 10 0% ← (1) ÷ [2]
% %
% 3 8 13 0%
% %
%−2 −6 −10 0%
% %
% 1 3 5 0%%
%
% 3 8 13 0% ← (2) − 3 × (1)
% %
%−2 −6 −10 0% ← (3) + 2 × (1)
% %
%1 3 5 0%
% %
%0 −1 −2 0% ← (2) ÷ [−1]
% %
%0 0 0 0%
% %
%1 3 5 0% ∴ 1 x + 3 y + 5 z = 0
% %
%0 1 2 0% ∴ 0 x + 1 y + 2 z = 0
% %
%0 0 0 0% ∴ z: free variable ! ∴ z = t

From the second and first equations we calculate

y + 2z = 0 ∴ y = −2z ∴ y = −2t

x + 3y + 5z = 0
∴
x − 6t + 5t = 0 ∴ x − t = 0 ∴ x = t

Thus we choose a convenient scale as

⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
x t 1 1
v#»2 = ⎣y ⎦ = ⎣−2t ⎦ = t ⎣−2⎦ = ⎣−2⎦
z t 1 1

Case λ = 4 : thus, we write,

% %
% 0 6 10 0% ← (1) ↔ (2)
% %
% 3 6 13 0%
% %
%−2 −6 −12 0%
% %
% 3 6 13 0% ← (1) ÷ [3]
% %
% 0 6 10 0%
% %
%−2 −6 −12 0% ← (3) ÷ [−2]
% %
%1 2 13/3 0%
% %
%0 6 10 0% ← (2) ÷ [6]
% %
%1 3 6 0% ← (3) − (1)
% %
%1 2 13/3 0%
% %
%0 1 5/3 0%
% %
%0 1 5/3 0% ← (3) − (2)
Solutions 183

% %
%1 2 13/3 0%% ∴ 1 x + 2 y + 13/3 z = 0
%
%0 1 5/3 0%% ∴ 0 x + 1 y + 5/3 z = 0
%
%0 0 0 0% ∴ z: free variable ! ∴ z = t

From the second and first equations we calculate

5 5 5
y+ z=0 ∴ y=− z ∴ y=− t
3 3 3
13
x + 2y + z=0
3
∴
10 13
x− t + t = 0 ∴ x + t = 0 ∴ x = −t
3 3

Thus we choose a convenient scale as

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
x −t −3
v#»3 = ⎣y ⎦ = ⎣−5t/3⎦ = ⎣−5⎦ (we choose t = 3)
z t 3

Good practice is to verify results, for example, in case of v#»2 and λ = 2 we calculate,
⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
4 6 10 1 (4 − 12 + 10) 2 1
A v#»2 = ⎣ 3 10 13⎦ ⎣−2⎦ = ⎣ (3 − 20 + 13) ⎦ = ⎣−4⎦ = 2 ⎣−2⎦ = λ2 v#»2
−2 −6 −8 1 (−2 + 12 − 8) 2 1

Exercise 7.10, page 136

1. Calculation of a matrix inverse involves similar assumptions as to the product of a non-zero

number and its inverse that equals one. For a matrix we write

A A−1 = A−1 A = I, if ΔA = 0 (7.11)

First, we verify if determinant equals zero or not, which answers the question of the inverse matrix
existence.
% %
%4 −2%
Δ = %% % = 4 × 3 − 2 × (−2) = 16 = 0 ∴ A−1 exists
2 3%

Technique used to calculate the inverse matrix is based on “Gauss” form of (7.11). First, we start
with the A A−1 part with the objective to use matrix transformations and derive the A−1 A side.

4 −2 1 0 ← (1) ÷ [4]
2 3 0 1

   
A I
184 7 Linear Algebra

1 −1/2 1/4 0
2 3 0 1 ← (2) − 2 × (1)
1 −1/2 1/4 0
0 4 −1/2 1 ← (2) ÷ [4]
1 −1/2 1/4 0 ← (1) + 1/2 × (2)
0 1 −1/8 1/4

10 3/16 1/8

01 −1/8 1/4

     
I A−1

Which is the right side of (7.11), therefore,

3/16 1/8
A−1 = −1/8 1/4

To verify, we calculate

4 −2 3/16 1/8 10
AA−1 = −1/8 1/4
=
2 3 01

Because:

4 −2 3/16 4 −2 (12/16 + 1/4) 1

= 3/16 + −1/8 = =
2 3 −1/8 2 3 (6/16 − 3/8) 0
4 −2 1/8 4 −2 (4/8 − 1/2) 0
= 1/8 + 1/4 = 2 =
2 3 1/4 2 3 ( /8 + 3/4) 1

Exercise 7.11, page 137

1. Calculation of a matrix powers involves several steps. By definition, a square matrix A is

diagonalizable if it can be written as

A = P D P −1 (7.12)

where:

1. P is a matrix that has its inverse, i.e. P P −1 = I exists.

2. D is a diagonal matrix

In this form, for example, we write

)( P D 
P −1
A5 = AAAAA = (P D  )( P D 
P −1 )( P D 
P −1 )( P D P −1 )
P −1
= P DDDDDP −1 = P D 5 P −1
Solutions 185

That is to say, the problem of calculating A5 is replaced with the problem of calculating D 5 and
calculation of P −1 matrix. For higher order powers, this approach is faster relative to the trivial
repeated multiplication. This is because the powers of a diagonal matrix are simple to calculate.
(a) Matrix D is diagonal, thus we write

20 22 0 40
D= ∴ D2 = =
02 0 22 04

(b) Matrix D is diagonal, thus we write

20 25 0 32 0
D= ∴ D5 = =
05 0 55 0 3125

(c) Powers of non-diagonal matrices may be done with the following technique.

1. Calculate determinant to verify if det A = 0?

% %
% 2 0%
det A = %% % = 2 × 3 − (−1) × 0 = 6 = 0
−1 3%

2. Calculate the characteristic polynomial

% % % %
% 20 1 0 %% %%2 − λ 0%%
det(A − λI ) = %% −λ = = (2 − λ)(3 − λ) = 0
−1 3 0 1 % % −1 3 − λ%

Therefore, λ1 = 2 and λ2 = 3.
3. Calculate eigenvectors,

20 10 2−2 0 00
λ1 = 2 : A − λ1 I = 0 ∴ −2 = =
−1 3 01 −1 3 − 2 −1 1
000 ←x=t =1
∴
−1 1 0 ← −x + y = 0 ∴ x = y = 1

∴ v#»1 =
1
1
20 10 2−3 0 −1 0
λ1 = 3 : A − λ1 I = 0 ∴ −3 = =
−1 3 01 −1 3 − 3 −1 0
−1 0 0 ← −x = 0 ∴ x = 0
∴
−1 0 0 ← y = t ∴ y = 1

∴ v#»2 =
0
1
186 7 Linear Algebra

4. Write diagonal matrix, by definition we write

λ1 0 20
D= =
0 λ2 03

1. Calculate P and P −1 , by definition we write

# $
P = v#»1 v#»2 =
10
11

and by using Gauss form we find

1010
PI=
1101 ← (2) − (1)
10 10 10
= = I P −1 ∴ P −1 =
0 1 −1 1 −1 1

(d) Calculate (7.12), as

10 25 0 10 10 32 0 10
A5 = P D 5 P −1 = =
11 0 35 −1 1 11 0 243 −1 1
32 0 10 32 0
= =
32 243 −1 1 −211 243
 Limits
8

Important to Know

Basic rules for calculating limits Infinite limits

 lim x n = +∞
x→+∞
lim C f (x) = C lim f (x)
x→x0 x→x0 lim x n = −∞ n is odd
 x→−∞

lim f (x) ± g(x) = lim f (x) ± lim g(x) lim x n = +∞ n is even

x→x0 x→x0 x→x0 x→−∞

1
lim f (x) g(x) = lim f (x) lim g(x) lim =0
x→x0 x→x0 x→x0 x→±∞ xn

f (x) limx→x0 f (x) lim ln x = −∞
lim = x→0
x→x0 g(x) limx→x0 g(x)
  lim ln x = +∞
n n x→∞
lim f (x) = lim f (x)
x→x0 x→x0 lim ex = 0
x→−∞

lim ex = +∞
x→∞

Some of the “famous” limits

 x
1 ex − 1
lim 1+ =e lim =1
x→±∞ x x→0 x
sin x ln(1 + x)
lim =1 lim =1
x→0 x x→0 x

© Springer Nature Switzerland AG 2021 187

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_8
188 8 Limits

8.1 Exercises

8.1 * Limits

1. Determine the following limits:


3 (x > 2)
(a) lim f (x), f (x) = x (b) lim f (x), f (x) =
x→2 x→2 1 (x < 2)
|x − 2| x 2 − 2x
(c) lim f (x), f (x) = (d) lim f (x), f (x) =
x→2 x−2 x→0 x
√
x 2 + 2x − 15 x−1
(e) lim f (x), f (x) = (f) lim f (x), f (x) =
x→−5 x 2 + 8x + 15 x→1 x−1
8.2 ** Limits: Undefined Functions

1. Determine the following limits:

x 2 + 2x − 3 sin(x)
(a) lim (b) lim f (x), f (x) =
x→1 x−1 x→∞;x→0 x
x + sin(x) sin(2x)
(c) lim f (x), f (x) = (d) lim f (x), f (x) =
x→0 x x→0 x
4 sin(5x) sin(2x)
(e) lim f (x), f (x) = (f) lim f (x), f (x) =
x→0 sin(4x) x→0 sin(3x)
 
1 x+5 1 3x
(g) lim f (x), f (x) = 1 + (h) lim f (x), f (x) = 1 +
x→∞ x x→∞ x
 
2 x x x
(i) lim f (x), f (x) = 1 + (j) lim f (x), f (x) =
x→∞ x x→∞ 1+x

x + 3 x+1
(k) lim f (x), f (x) =
x→∞ x−1
8.3 *** Asymptotes

1. Calculate limits of the following functions:

√
x−1 x+1−3 x3
(a) f (x) = (b) f (x) = (c) f (x) =
x 2 − 5x + 6 x−8 x2 − 4
Solutions 189

Solutions

Exercise 8.1, page 188

1. There are various possibilities that determine whether the limiting value exists or not.
(a) A continuous function is defined in all points, including for x = 2, and therefore calculating
any limiting value is trivial. For example,

lim f (x) = lim x = 2

x→2 x→2

because both left side (i.e. limx→2− ) and right side (i.e. limx→2+ ) limits point to the same value of
function, i.e. f (2) = 2, Fig. 8.1.

Fig. 8.1 Example 8.1-1(a)

f (x)
limx→2+ f (x)
2
limx→2− f (x)

0 x

0 2

(b) A non-continuous function may or may not have defined limit. For example, function f (x)
is not defined for x = 2. When calculating limits, we find

lim f (x) = 1
x→2−

lim f (x) = 3
x→2+

For the reason that the left side and right side limits are not equal, consequently, limit limx→2 f (x)
does not exist, Fig. 8.2.
190 8 Limits

Fig. 8.2 Example 8.1-1(b)

f (x)
3
limx→2+ f (x)

1
limx→2− f (x)
0 x

0 2

(c) Similarly, function

|x − 2|
f (x) =
x−2

is not defined for x = 2, it equals 0/0. Consequently, the first limitation is x = 2. We analyse
separately left and right sides of x = 2 point, Fig. 8.3.

1. if x > 2, then

|x − 2| = x − 2
∴ (x = 2)
x−
 2
lim  =1
x−2
x→2+ 

2. if x < 2, then

|x − 2| = −(x − 2)
∴ (x = 2)
 
−(x
 − 2)
lim  = −1
x→2− x−
 2
Solutions 191

Therefore, for the reason that the left and right side limits are not equal, the conclusion is
that limx→2 f (x) does not exist.

Fig. 8.3 Example 8.1-1(c)

limx→2+ f (x)
1

0 x

−1
limx→2− f (x)

0 2

(d) For a special case of rational function, Fig. 8.4, after excluding division by zero, i.e. x = 0,
we calculate the limit as

x 2 − 2x
lim f (x) =
x→0 x
x(x − 2)
= lim

x→0 x

= −2

which is equal from both sides of x = 0. In conclusion, even though x = 0, the limit exists in that
point because both left and right side limits are equal (search for “Two Policemen and a Drunk”
or “Sandwich” theorem).

Fig. 8.4 Example 8.1-1(d)

f (x)
0 x

limx→0+ f (x)
−2
limx→0− f (x)

0 2
192 8 Limits

(e) Similar example of rational function where x = −5 also permits the existence of limit as

x 2 + 2x − 15 x 2 − 3x + 5x − 15
lim f (x) = lim = lim 2
x→−5 x→−5 x + 8x + 15
2 x→−5 x + 3x + 5x + 15

x(x − 3 + 5(x − 3)
= lim
(x = 0)

+ 3) + 5(x + 3)
x→−5 x(x

+
(x
(x − 3) 
5)
= lim 
 (x = −5)
(x
x→−5 (x + 3) + 5)
−5 − 3
= =4
−5 + 3

In conclusion, even though the function is not defined for x = −5, Fig. 8.5, limit limx→−5 f (x) =
4. This case is also known as “pole–zero cancellation” because both numerator and denominator
have one equal root (i.e. x = −5) that is cancelled, and therefore value of the function for x = −5
is neither zero nor infinity. We note that is not the case for x = −3 where the limit does not exist
because

lim f (x) = +∞ and lim f (x) = −∞

x→−3− x→−3+

that is to say, they are not equal, and thus the limit does not exist.

Fig. 8.5 Example 8.1-1(e)

f (x)

limx→−5− f (x) limx→−5+ f (x)

4

0 x

−5 −3 0
Solutions 193

(f) Rational functions that contain radicals may also take advantage of pole–zero cancellation
(and, with the help of a 2 − b2 = (a − b)(a + b) identity) as
√
x−1
lim f (x) = lim (x ≥ 0)
x→1 x→1 x − 1
√  
x− 1
= lim √  √ (x = 1)
( 
x→1  1 − 1)( 1 + 1)
1 1
= lim √ =
x→1 1+1 2

In conclusion, even though the function is not defined for x = 1, due to pole–zero cancellation
the limit exists, Fig. 8.6.

Fig. 8.6 Example 8.1-1(f)

f (x)
1

limx→1+ f (x)
limx→1− f (x)

x≥0
0 x
0 1

Exercise 8.2, page 188

1. Limits based on number e and sinc function are important.

(a) Knowing that sin function is limited to ±1, we can deduce that
sin(x) (≤ 1)
lim sinc(x) = lim = =0
x→∞ x→∞ x ∞

 1
(b) x + sin(x) >


sin(x)
lim = lim 1+  =2
x→0 x x→0 x
*
 1
(c) sin(2x)
= lim
2 sin(2x)
= 2 lim
sin(2x)
(t = 2x) = 2 lim 
sin t
=2
lim
x→0 2  
x→0 x x x→0 2x x→0 t
1
5x 
>
sin(5x)
4 sin(5x) sin(5x) 
(d) lim = lim 5x = lim 5x
 5x =5
x→0 sin(4x) x→0 sin(4x) x x→0 x 1

sin(4x)>

4 x 

 4x
194 8 Limits

⎧ ⎫ 1
⎪ x = 0, ⎪ 
sin(t)
2x ⎪
⎨ ⎪
⎬
sin(2x) sin(2x) 2x 2
(e) lim = lim 2x
= t = 2x, = lim
t =
x→0 sin(3x) x→0 3x
sin(3x) ⎪
⎪ ⎪
⎪ x→0 3x


1 3
3x ⎩ ⎭ 
p = 3x sin(p)
p
(f) Forms that can be forced into limit of e,
 !  "
1 x+5 1 x 1 5
lim 1 + = lim 1+ 1+
x→∞ x x→∞ x x
  5
1 x 1
= lim 1 + lim 1 +
x→∞ x x→∞ x
 x e
0
1
: 1
> 5
= lim 
 1 + x
x→∞
1 + lim


x→∞ x

=e
  
1
:
3x x 3 x e 3
1 1
(g) lim 1+ = lim 1+ = lim
  
1 + = e3
x→∞ x x→∞ x 
x→∞ x
(h) Similarly, a non-one numerator, for example, “2” in this example, may be temporarily
“hidden” into another variable so that
⎧ ⎫
 ⎪
⎨ 2 1 ⎪
⎬ 
2 x = ∴ x = 2t, 1 2t
lim 1 + = x t = lim 1 +
x→∞ x ⎩x → ∞ ∴ t → ∞ ⎪
⎪ ⎭ t→∞ t

⎡ e ⎤2
 
*
⎢ 1 t
⎥
= ⎣ lim 1 + ⎦ =e
2
t→∞ t


(i) Sometimes, rational forms may also be forced into

⎛ ⎞x
 1
x x
⎜ 1 ⎟ 1
>
x
= lim ⎝ = lim 
1+x ⎠
lim
x→∞ 1 + x x→∞ x→∞ 1+x x
x x
1 1
=  :
x e =
1+ x e

x→∞x
lim
Solutions 195

(j) Or, the combination of the above cases

   
x + 3 x+1 x − 1 = t ∴ x = t + 1, t +4 t+2
lim = = lim
x→∞ x − 1 x→∞ ∴ t →∞ t→∞ t

*1


  0
+ 4 )2
t
4
= lim 1 + (1
lim
t→∞ t 
t→∞ t
⎧ ⎫
⎪
⎨ =
4 1
∴ t = 4p, ⎪⎬  4p
t p 1
= = lim 1 +
⎪
⎩ ⎪
⎭ p→∞ p
t →∞ ∴ p→∞
!  "
p e 4
:

1
= lim 

p→∞1 + p = e4

Exercise 8.3, page 188

1. Limits are systematically found around the domain extremes and the breaking points (i.e. vertical
asymptotes).
(a) Factorize f (x)
x−1 x−1 x−1
f (x) = 2 = 2 =
x − 5x + 6 x + 2x + 3x + 6 x(x + 2) + 3(x + 2)
x−1 P (x)
= =
(x + 2)(x + 3) Q(x)
Vertical asymptotes are found as roots of Q(x) = 0:
i.e. (x + 2)(x + 3) = 0 ∴ x1 = −2, x2 = −3
The only zero of f (x) is when P (x) = 0 ∴ x − 1 − 0 ∴ x = 1, which is different from
the two poles, and therefore there is no pole–zero cancellation. Two vertical asymptotes define,
in total, six limits to be calculated, see Fig. 8.7.

Fig. 8.7 Example 8.3-1(a)

x2 x1
0 x

1 3 4 5 6 2

−3 −2 0
196 8 Limits

0
1
x−1
− x )
x(1 1 1
lim = lim  = lim ==
1 x→−∞ x 2 + 5x + 6 x→−∞ 0 x→−∞ x−5 −∞
6

+ 5 + x )
x(x

= 0− (i.e. below the horizontal axis)
0
1
x−1
− x )
x(1 1 1
lim = lim  = lim ==
2 x→+∞ x 2 + 5x + 6 x→+∞ 0 x→+∞ x − 5 +∞
6

+ 5 + x )
x(x

= 0+ (i.e. above the horizontal axis)
x−1 −3 − 1 −4
lim = lim = lim
x→−3− (x + 2)(x + 3) x→−3− (−3 + 2)(x + 3) x→−3− −(x + 3)
 
3 = if x < −3 ∴ (x + 3) < 0
−4
= = −∞
+0
x−1 −3 − 1 −4
lim = lim = lim
x→−3+ (x + 2)(x + 3) x→−3+ (−3 + 2)(x + 3) x→−3+ −(x + 3)
 
(a) 4 = if x > −3 ∴ (x + 3) > 0
−4
= +∞
=
−0
x−1 −2 − 1 −3
lim = lim = lim
x→−2− (x + 2)(x + 3) x→−2− (x + 2)(−2 + 3) x→−2− (x + 2)
 
5 = if x < −2 ∴ (x + 2) < 0
−3
= = +∞
−0
x−1 −2 − 1 −3
lim = lim = lim
x→−2+ (x + 2)(x + 3) x→−2+ (x + 2)(−2 + 3) x→−2+ (x + 2)
 
6 = if x > −2 ∴ (x + 2) > 0
−3
= = −∞
+0
Plot of this function shows (not to y-scale), see Fig. 8.8, that knowing limits we can already
deduce the shape of the function (the complete function f (x) is plot in the background).
Solutions 197

Fig. 8.8 Example 8.3-1(a)

4 5 f (x)

x2 x1 2
0
1

3 6

−3−2

√
x+1−3 P (x)
(b) f (x) = =
x−8 Q(x)
Domain is defined for x + 1 ≥ 0 ∴ x ≥ −1. Furthermore, zero is found when P (x) = 0 ∴
√
x + 1 − 3 = 0 ∴ x = 8. However, vertical asymptote is found at Q(x) = 0 ∴ x − 8 =
0 ∴ x = 8. The conclusion is that the only pole–zero pair is cancelled, that is to say f (8)
equals neither zero nor infinity.
One break point x = 8 defines four limits to be calculated, see Fig. 8.9.

Fig. 8.9 Example 8.3-1(a)(b)

f (x)

1 2 3 4

−1 0 8

√ √
x+1−3 −1 + 1 − 3 −3 1
1 lim = = =
x→−1+
√x − 8 √ −1 − 8 −9 3
x+1−3 x+1 (x + 1) − 9
f (x) = √ = √  (x = 8)
x−8 x+1 (x − 8) x + 1 + 3
2 
x−
  8
= 
 √ 

(x 
− 8) x +1+3
198 8 Limits

Therefore,
1 1 1
lim √ =√ = {x ≥ −1} =
x→8− x+1+3 8+1+3 6
1 1 1
3 lim √ =√ = {x ≥ −1} =
x→8+ x+1+3 8+1+3 6
1 1
lim √ =√ = 0+
x→+∞ x + 1 + 3 +∞ + 1 + 3
4
(i.e. above the horizontal axes)

(c) Vertical asymptotes are found as roots of Q(x) = 0:

Q(x) = 0 ∴ x 2 − 4 = 0 ∴ x1 = −2, x2 = 2

Third order polynomial must have at least one real root; thus, real zeros of f (x) are found as
roots of P (x) = 0

P (x) = 0 ∴ x 3 = x x x = 0 ∴ x1,2,3 = 0

That is to say, all three roots of P (x) are real and identical (i.e. “triple zero”). The zeros and
poles of f (x) are not equal, and thus there are no pole–zero cancellations. These two vertical
asymptotes determine the following six limits to calculate:

x3 x2 x
1 lim = lim = −∞
x→−∞ x 2 − 4 x→−∞ 0
x 2 (1 − x47 )
3

3
2

x x (−2)3
lim 2 = lim = lim
x→−2− x − 4 x→−2− (x − 2)(x + 2) x→−2− (−2 − 2)(x + 2)

−8  
2 = lim = if x < −2 ∴ (x + 2) < 0
x→−2− −4(x + 2)

−8
= = −∞
−4(−0)
3
x x3 (−2)3
lim 2 = lim = lim
x→−2+ x − 4 x→−2+ (x − 2)(x + 2) x→−2+ (−2 − 2)(x + 2)

−8  
3 = lim = if x > −2 ∴ (x + 2) > 0
x→−2+ −4(x + 2)

−8
= = +∞
−4(+0)
3
x x3 (2)3
lim 2 = lim = lim
x→2− x − 4 x→2− (x − 2)(x + 2) x→2− (x − 2)(2 + 2)

8  
4 = lim = if x < 2 ∴ (x − 2) < 0
x→2− 4(x − 2)

8
= = −∞
4(−0)
Solutions 199

x3 x3 (2)3
lim = lim = lim
x→2+ x2− 4 x→2+ (x − 2)(x + 2) x→2+ (x + 2)(2 + 2)
8  
5 = lim = if x > 2 ∴ (x − 2) > 0
x→2+ 4(x + 2)

8
= = +∞
4(+0)
x3 x2 x
6 lim = lim =∞
x→∞ x 2 − 4 x→∞ 0
4
7
x (1 −
2
)
x 2
(d) In addition, we search for oblique asymptote(s) yaa whose linear equation is found in form

f (x)
yaa = ax + b where, a = lim and b = lim (f (x) − ax)
x→∞ x x→∞

Thus, we calculate

x3
x2 − 4 = lim x3 x3
a = lim = lim =1
x→∞ x x→∞ x(x − 4)
2 x→∞ 0
x 3 (1 − x47
)
 2
x3 x 3 − x(x 2 − 4) x3 −x3 + 4x
b = lim −x = lim = lim
x→∞ x2 − 4 x→∞ x2 − 4 x→∞ x2 − 4
4x
4
= lim 0
= lim =0
x→∞ x→∞ x
47

−
x(x x
)

Therefore, there is one oblique asymptote yaa = ax + b = x. Knowledge of the function’s limits
is sufficient to sketch its form, see Fig. 8.10.

Fig. 8.10 Example 8.3-1(c)

f (x) 3
5 6

0 x
a
ya

1 2 4
−2 0 2
 Derivatives
9

Important to Know

Basic tabular derivatives

f (x) f  (x)
a = const. 0
xn n x n−1
√ 1 −1/2
x = x 1/2 x
2
1
ln x (x = 0)
x
ex ex
ax a x ln a
|x| sign (t) (x = 0)
sin x cos x
cos x − sin x
1
tan x = 1 + tan2 x
cos2 x
1
arcsin x √
1 − x2
1
arccos x −√
1 − x2
1
arctan x
1 + x2

© Springer Nature Switzerland AG 2021 201

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_9
202 9 Derivatives

Basic rules of derivations


C f (x) = C f  (x)


f (x) ± g(x) = f  (x) ± g  (x)


f (x) g(x) = f  (x) g(x) + f (x) g  (x)


f (x) f  (x) g(x) − f (x) g  (x)
=
g(x) g 2 (x)

f (g(x)) = f  (g(x)) g  (x)

9.1 Exercises

9.1 Tabular Derivatives: Power Rule (kx n ) = k nx n−1

1. Calculate derivatives of the following functions.

1 1
(a) f (x) = x 5 (b) h(t) = x 3 + 2x + 1 (c) g(x) = − x 3 + x 3
27 18
4 1 1 4 3
(d) h(x) = − x −3 (e) a(x) = x2 (f) h(x) = x − 4
x3 2 3
2. Calculate derivatives of the following functions.
√ √ √ √
(a) f (x) = x (b) h(x) = x 2
3
g(x) = 5 4 x − 2 x 3
3
(c)
√
5 1 1 t
(d) b(x) = √ 4
(e) t (x) = √ 3
(f) g(t) = t 3 − +2
x x4 3 3

9.2 Tabular Derivatives: Exponent (a x ) = a x ln a

1. Calculate derivatives of the following functions.

(a) f (x) = 1x (b) h(x) = 2x (c) g(x) = ex
(d) b(x) = 10x (e) g(x) = a x + x a (f) f (x) = ex + x 3

 
9.3 * Tabular Derivatives: Logarithm loga x = 1
x ln a

1. Calculate derivatives of the following functions.

(a) f (x) = ln x (b) h(x) = log x (c) g(x) = log5 x
9.1 Exercises 203

9.4 ** Composite Functions: Chain Rule

Composite functions follow chain rule of derivation:

[h(x)] = [f (g(x))] = f  (g(x)) g  (x)

That is to say, first we calculate derivative of the “external” function f (x), then we multiply it with
derivative of its argument function g(x).

1. Calculate derivatives of simple composite functions.

 
(a) f (x) = e5x (b) h(x) = sin x 2 (c) g(x) = sin(abcx)

2. Calculate derivatives of composite functions by using the power rule.

 √
(a) f (x) = (1 − x)5 (b) h(x) = x 2 + 2 (c) g(x) = 3
2x − 5

3. Calculate derivatives of composite functions by using the exponent rule.

2
+2x−1 2
(a) f (x) = e1−x (b) h(x) = 2x (c) h(x) = ex
√ 1 1
g(x) = e6x−3 g(x) = g(x) =
3
(d) (e) (f) √
e−x 2 /2 e 2−3x

4. Calculate derivatives of composite functions by using the logarithmic rule.

 
(a) f (x) = ln(1 − x) (b) h(x) = log2 x 2 + 2x + 1
√ 1
(c) g(x) = ln 2−x (d) g(t) =
log(2t − 3)
5. Calculate derivatives of composite functions that include trigonometric functions.
√
(a) f (x) = sin(2x − 1) (b) h(x) = sin x
1
(c) g(x) = cos(sin(1 − x)) (d) g(x) =
ln(sin x)
   
1
(e) g(x) = sin (f) g(x) = ln sin x + 1 + sin2 x
x
 
2 2π 2 2π
(g) g(x) = sin x + sin
2
+ x + sin −x
3 3

9.5 ** Product Rule (f (x)g(x)) = f  (x)g(x) + f (x)g  (x)

1. Calculate derivatives by using the product rule.

(a) h(x) = x sin x (b) g(x) = ex cos x
√
(c) f (x) = (2x − 1)(1 − x) (d) f (x) = x ln x
(e) f (x) = sin(x − 1) cos(1 − x) (f) f (x) = log(x − 1) cos(1 − x)
(g) h(t) = t 2 3t
204 9 Derivatives

 
f (x) f  (x)g(x)−f (x)g  (x)
9.6 ** Ratio Rule g(x) = g 2 (x)

1. Calculate derivatives by using the ratio rule for the following functions in points where the
derivatives exist.
x sin x 2x − 1
(a) h(x) = (b) f (x) = (c) f (x) =
sin x ex 1−x
√
x sin(x − 1) log(x − 1)
(d) f (x) = (e) f (x) = (f) f (x) =
ln x cos(1 − x) cos(1 − x)

9.7 * Taylor Polynomial

1. Develop the following functions into its equivalent Tylor polynomials, calculate only the first four
terms of the polynomials.
(a) f (x) = ex , x0 = 0 (b) g(x) = cos x, x0 = 0
√
(c) f (x) = sin x, x0 = 0 (d) f (x) = x, x0 = 1
2
(e) f (x) = ln(1 + x), x0 = 0 (f) f (x) = e2x−x , x0 = 0
x
(g) f (x) = x−1 , x0 = 1
e

9.8 * L’Hôpital’s Rule

1. Evaluate the following limits:

x 2 + 2x − 3 sin(x) x + sin(x)
(a) lim (b) lim (c) lim
x→1 x−1 x→0 x x→0 x
sin(2x) 4 sin(5x) sin(2x)
(d) lim (e) lim (f) lim
x→0 x x→0 sin(4x) x→0 sin(3x)

2. Evaluate the following limits:

ln(sin x)
(a) lim x x (b) lim x 1/x (c) lim
x→0 x→∞ x→π/2 cos x

lim (sin x)tan x 1 1
(d) x→π/2
(e) lim − (f) lim x ln x
x→0
x→0 x sin x

lim x ex 1 x+5 ex
(g) x→−∞
(h) lim 1 + (i) lim
x→∞ x x→∞ x 2
Solutions 205

Solutions

Exercise 9.1, page 202

1. The power rule results in the following answers.

 
(a) f  (x) = x 5 = 5x 5−1 = 5x 4
 
(b) h (t) = x 3 + 2x + 1 = 3x 2 + 2
 
(c) g (x) = − x 3 + x 3 = −
x 2 +
x 2 = 0
 1 1 31 21
27 18 
279 

189
 
4    9
(d) h (x) = − x −3 = 4x −3 − x −3 = −12x −4 + 3x −4 = − 4
x3 x

1 1  1 1 ( 1 −1) 1 1 1
(e) a  (x) = = x (− 2 ) =
1
x2 = x 2 1 = √
2 22 4 4x 2 4 x
  
4 −3 4
3
1 1
(f) h (x) = x − 4 −1 = −x − 4 = − 7 = − √
3 7
x 4 = −
3 3
4
x4
4
x7
2. Radicals are equivalent to fractional powers.
√   1  1 1 1 1 1
(a) f  (x) = x = x 2 = x 2 −1 = x − 2 = √
2 2 2 x
√   2  2 2 2 2
(b) h (x) = x 2 = x 3 = x 3 −1 = x − 3 =
3 2 1
1 = √
3 3 3x 3 33x
 √ √   1 
3  5 3 5
(c) g  (x) = 5 4 x − 2 x 3 = 5x 4 − 2x
3 = x − 4 − 2 = √
3
4
−2
4 4 x3
  
5  5   − 1  1 5
(d) b (x) = √ x− 4 = − √
5
= 1 = 5x 4 = 5 − 4
4
x x4 4 4 x5
   
1 4  4 7 4
(e) t  (x) = √ 3
= x− 3 = − x− 3 = − √ 3
x 4 3 3 x7
 √ 
1 3 t 1
(f) g  (t) = t − + 2 = t2 − √
3 3 6 t

Exercise 9.2, page 202

1. Derivatives of a x functions include ln a term.

(a) f  (x) = 1x  =  *0 = 0
1
1
>
x 
ln(1)

 
(b) h (x) = 2x = 2x ln 2
  1
(c) g  (x) = ex = ex  ln 
*
e = ex
 
(d) b (x) = 10x = 10x ln 10
 
(e) g  (x) = a x + x a = a x ln a + ax a−1
(f) f  (x) = (ex + x 3 ) = (ex ) + (x 3 ) = ex + 3x 2
206 9 Derivatives

Exercise 9.3, page 202

1. Derivatives of loga x functions include ln a term.

1 1
(a) f  (x) = (ln x) = =
ln *
 1 x
x e
1
(b) h (x) = (log x) =
x ln 10
   1
(c) g  (x) = log5 x =
x ln 5

Exercise 9.4, page 203

1. Derivatives of composite functions follow the “chain rule”. First, we use derivative of the
“external” function and keep its argument as is, then we multiply it with derivative of the argument
function.
# $
(a) f  (x) = e5x = e5x (5x) = 5e5x
#  $      
(b) h (x) = sin x 2 = cos x 2 x 2 = 2x cos x 2
(c) g  (x) = [sin(abcx)] = cos(abcx) (abcx) = abc cos(abcx)

2. Derivatives of composite functions follow the “chain rule”.

# $
(a) f  (x) = (1 − x)5 = 5 (1 − x)5−1 (1 − x) = 5 (1 − x)4 (−1) = −5 (1 − x)4

   1/2  1 2 −1/2  2 
(b) h (x) = x2 + 2 = x2 + 2 = x +2 x +2 
2
2
x x
= √ =√
2
x 2 + 2 x2 + 2

     
√ 1 1  1 
g  (x) =
3
(c) 2x − 5 = (2x − 5) 2 3 = (2x − 5) 6
1 2
1 1
(2x − 5)− 6 (2x − 5) = (2x − 5)− 6 = 
5 5
=
6 6
3 3 (2x − 5)5
6
Solutions 207

3. Derivatives of composite exponential functions. First we use the exponent rule, then multiply it
with the derivative of its argument.
  1
(a) f  (x) = e1−x = e1−x  ln 
*e (1 − x) = −e1−x
 2 
h (x) = 2x +2x−1 = 2x +2x−1 ln(2) (x 2 + 2x − 1) = ln(2) (2x + 2) 2x +2x−1
2 2
(b)
 2  1  2 
h (x) = ex = ex  ln 
*
2 2
(c) e x = 2x ex
√   1/3   2x−1  1
g  (x) = ln 
*
3
(d) e6x−3 = e6x−3 = e = 2 e2x−1  e
   2 
1
g  (x) =
2
(e) = ex /2 = x ex /2
e−x /2
2

   √   
1 1/2 
(f) g  (x) = √ = e− 2−3x = e−(2−3x)
e 2−3x
1 1# $
= √ ln 
 *e −(2 − 3x)1/2
e 2−3x

1 1
= √ − (2 − 3x)−1/2 (2 − 3x)
e 2−3x 2
3
= √ √
2e 2−3x (2 − 3x)
4. Derivatives of composite loga x functions include ln a term, and derivative of its argument.
1 1 1
(a) f  (x) = (ln(1 − x)) = ln 
 *
e (1 − x) = −
1−x 1−x

   1
(b) h (x) = log2 x 2 + 2x + 1 = (x 2 + 2x + 1)
+ 2x + 1) ln(2)
(x 2
2x + 2 2(x+
1) 2
= = =
ln(2) (x + 2x + 1)
2
ln(2) (x + 1)2 ln(2) (x + 1)

 √  1 √ 
(c) g  (x) = ln 2 − x = √ 2−x
ln 
* 1
2−x  e
1 # $
1/2  1
=√ (2 − x) =√ (2 − x)−1/2 (2 − x)
2−x 2−x
1 1
= √ √ (−1) = −
2 2−x 2−x 2 (2 − x)

 #
 1 
$−1 
(d) g (t) = = log(2t − 3)
log(2t − 3)
# $−2
= −1 log(2t − 3) (log(2t − 3))
1 1
=− (2t − 3)
log (2t − 3) (2t − 3) ln 10
2

2 1
=−
ln 10 log (2t − 3)
2 (2t − 3)
208 9 Derivatives

5. Derivatives of composite trigonometric functions include derivative of its argument.

(a) f  (x) = (sin(2x − 1)) = cos(2x − 1) (2x − 1) = 2 cos(2x − 1)
√  √ √ √ √ 1
(b) h (x) = (sinx) = cos( x) ( x) = cos( x) (x 1/2 ) = cos( x) x 1−1/2
√ 2
√ 1 −1/2 cos( x)
= cos( x) x = √
2 2 x

(c) g  (x) = [cos(sin(1 − x))] = − sin(sin(1 − x)) [sin(1 − x)]

= − sin(sin(1 − x)) cos(1 − x) (1 − x) = sin(sin(1 − x)) cos(1 − x)
 
1 # $
(d) 
g (x) = = (ln(sin x))−1 = (−1) (ln(sin x))−2 (ln(sin x))
ln(sin x)
1 1 1 cos x cot x
=− 2 (sin x) = − 2 =− 2
ln 
* 1
ln (sin x) sin x  e ln (sin x) sin x ln (sin x)
     
1 1 1
(e) g  (x) = sin = cos
x x x  
 
1  −1  1 −2
cos x1
= cos x = cos (−1) x = −
x x x2
   
(f) g  (x) = ln sin x + 1 + sin2 x
1   
=  sin x + 1 + sin2 x
sin x + 1 + sin2 x
 
1 2
sin x cos x
=  cos x + 
sin x + 1 + sin2 x 2
1 + sin2 x
 ((( (
cos x ( (2(
1(+(sin x + sin x cos x
=  ((( (  =
x(
sin( (
+( 1 + sin x
2 1 + sin x
2
1 + sin2 x
(

(g)
  
 2π 2 2π
g (x) = sin x + sin
2 2
+ x + sin −x
 3 3
sin(x ± y) = sin x cos y ± cos x sin x
!  2
2π 2π
= sin x + sin
2
cos x + cos sin x
3 3

 "
2 
2π 2π
+ sin cos x − cos sin x
3 3
Solutions 209

⎡ √ 2  √ 2 ⎤
3 1 3 1
= ⎣sin2 x + cos x − sin x + cos x + sin x ⎦
2 2 2 2

√  √ 
3 1 3 1
= 2 sin x cos x + 2 cos x − sin x cos x − sin x
2 2 2 2

√  √ 
3 1 3 1
+2 cos x + sin x cos x + sin x
2 2 2 2

11 √ √
= 2 sin x cos x + 2
( 3 cos x − sin x)(− 3 sin x − cos x)
2
2

11 √ √
+ 2
( 3 cos x + sin x)(− 3 sin x + cos x)
2
2

1 √ 2 √ 2 
= 2 sin x cos x + (−3 cos x sin x + 3sin x − 3cos x + sin x cos x
2
√ 2 √ 2 
− 3 cos x sin x − 3sin x + 3cos x + sin x cos x)
1
= 2 sin x cos x + (−4 sin x cos x) = 0
2

Exercise 9.5, page 203

1. Derivative of two functions product follows formula

(f (x)g(x)) = f  (x)g(x) + f (x)g  (x)

(a) h (x) = (x sin x) = x  sin x + x(sin x) = sin x + x cos x

   
(b) g  (x) = ex cos x = ex cos x + ex (cos x) = ex cos x − ex sin x
= ex (cos x − sin x)
(c) f  (x) = [(2x − 1)(1 − x)] = (2x − 1) (1 − x) + (2x − 1)(1 − x)
= 2(1 − x) − (2x − 1) = 2 − 2x − 2x + 1 = −4x + 3
√

√  ln x x ln x 1
(d) f (x) = x ln x = √ + = √ +√
2 x x 2 x x
(e) f  (x) = [sin(x − 1) cos(1 − x)]
= [sin(x − 1)] cos(1 − x) + sin(x − 1) [cos(1 − x)]
= cos(x − 1)(x − 1) cos(1 − x) − sin(x − 1) sin(1 − x)(1 − x)
= cos(x − 1) cos(1 − x) + sin(x − 1) sin(1 − x)
 
cos(x − y) = cos x cos y + sin x sin y
:

sin
= cos(2x − 2)+ 0
0
210 9 Derivatives

# $
(f) f  (x) = log(x − 1) cos(1 − x)
1
= cos(1 − x) + log(x − 1) sin(1 − x)
(x − 1) ln 10
(g) h (t) = (t 2 3t ) = (t 2 ) 3t + t 2 (3t ) = 2t 3t + t 2 3t ln 3 = 3t t (2 + t ln 3)

Exercise 9.6, page 204

1. Derivative of two functions ratio follows formula

f  (x)g(x) − f (x)g  (x)

(f (x)g(x)) =
g 2 (x)
 x  sin x − x cos x
(a) h (x) = =
sin x sin2 x

sin x  cos x ex − sin x ex cos x − sin x cos x − sin x
(b) f  (x) = = = ex =
ex e2x e2x ex
 
2x − 1 2(1 − x) + (2x − 1) 1
(c) f  (x) = = =
1−x (1 − x) 2 (1 − x)2
√
√  ln√x
− xx ln√x
− √1x
(d)  x 2 2 2 2 ln x − 2
f (x) = = 2
= 2
= √ 2
ln x ln x ln x 2 2 ln x
sin(x − 1)  cos(x − 1) cos(1 − x) − sin(x − 1) sin(1 − x) (−1)
(e) f  (x) = =
 cos(1 − x) cos (1 − x)
2

cos(−x) = cos x and sin(−x) = − sin(x)

cos(x − 1) cos(x − 1) + sin(x − 1) sin(x − 1)
=
cos2 (1 − x)
cos2 (x − 1) + sin2 (x − 1) 1
= =
cos2 (1 − x) cos2 (1 − x)

log(x − 1)  + log(x − 1) sin(1 − x)(−1)
cos(1−x)
 (x−1) ln 10
(f) f (x) = =
cos(1 − x) cos2 (1 − x)
cos(1 − x) − ln 10 log(x − 1) sin(1 − x) (x − 1)
=
ln 10 (x − 1) cos2 (1 − x)
 
ln a
log a =
ln 10
cos(1 − x) − ln(x − 1) sin(1 − x) (x − 1)
=
ln 10 (x − 1) cos2 (1 − x)

Exercise 9.7, page 204

1. Tylor formula gives an approximation of a function f (x) in the form of polynomial of nth order
as


n
f (k) (x0 )
f (x) = (x − x0 )k + O(x − x0 )n
k=0
k!
Solutions 211

where f (k) (x0 ) is k th derivative in x = x0 , and O(x − x0 )n is the reminder term so that
limx→x0 O(x − x0 )n = 0.
(a) First four terms of Tylor polynomial are found as

f (x) = ex , x0 = 0 ∴ f (0) = f  (0) = f  (0) = f  (0) = e0 = 1

∴
f (0) (0) f  (0) f  (0) f  (0)
T (x) = (x − 0)0 + (x − 0)1 + (x − 0)2 + (x − 0)3
0! 1! 2! 3!
x2 x3
=1+x+ +
2 6

Around the point x = 0, obviously, f (x) = T (x) = 1 and difference is relatively small in
within [−1, 1] interval, while further away T (x) becomes less accurate approximation of f (x),
i.e. Δ = f (x) − T (x) becomes large, see Fig. 9.1.

Fig. 9.1 Example 9.7-1(a)

f (x)

x
0
ex
T (x)
ex − T (x)
0

(b) Trigonometric functions are periodic, however, polynomial approximation within an interval
inferior to the period of an even function is
g(x) = cos x, x0 = 0 ∴
f (0) = cos 0 = 1, f  (0) = − sin 0 = 0,
f  (0) = − cos 0 = −1, f  (0) = sin 0 = 0,
therefore,

f (0) (0) f  (0) f  (0) f  (0)

T (x) = (x − 0)0 + (x − 0)1 + (x − 0)2 + (x − 0)3
0! 1! 2! 3!
x2
=1−
2

Effectively, there are only two even order terms in T (x) (odd order terms are cancelled) that
reasonably well
212 9 Derivatives

(c) approximate cos x within [−1, 1] interval, see Fig. 9.2.

Fig. 9.2 Example 9.7-1(b)
cos x
T (x)
1 cos x − T (x)

0
(d) Approximation of an odd function is
f (x) = sin x, x0 = 0 ∴
f (0) = sin 0 = 0, f  (0) = cos 0 = 1,
f  (0) = − sin 0 = 0, f  (0) = − cos 0 = −1,
therefore,

f (0) (0) f  (0) f  (0) f  (0)

T (x) = (x − 0)0 + (x − 0)1 + (x − 0)2 + (x − 0)3
0! 1! 2! 3!
x3
=x−
6
Effectively, there are only two odd order terms in T (x) (even order terms are cancelled) that
reasonably well approximate sin x within [−1, 1] interval, see Fig. 9.3.

Fig. 9.3 Example 9.7-1(d)

sin x f (x)
1 T (x)
sin x − T (x)

0
x

−1

0
Solutions 213

(e) We systematically calculate derivatives at x0 = 1 as

√
f (x) = x ∴ f (1) = 1
1 1
f  (x) = √ ∴ f  (1) =
2 x 2
1 1
f  (x) = − √ ∴ f  (1) = −
4 x3 4
3 3
f  (x) = √ ∴ f  (1) =
8 x 5 8

Therefore,

f (0) (1) f  (1) f  (1) f  (1)

T (x) = (x − 1)0 + (x − 1)1 + (x − 1)2 + (x − 1)3
0! 1! 2! 3!
1 1 3
= 1 + (x − 1) − (x − 1)2 + (x − 1)3
2 4 × 2! 8 × 3!
x − 1 (x − 1)2 (x − 1)3
=1+ − +
2 8 16

Comparison between f (x) and T (x) is shown in Fig. 9.4.

Fig. 9.4 Example 9.7-1(e) √

x
T (x)
√
x − T (x)
1

0
x

0 1
214 9 Derivatives

(f) f (x) = ln(1 + x), x0 = 0

f (x) = ln(1 + x) ∴ f (0) = 0

1
f  (x) = ∴ f  (0) = 1
1+x
1
f  (x) = − ∴ f  (0) = −1
(1 + x)2
2
f  (x) = ∴ f  (0) = 2
(1 + x)3

Therefore,

f (0) (0) f  (0) f  (0) f  (0)

T (x) = (x − 0)0 + (x − 0)1 + (x − 0)2 + (x − 0)3
0! 1! 2! 3!
1 1
= x − x2 + x3
2 3

2
(g) f (x) = e2x−x , x0 = 0

2
f (x) = e2x−x ∴ f (0) = 1

f  (x) = e2x−x (2 − 2x) ∴ f  (0) = 2

2

f  (x) = e2x−x (4x 2 − 8x + 2) ∴ f  (0) = 2

2

f  (x) = e2x−x (−8x 3 + 24x 2 − 12x − 4) ∴ f  (0) = −4

2

Therefore,

f (0) (0) f  (0) f  (0) f  (0)

T (x) = (x − 0)0 + (x − 0)1 + (x − 0)2 + (x − 0)3
0! 1! 2! 3!
2
= 1 + 2x + x 2 − x 3
3
Solutions 215

x
(h) f (x) = , x0 = 1
ex−1

x
f (x) = ∴ f (1) = 1
ex−1
1−x
f  (x) = x−1 ∴ f  (1) = 0
e
x−2
f  (x) = x−1 ∴ f  (1) = −1
e
3−x
f  (x) = x−1 ∴ f  (1) = 2
e

Therefore,

f (0) (1) f  (1) f  (1) f  (1)

T (x) = (x − 1)0 + (x − 1)1 + (x − 1)2 + (x − 1)3
0! 1! 2! 3!
1 1
= 1 − (x − 1)2 + (x − 1)3
2 3

Exercise 9.8, page 204

1. L’Hôpital’s rule is a very powerful technique to calculate limits that take the following
indeterminate forms:

0 ∞
, , 0 · ∞, 00 , ∞0 , ∞ − ∞
0 ∞

Note that L’Hôpital’s rule is applicable only in the case of rational functions that take the form of

f (x) f (x0 ) 0 ∞
lim = = or,
x→x0 g(x) g(x0 ) 0 ∞

Then, the rational function’s limit may be calculated after taking derivates of its numerator and
denominator separately, that is to say

f (x) 0 ∞ f (x) f  (x)

if lim = or ⇒ lim = lim 
x→x0 g(x) 0 ∞ x→x0 g(x) x→x0 g (x)

Some of the following examples may be solved by other methods (see Example 8.2-1), here we
illustrate the use of L’Hôpital’s rule.

x 2 + 2x − 3 0 L.H. (x 2 + 2x − 3) 2x + 2
(a) lim = = lim = lim =4
x→1 x−1 0 x→1 (x − 1) x→1 1

sin(x) 0 L.H. (sin(x)) cos(x)
(b) lim = = lim = lim =1
x→0 x 0 x→0 (x) x→0 1
216 9 Derivatives


x + sin(x) 0 L.H. (x + sin(x)) 1 + cos(x)
(c) lim = = lim 
= lim =2
x→0 x 0 x→0 (x) x→0 1

sin(2x) 0 L.H. (sin(2x)) 2 cos(2x)
(d) lim = = lim 
= lim =2
x→0 x 0 x→0 (x) x→0 1

4 sin(5x) 0 L.H. (4 sin(5x)) 4
× 5 cos(5x)
(e) lim = = lim 
= lim =5
x→0 sin(4x) 0 x→0 (sin(4x)) x→0 4
cos(4x)

sin(2x) 0 L.H. (sin(2x)) 2 cos(2x) 2
(f) lim = = lim 
= lim =
x→0 sin(3x) 0 x→0 (sin(3x)) x→0 3 cos(3x) 3

2. Indeterminate forms (00 ), (0 · ∞), (∞ − ∞), etc. should be converted into the forms (0/0) or
(∞/∞), then L’Hôpital’s rule applies. The first following example illustrates technique that helps
us to resolve the exponential forms, which is used in the subsequent examples.
  x
(a) x→0lim x x = 00 = lim eln x = lim ex ln x = elimx→0+ x ln x = e0 = 1
+ x→0+ x→0+
 0   ln a     
(b) x→∞lim x = ∞ = e = a = lim exp ln x 1/x = ln a b = b ln a
1/x
 
x→∞
∞
ln x ln x
= lim exp = exp lim = exp
x→∞ x x→∞ x ∞
   
(ln x) 1/x 0
= exp lim = = = e0 = 1
L.H.
exp lim exp
x→∞ x x→∞ 1 1

ln(sin x) 0 L.H. (ln(sin x))
(c) lim = = lim
x→π/2 cos x x→π/2 (cos x)
0 
= [f (g(x))] = f  (g(x)) g  (x)
(cos x/ sin x) cos x 0
= lim = lim = =0
x→π/2 (− sin x) x→π/2 − sin2 x −12
   
(d) lim (sin x)tan x = 1∞ = eln a = a, ln a b = b ln a
x→π/2
 
= lim exp ln(sin x)tan x = lim exp (tan x ln(sin x))
x→π/2 x→π/2

x
sin: 1

= exp lim ln(sin x)
x→π/2 cos x

ln(sin x)  
= exp lim = see Example 9.8-2(c)
x→π/2 cos x

= e0 = 1
  
1 1 sin x − x 0
(e) lim − = (∞ − ∞) = lim =
x→0 x
 sin x x→0 x sin x 0
(sin x − x)
= lim
L.H.

x→0 (x sin x)

 
  
= [f (x) g(x)] = f (x)g(x) + f (x)g (x)
  
cos x − 1 0 (cos x − 1)
= lim = = lim
L.H.

x→0 sin x + x cos x 0 x→0 (sin x + x cos x)


− sin x 0
= lim = =0
x→0 cos x + cos x − x sin x 2
Solutions 217

ln x ∞ (ln x) 1/x

lim x ln x = (0 · ∞) = lim = = lim  = lim
L.H.
(f)
x→0 x→0 1/x ∞ x→0 (1/x) x→0 −1/x 2

= lim (−x) = 0
x→0

x ∞ (x)
lim x ex = (∞ · 0) = lim = = lim
L.H.
(g)
x→−∞ x→−∞ e−x ∞ x→−∞ (e−x )
1
= lim =0
x→−∞ −e−x
 !  "
1 x+5
 ∞ 1 x+5
(h) lim 1+ = 1 = lim exp ln 1 +
x→∞ x x→∞ x
 
= ln a = b ln a, e = x
b ln x


1
= exp lim (x + 5) ln 1 +
x→∞ x
= (exp(∞ · 0))
!  " 
ln 1 + x1  ∞ 
= exp lim = exp
x→∞ 1
x+5
∞
! #  $ "
ln 1 + x1
= exp lim
L.H.
# 1 $
x→∞
x+5
⎡ ⎤
−1/x 2
⎦ = exp lim (x + 5)
2
= ⎣
exp lim
1+1/x
x→∞ 1 x→∞ x 2 + x
(x+5)2

2(x + 5) L.H. 2

= exp lim = exp lim
L.H.

x→∞ 2x + 1 x→∞ 2

= e1 = e
ex ∞ (ex ) ex ∞ (ex )
= = = = =
L.H. L.H.
(i) lim lim lim lim
x→∞ x 2 ∞ x→∞ (x 2 ) x→∞ 2x ∞ x→∞ (2x)
ex
= lim =∞
x→∞ 2
 Function Analysis
10

Important to Know

Function analysis is usually done in the following steps

1. Domain: set of x where the function is defined,

2. Parity: if f (x) = f (−x) (even), if f (x) = −f (−x) (odd),
3. Zeros and sign: solve for f (x) = 0,
4. Vertical asymptotes: if around point x0 it is true that limx→x0 f (x) = ±∞,
5. Horizontal asymptotes: if limx→±∞ f (x) = a, where a = const.,
6. Oblique asymptote fbb (x): the asymptote equitation fbb (x) = ax + b, where a =
limx→∞ (f (x)/x) and b = limx→∞ (f (x) − ax),
7. Critical points: solve for f  (x) = 0 and f  (x) = 0,
8. The function graph.

10.1 Exercises

10.1 Basic Functions: Review

1. Sketch graphs and list the main properties of the following basic functions.
1 √
(a) f (x) = (b) f (x) = n x
x
(c) f (x) = a x (d) f (x) = loga x
(e) f (x) = sin x and f (x) = cos x (f) f (x) = tan x
(g) f (x) = arctan x

© Springer Nature Switzerland AG 2021 219

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_10
220 10 Functions

10.2 Composite Functions

1. Given that x = g(t) and y = f (x) derive composite function y = f (g(t)).

√ 1
(a) x = t 2 + 1, y = x (b) x = , y = ln x
t +1
√
(c) x = −(t 2 + 1), y = x (d) x = −t 2 , y = ln x

2. Assuming a composite function is in the form F (x) = f (g(x)), determine f (x) and g(x)
functions,
(a) F (x) = (2x + x 2 )4 (b) F (x) = cos2 x
√3
x x
(c) F (x) = √ (d) F (x) = 3

1+ 3 x 1+x
3. Assuming a composite function is in the form F (x) = f (g(h(x))), determine f (x), g(x), and
h(x) functions,
2 
(a) F (x) = 1 − 3x (b) F (x) = 4 1 + |x|

4. Derive f (x) given that:

(a) f (x + 1) = 1 − x (b) f (2x − 1) = x 2 − 2x + 1

2x − 1   9
(c) f =x (d) f x2 =
x x
5. Given a graph of ln function, without using calculator or graphic tools, sketch the graph of y =
ln(x − 2) − 1 function.

10.3 *** Analysis of Functions

1. Do complete study of the following function:

x3
f (x) =
(x − 1)2

2. Do complete study of the following function

2(1 − x)
f (x) = 2 −
x2 + 1
Solutions 221

3. Given function f (x) studied in Example 10.3-2,

(a) Derive the equation of a tangent T1 in point x = 1.
(b) Determine coordinates of point B so that tangent T2 to f (x) crossing B is parallel to line
y = −x. What is the relationship between T1 and T2 ?

4. Do complete study of the following function

x
f (x) =
(x + 1)(x − 4)

10.4 *** Functions: Study

1. Analyse and plot graphs of the following functions.

4x
(a) f (x) = x 4 − 2x 2 (b) f (x) =
4 − x2
1
(c) f (x) = e x (d) f (x) = (2 − x 2 )ex
 
1 1
(e) f (x) = x ln2 x (f) f (x) = x − ln 1 −
 4 x
(g) f (x) = x 2 − x 3
3
(h) f (x) = x x

Solutions

Exercise 10.1, page 219

1. Quick review of basic functions.

(a) Inverse, Fig. 10.1.
±k
f (x) =
x
Domain: x ∈ R, x = 0
Asymptotes: (H): y = 0, (V): x = 0
Derivatives:

1 2
f  (x) = ∓ ; f  (x) = ±
x2 x3

Critical points: none

222 10 Functions

Fig. 10.1 Example 10.1-1(a) f (x)

+c
x
−c
x

(b) Radicals, Fig. 10.2.

√
f (x) = n x

x ≥ 0 (n = 2k),
Domain:
x ∈ R (n = 2k + 1)
Derivatives:
√ √

n
x  1−n n x
f (x) = ; f (x) =
nx n2 x 2

Critical points:
zero: (0, 0), inflection: (0, 0)
Fig. 10.2 Example 10.1-1(b)
f (x)

√
2k
x
√
2k+1
x
Solutions 223

(c) Exponent, Fig. 10.3.

f (x) = a x

Domain: x ∈ R
Asymptotes: (H): y = 0
Derivatives:

f  (x) = a x ln a; f  (x) = a x ln2 a

Critical points: none

Y intercept point: (0, 1)

Fig. 10.3 Example 10.1-1(c)

ax (a > 1) f (x)
ax (0 < a < 1)
ax (a = 1)

(d) Logarithm, Fig. 10.4.

f (x) = loga x

Domain: (x, a) > 0, a = 1

Asymptotes: (V): x = 0
Derivatives:

1 1
f  (x) = ; f  (x) = −
x ln a x2 ln a

Critical points: zero: (0, 1)

Fig. 10.4 Example 10.1-1(c)
f (x)

loga x (a > 1)
loga x (0 < a < 1)
224 10 Functions

(e) Sine and cosine, Fig. 10.5.

f (x) = sin x and f (x) = cos x

Domain: x ∈ R
Period: T = 2π k, k ∈ Z
Derivatives:

sin x : f  (x) = cos x; f  (x) = − sin x

cos x : f  (x) = − sin x; f  (x) = − cos x
Critical points: zeros: x0 + kπ
(sin x : x0 = 0, cos x : x0 = π/2)

Extreme points: min/max: x0 + kπ

(sin x : x0 = π/2, cos x : x0 = 0)
f (max/min) = ±1

Fig. 10.5 Example 10.1-1(e)

f (x)
1

0 x

sin x
−1 cos x
0 p 2p

(f) Tangent, Fig. 10.6.

sin x
f (x) = tan x =
cos x
Domain: x = kπ
Period: T = 2π k, k ∈ Z
Asymptotes: (V): x = π/2 + kπ
Derivatives:

1 2 tan x
f  (x) = 2
; f  (x) =
cos x cos2 x

Critical points: zeros and inflection points: x = kπ

Solutions 225

Fig. 10.6 Example 10.1-1(f)

f (x)

0 x

0 p 2p

(g) Arctangent, Fig. 10.7.

f (x) = arctan x

Domain: x ∈ R
Asymptotes: (H): y = ±π/2
Derivatives:

1
sin x : f  (x) = ;
x2 + 1
2x
f  (x) = − 2
(x + 1)2
Critical points: zero: (0, 0)

Extreme points: none

Fig. 10.7 Example 10.1-1(g) p f (x)

2

0 x

p
−
2

Exercise 10.2, page 220

1. Composite functions are created as follows.


x = t2 + 1 
(a) √ ∴ y = f (g(t)) = f (t 2 + 1) = t 2 + 1
y = x
226 10 Functions

 
x = 1
t+1 1 1
(b) ∴ y = f (g(t)) = f = ln , (t > −1)
y = ln x t +1 t +1

x = −(t 2 + 1) 
√ ∴ y = f (g(t)) = f (−(t 2 + 1)) = −(t 2 + 1)
(c) y = x
∴ f (t) not defined in R : −(t 2 + 1) < 0 ∀x

x = −t 2
∴ y = f (g(t)) = f (−t 2 ) = ln(−t 2 )
(d) y = ln x
∴ f (t) not defined in R : −t 2 < 0 ∀x

2. By “decomposing” mathematical forms into its basic operations and clarifying the order of
operations we can resolve more complicated structures.
(a) F (x) = f (g(x)) = (2x + x 2 )4 where f (x) = x 4 and g(x) = 2x + x 2

(b) F (x) = f (g(x)) = cos2 x where f (x) = x 2 and g(x) = cos x

√3
x x √
(c) F (x) = f (g(x)) = √ where f (x) = and g(x) = 3 x
1+ x 3
1+x

x  x
(d) F (x) = f (g(x)) = 3
where f (x) = 3 x and g(x) =
1+x 1+x
3. By inspection of operations and their order we write.
2
(a) F (x) = f (g(h(x))) = 1 − 3x where f (x) = 1 − x, g(x) = 3x , h(x) = x 2
 √
(b) F (x) = f (g(h(x))) = 4 1 + |x| where f (x) = 4 x, g(x) = 1 + x, h(x) = |x|

4. Quick review of basic functions.

(a) Given f (x + 1) = 1 − x, therefore the argument mapping is
(x + 1) → x ∴ (x + 1) − 1 → x − 1 ∴ x → x − 1. That is to say, if argument x + 1 is
replaced with the argument x, then on the right side of the equation each occurrence of x must be
replaced with x − 1, i.e.

f (x) = 1 − (x − 1) = 2 − x

(b) Similarly, f (2x − 1) = x 2 − 2x + 1, therefore

(2x − 1) → x ∴ x → (x + 1)/2, i.e.
 2 
x+1 x+1 1 2
f (x) = −2 +1= (x − 2x + 1)
2 2 4
Solutions 227


2x − 1 1
(c) Given f = x, therefore x → − , i.e.
x x−2

1
f (x) = −
x−2

  9 9
(d) f x2 = ∴ f (x) = √ , (x > 0)
x x
5. Logarithmic function crosses the horizontal axis when its argument equals one. That is to say,
the root of ln(1) = 0 function, see Fig. 10.8. Equally, if the function’s argument is an expression,
again it has to equal one. Therefore, we write x − 2 = 1, that is x = 3 and simply translate
horizontally the ln x graph, see Fig. 10.8. Adding a constant affects the y value after we account
for the function’s argument. That is to say, causes the vertical shift of ln(x − 2), see Fig. 10.8.

Fig. 10.8 Example 10.2-5

0
−1 ln x
ln(x − 2)
ln(x − 2) − 1
0 1 2 3

Exercise 10.3, page 220

1. Given rational function, a complete study may be systematically done as follows.

x3 P (x)
f (x) = =
(x − 1)2 Q(x)

(a) Domain of definition D: there is one double pole of f (x) calculated as

Q(x) = 0 ∴ (x − 1)2 = 0 ∴ x1,2 = 1

therefore, f (x) is defined for all real numbers except number one where the function becomes
infinite (vertical asymptote), i.e. D : R, x = 1
228 10 Functions

(b) Function’s parity: functions can be odd, even, or none.

if: f (x) = f (−x) ∴ even

if: f (x) = −f (−x) ∴ odd

We write,

(−x)3 −x 3 −x 3 x3
f (−x) = = = = −
((−x) − 1)2 (−x − 1)2 (−1)2 (x + 1)2 (x + 1)2
= f (x) ∴ not even

x3 x3
−f (−x) = − − = = f (x) ∴ not odd
(x + 1)2 (x + 1)2

In conclusion, there is no symmetry relative to neither y-axis (even) nor to the origin point (odd).

(c) Function’s sign: There is one triple zero of f (x) found when

f (x) = 0 ∴ P (x) = 0 ∴ x 3 = 0 ∴ x1,2,3 = 0

Reminder: in order to determine sign of an expression, we use rules for products and ratios, for
example,

(neg)
(neg) × (neg) = (pos) or, = (neg) etc.
(pos)

A function may change its sign (i.e. positivity/negativity) at zero and break points, which divide
real domain D into the following intervals

x (−∞; 0) 0 (0; 1) (1; +∞)

x3 − 0 + +
(x − 1)2 + + + +
f (x) − 0 + +

In conclusion, f (x) is negative for x < 0, changes its sign at x = 0, and it is positive for x > 0.
Solutions 229

(d) Limits: having one vertical asymptote at x = 1, there are four boundary limits to calculate.
At far extremes, we find

x3 x2 x
lim = lim = lim x = −∞
x→−∞ x 2 − 2x + 1 x→−∞ 
0 0
 x→−∞
x2 1 − x2
+ x1
7

2
x3 x2 x
lim = lim = lim x = ∞
x→∞ x 2 − 2x + 1 0 0
x→∞
 2

 x→∞
x 2
1−  + 1
7

x x 2
Around vertical asymptote, we find

x3   (1)3
lim = for x = 1 ∴ (x − 1)2
> 0 regardless = = +∞
x→1− (x − 1)2 +0
x3   (1)3
lim = for x = 1 ∴ (x − 1)2
> 0 regardless = = +∞
x→1+ (x − 1)2 +0

(e) Oblique asymptote: two limits are calculated to determine coefficients in

yaa = ax + b

where

f (x) x3 ∞ (x 3 )
a = lim = lim = =
L.H.
lim
x→∞ x x→∞ x(x − 1)2 ∞ x→∞ (x 3 − 2x 2 + x)

3x 2 ∞ 6x ∞ 6
= lim = = = = lim
L.H. L.H.
lim
x→∞ 3x 2 − 4x + 1 ∞ x→∞ 6x − 4 ∞ x→∞ 6

=1
x3 x 3 − x(x − 1)2
b = lim (f (x) − x) = lim − x = lim
x→∞ x→∞ (x − 1)2 x→∞ (x − 1)2
! "
x 3 − x 3 + 2x 2 − 1 ∞ 4x ∞ 4
= lim = = = = lim
L.H. L.H.
lim
x→∞ x 2 − 2x + 1 ∞ x→∞ 2x − 2 ∞ x→∞ 2

=2

In conclusion, oblique asymptote is in form yaa = x + 2.

230 10 Functions

(f) Critical points: minimum/maxim and convex/concave points are found by solving f  (x) = 0
and f  (x) = 0 equations.

x3

3x 2 (x − 1)2 − 2x 3 (x − 1) x2 −
(x 
1)[3(x − 1) − 2x]
f  (x) = = =
(x − 1)2 (x − 1)4
(x − 1)4 3
x 2 (x − 3)
= = 0 ∴ x1,2 = 0, x3 = 3
(x − 1)3

We calculate f (x) at these two points as, f (0) = 0 and f (3) = 27/4.


 x 2 (x − 3)
  [x 2 (x − 3)] (x − 1)3 − x 2 (x − 3) [(x − 1)3 ]
f (x) = [f (x)] = =
(x − 1)3 (x − 1)6
(3x 2 − 6x)(x − 1)3 − 3x 2 (x − 3)(x − 1)2
=
(x − 1)6

(x
 − 1)2 [(3x 2 − 6x)(x − 1) − 3x 2 (x − 3)]
=
(x − 1)6 4
3x3 − 
 3x3 + 
9x2 + 6x −  9x2 6x
= = =0 ∴ x=0
(x − 1)4 (x − 1)4

Minimum/maxim and convex/concave points are determined by change of sign of critical points,
we summarize as

x (−∞; 0) 0 (0; 1) 1 (1; 3) 3 (1; +∞)

x2 + 0 + + 0 +
x−3 − − − − 0 +
(x − 1)3 − − − + + +
f  (x) + 0 + − 0 +
f (x)  0   27/4 
(min.)

x (−∞; 0) 0 (0; 1) 1 (1; +∞)

6x − 0 + +
(x − 1)4 + + + +
f  (x) − 0 + +
f (x) ∩ 0 ∪ ∪
f (x) (0, 0)

In conclusion, there is one minimum at min = (3, 27/4) and on inflection point at m = (0, 0).
Solutions 231

(g) Graphical representation: the calculated results are sufficient to plot f (x) graph, see
Fig. 10.9.

Fig. 10.9 Example 10.3-1

f (x)

min.

2
x+
=
y aa x

0 1

2. Following the same methodology as in Example 10.3-1, we analyse this function as follows,

2(1 − x) 2(x 2 + 1) − 2(1 − x) 2x 2 + 2
− 2
+ 2x) 2x(x + 1)

f (x) = 2 − = = =
x +1
2 x +1
2 x2 + 1 x2 + 1

(a) Domain of definition D: there are no vertical asymptotes, x ∈ [−∞, +∞]

(b) Function’s parity: We write,

2(−x)[(−x) + 1] 2x(1 − x)
f (−x) = = = f (x) ∴ not even
(−x) + 1
2 x2 + 1
2x(1 − x) 2x(x − 1)
−f (−x) = − = = f (x) ∴ not odd
x2 + 1 x2 + 1
232 10 Functions

(c) Function’s sign: There are two zeros of f (x) found as

f (x) = 0 ∴ P (x) = 0 ∴ 2x(x + 1) = 0 ∴ x1 = 0, x2 = −1

Denominator is x 2 + 1 > 0, therefore the sign of f (x) is determined by numerator only.

x (−∞; −1) −1 (−1; 0) 0 (0; +∞)

2x − − +
x+1 − + +
f (x) + − +

(d) Limits: we find one horizontal asymptote at y = 2 as

2x 2 + 2x ∞ 4x + 2 L.H. 4
= = lim = lim = 2
L.H.
lim
±∞ x +1
2 ∞ ±∞ 2x ±∞ 2

(e) Oblique asymptote: by definition we write

yaa = ax + b

where

f (x) 2x + 2  ∞  L.H. 2
1
a = lim = lim 2 = = lim =0
x→∞ x x→∞ x + 1 ∞ ±∞ 2
x
Therefore, there is no oblique asymptote.

(f) Critical points: minimum/maxim and convex/concave points are found by solving f  (x) = 0
and f  (x) = 0 equations.
  
 P (x) P  (x)Q(x) − P (x)Q (x)
f (x) = =
Q(x) Q2 (x)
∴
−2(x 2 − 2x − 1)
=
(x 2 + 1)2

Therefore,
√
 2± 8 √
f (x) = 0 ∴ x − 2x + 1 = 0 ∴ x1,2
2
= =1± 2
2
Solutions 233

Therefore, we write the first derivative in its factorized form as

−2 (x − x1 ) (x − x2 )
f  (x) =
(x 2 + 1)2
 √  √ 
−2 x − (1 − 2) x − (1 + 2)
=
(x 2 + 1)2

Denominator of f  (x) is greater than zero, thus we tabulate the corresponding critical points as
follows.

x (−∞; x1 ) x1 (x1 , x2 ) x2 (x2 ; +∞)

−2 − − −
x − x1 − 0 + +
x − x2 − − 0 +
f  (x) − 0 + 0 −
f (x)  (min.)  (max.) 

Where coordinates of minimum and maximum points of f (x) are calculated as

√ √
√ 2(1 − 2)(1 − 2 + 1) √
(min.): x1 = 1 −2 ∴ f (x1 ) = √ =1− 2
(1 − 2)2 + 1
√ √
√ 2(1 + 2)(1 + 2 + 1) √
(max.): x2 = 1 + 2 ∴ f (x1 ) = √ =1+ 2
(1 + 2) + 1
2

Similarly, convex/concave regions and coordinates and the corresponding inflection points are
found as follows.

−4(−x 3 + 3x 2 + 3x − 1)
f  (x) =
(x 2 + 1)3
∴
f  (x) = 0 ∴ −x 3 + 3x 2 + 3x − 1 = 0

A third order polynomial must have at least one real zero. By factor theorem we search the first
solution among the factors of the x 0 term, here it is ±1. We find that f  (−1) = 1 + 3 − 3 − 1 = 0,
therefore f  (x) is divisible by x + 1, i.e. x1 = −1. Consequently,
√ √
(−x 3 + 3x 2 + 3x − 1) ÷ (x + 1) = −x 2 + 4x − 1 = −(x − 2 + 3)(x − 2 − 3)
√ √
Therefore, we conclude that x1 = −1, x2 = 2 − 3, x3 = 2 + 3 and we write the second
derivative in its factorized form as
√ √
 4(x + 1)(x − 2 + 3)(x − 2 − 3)
f (x) = where (x 2 + 1)3 > 0 ∀x
(x 2 + 1)3
234 10 Functions

Denominator of f  (x) is greater than zero, thus we tabulate the corresponding critical points as
follows.

x (−∞; x1 ) x1 (x1 , x2 ) x2 (x2 , x3 ) x3 (x3 ; +∞)

x − x1 − 0 + + +
x − x2 − − 0 + +
x − x3 − − − 0 +
f  (x) − 0 + 0 − 0 +
f (x) ∩ (yp1 ) ∪ (yp2 ) ∩ (yp3 ) ∪

Where coordinates of inflection points of f (x) are calculated as

2(−1)(−1 + 1)
(yp1 ) : x1 = −1 ∴ f (x1 ) = =0
(−1)2 + 1
√ √ √
√ 2(2 − 3)(2 − 3 + 1) 3− 3
(yp2 ) : x2 = 2 − 3 ∴ f (x2 ) = √ =
(2 − 3)2 + 1 2
√ √ √
√ 2(2 − 3)(2 − 3 + 1) 3+ 3
(yp3 ) : x2 = 2 + 3 ∴ f (x3 ) = √ =
(2 − 3)2 + 1 2

(g) Graphical representation: in summary of function analysis, plot of f (x) is shown in

Fig. 10.10.

Fig. 10.10 Example 10.3-2

f ( x ) max.
y p3
2

y p2

0 y p1 x
min.

0
Solutions 235

3. We use results of the analysis in Example 10.3-2.

(a) Given condition that x = 1, we find that

2(1 + 1)
f (1) = = 2 ∴ A = (1, 2)
12 + 1

Slope m1 of tangent T1 in x = 1 is calculated as

−2(12 − 2 − 1)
f  (1) = = 1 ∴ m1 = 1
(12 + 1)2

Therefore, equation of tangent is T1 = m1 x + b1 = x + b1 , where constant b1 is found from the

condition that T1 crosses point A = (1, 2). That is to say

x = 1 : ∴ T 1 = 2 ∴ 2 = 1 + b 1 ∴ b 1 = 1 ∴ T1 = x + 1

We show this tangent in Fig. 10.11.

Fig. 10.11
Example 10.3-3(a) f (x)

2
1
+
x
=
1
T

0 x

−1 0 1
236 10 Functions

(b) Given line y = −x we know that its slope is m2 = −1, therefore all parallel lines have the
same slope. As a consequence

T2 = −x + b2

Now, the problem is to find at what x the slope (i.e. the first derivative) of f (x) equals f  (x) =
m2 = −1? We write

−2(x 2 − 2x − 1)
f  (x) = = −1
(x 2 + 1)2
∴
−2(x 2 − 2x − 1) = −(x 4 + 2x 2 + 1) ∴ x 4 + 4x + 3 = 0

By exploiting the factor theorem, and factors of number 3, we find that

P (x) = x 4 + 4x + 3 ∴ P (−1) = (−1)4 + 4(−1) + 3 = 0

Therefore, P (x) is divisible by x + 1 binomial and we factorize P (x) by divisions

(x 4 + 4x + 3) ÷ (x + 1) = x 3 − x 2 + x + 3

We repeat again the factor theorem to find

Q(x) = x 3 − x 2 + x + 3 ∴ Q(−1) = (−1)3 − (−1)2 + (−1) + 3 = 0

∴
(x 3 − x 2 + x + 3) ÷ (x + 1) = x 2 − 2x + 3

Therefore, after noting that this quadratic polynomial does not heave real zeros, we write: x 4 +
4x + 3 = (x + 1)2 (x 2 − 2x + 3). Now we solve the fourth order equation as

(x + 1)2 (x 2 − 2x + 3) = 0 ∴ x1,2 = −1

∴
f  (−1) = 0 and f (−1) = 0

We conclude the coordinates of B = (−1, 0) because B point must be at f (x) curve where the
first derivative equals zero.
Therefore, equation of T2 is derived as

T2 = −x + b2 ∴ −1 = 0 x + b2 ∴ b2 = −1 ∴ T2 = −x − 1

General condition for two lines to be orthogonal is that their slopes have relationship m1 =
−1/m2 , which is exactly the case with T1 and T2 (i.e. 1 = −1/(−1)). Therefore they are normal
to each other, as illustrated in Fig. 10.12.
Solutions 237

Fig. 10.12
Example 10.3-3(b) f (x)

1
T
y
2

=
−
T2

x
0 x
B

−1 0

4. Given rational function:

x
f (x) =
(x + 1)(x − 4)

(a) Domain of definition D:

x ∈ : (x + 1) = 0, ⇒ (x = −1) and,
x ∈ : (x − 4) = 0, ⇒ (x = 4)

Therefore, there are two vertical asymptotes at (x = −1) and (x = 4).

Also, we note the product: (x + 1)(x − 4) = x 2 − 3x − 4

(b) Function’s parity: functions can be odd, even, or none.

if: f (x) = f (−x) ∴ even

if: f (x) = −f (−x) ∴ odd

We write,

−x x
f (−x) = = = f (x) not even
(−x + 1)(−x − 4) (1 − x)(4 − x)
x
−f (−x) = − = f (x) not odd
(1 − x)(4 − x)
238 10 Functions

In conclusion, there is no symmetry relative to neither y-axis (even) nor to the origin point
(odd).

(c) Limits: we calculate limx→−∞ f (x) and limx→+∞ f (x)

x x ∞
lim f (x) = lim = lim 2 =
x→−∞ x→−∞ (x + 1)(x − 4) x→−∞ x − 3x − 4 ∞
1 1
= lim = = 0−
L.H.

x→−∞ 2x − 3 −∞

x x ∞
lim f (x) = lim = lim 2 =
x→+∞ x→+∞ (x + 1)(x − 4) x→+∞ x − 3x − 4 ∞
1 1
= lim = = 0+
L.H.

x→+∞ 2x − 3 +∞

(d) Limits at x = −1: we calculate limx→−1− f (x) et limx→−1+ f (x)

x −1 (−)
lim f (x) = lim = lim = = −∞
x→−1− x→−1− (x + 1)(x − 4) x→−1− (x + 1)(−5) (−)(−)
  
<0

x −1 (−)
lim f (x) = lim = lim = = +∞
x→−1+ x→−1+ (x + 1)(x − 4) x→−1+ (x + 1)(−5) (+)(−)
  
>0

(e) Limits at x = 4: we calculate limx→4− f (x) et limx→4+ f (x)

x 4 (+)
lim f (x) = lim = lim = = −∞
x→4− x→4− (x + 1)(x − 4) x→4− (5) (x − 4) (+)(−)
  
<0

x 4 (+)
lim f (x) = lim = lim = = +∞
x→4+ x→4+ (x + 1)(x − 4) x→4 + (5) (x − 4) (+)(+)
  
>0

(f) Function’s zeros: we calculate

x
f (x) = 0 ⇒ =0 ∴ x=0
(x + 1)(x − 4)
Solutions 239

(g) Derivatives and critical points:

x
f (x) =
(x + 1)(x − 4)
 
x (x 2 − 3x − 4) − x(2x − 3)  − 4 − 2x 2 + 
x2 − 
3x )
3x
f  (x) = = =
x − 3x − 4
2 (x − 3x − 4)
2 2 (x − 3x − 4)
2 2

x2 + 4 (+)
=− =0=− ∴ f  (x) < 0; (x ∈ )
(x 2 − 3x − 4)2 (+)


x2 + 4 2x(x 2 − 3x − 4)2 − (x 2 + 4) 2 (x 2 − 3x − 4)(2x − 3)
f  (x) = − = −
(x 2 − 3x − 4)2 (x 2 − 3x − 4)4
(( # $
x 2(
2(( −(3x(− 4) x(x 2 − 3x − 4) − (x 2 + 4)(2x − 3)
=−
(x 2 − 3x − 4)4 3
# 3 $
3x2 − 4x − 2x 3 − 8x + 
2 x − 3x2 + 12
=−
(x 2 − 3x − 4)3
x 3 + 12x − 12
=2
(x + 1)3 (x − 4)3

Without doing the numerical calculations, we can only estimate that f  (x) = 0 ⇒ N(x) =
x 3 + 12x − 12 = 0 as: N(0) = −12 < 0 and N(1) = 1 > 0, Therefore, there must be a real zero
x = x0 found within the interval [0, 1] very close to x0 ≈ 1. Therefore,

f (x) = 0; ⇒ x = 0, y = 0
f  (x) = 0; and f  (x) < 0
1
f  (x) = 0; ⇒ x0 ≈ 1; ∴ x0 ≈ 1, y0 ≈ −
6

(h) Summary of variations:

x (−∞, −1) (−1) (−1, 0) (0) (0, x0 ) (x0 ) (x0 , 4) (4) (4, +∞)
(−) (−) (+) (+) (+)
f (x) (−)(−) (+)(−) 0 (+)(−) ≈ −1/6 (+)(−) (+)(+)
f  (x) (−) (−) −1/4 (−) (−) (−) (−)
f (x)   0    
f  (x) (−)
(−)(−)
(−)
(+)(−) +3/8 (−)
(+)(−) 0 (+)
(+)(−)
(+)
(+)(+)
f (x) ∩ ∪ ∪ ∪ (x0 , y0 ) ∩ ∪
240 10 Functions

(i) Oblique asymptote: two limits are calculated to determine coefficients in

yaa = ax + b

where

x
f (x) (x+1)(x−4) 1 1
a = lim = lim = lim = =0
x→±∞ x x→±∞ x
x→±∞ (x + 1)(x − 4) ±∞

b = lim f (x) − * 0 = lim f (x) = (already calculated) = 0

ax
x→±∞ x→±∞

therefore, because (a = b = 0), there is no oblique asymptote. Instead, there is one horizontal
asymptote at y = 0.

(j) Graphical representation: the complete function graph is shown in Fig. 10.13.

Fig. 10.13 Example 10.3-4

f (x)

0 x
(x0 , y0 )

−1 0 4

Exercise 10.4, page 221

1. Some of the typical function forms to study.

(a) f (x) = x 4 − 2x 2 , Fig. 10.14.
Domain: x ∈ R
Derivatives:

f  (x) = 4x 3 − 4x;
f  (x) = 12x 2 − 4
Solutions 241

Critical points:
√ √
zeros: (− 2, 0), (0, 0), (0, 0), ( 2, 0)
inflection
√ points: √
(−1/ 3, −5/9), (1/ 3, −5/9)
extreme points:
min: (−1, −1), (1, −1) max: (0, 0)

Fig. 10.14
Example 10.4-1(a)
f (x)

0 x
−1

−1 0 1

4x
(b) f (x) = , Fig. 10.15.
4 − x2
Domain: x = ±2
Asymptotes:
(V): x = −2, x = 2, (H): y = 0
Derivatives:

4(4 + x 2 )2
f  (x) = ;
(4 − x 2 )2
8x(x 2 + 12)2
f  (x) =
(4 − x 2 )3
Critical points:
zeros: (0, 0)
inflection point: (0, 0)
242 10 Functions

Fig. 10.15
Example 10.4-1(b) f (x)

x
0

−2 0 2

1
(c) f (x) = e x , Fig. 10.16.
Domain: x = 0
Asymptotes:
(V): x = 0, x = 2, (H): y = 1
Derivatives:
1
 ex
f (x) = − 2 ;
x
1
 ex
f (x) = 4 (2x + 1)
x
Limits: lim f (x) = 0
x→0−
Critical points:
inflection point: (−1/2, 1/e2 )

Fig. 10.16
Example 10.4-1(c) f (x)

1
x
0

0
Solutions 243

(d) f (x) = (2 − x 2 )ex , Fig. 10.17.

Domain: x ∈ R
Derivatives:

f  (x) = −(x 2 + 2x − 2)ex ;

f  (x) = −x(x + 4)ex
Critical points:
√ √
zeros: (− 2, 0), ( 2, 0)
extreme
 points: √ 
√ 2(1 + 3)
min: −1 − 3, − √
e1+ 3
 √ 
√ 2(1 − 3)
max: −1 + 3, − √
e1− 3
inflection points: (0, 2), (−4, −14/e4 )

Fig. 10.17
Example 10.4-1(d) f (x)

√ √
−4 − 2 0 2

(e) f (x) = x ln2 x, Fig. 10.18.

Domain: x > 0
Derivatives:

f  (x) = (2 + ln x) ln x;
2(1 + ln x)
f  (x) =
x
Critical points:
zeros: (0, 0), (1, 0), (1, 0)
extreme points:
min: (1, 0)
max: (1/e2 , 4/e2 )
inflection points: (1/e, 1/e)
244 10 Functions

Fig. 10.18
Example 10.4-1(e) f (x)

x
0

0 1

 
1 1
(f) f (x) = x − ln 1 − , Fig. 10.19.
4 x
Domain: x < 0 and x > 1
Asymptotes:
(V): x = 0, x = 1, (H): y = −1
Derivatives:

 1 4x − 1
f (x) = ln 1 − +
x 4x(x − 1)
2x + 1
f  (x) = −
4x 2 (x − 1)2
Critical point:
inflection points: (−1/2, −3/4 ln 3))

Fig. 10.19
f (x) x→
Example 10.4-1(f)

−1

−1/2 0 1

In order to find the extreme points, i.e. min and max, we must solve f  (x) = 0 equation. However,
in this case the required equation is transcendental because it has the form of “ln h(x) = g(x)”,
i.e.

 1 4x − 1
f (x) = ln 1 − + =0
x 4x(x − 1)

1 4x − 1
ln 1 − =−
x 4x(x − 1)
Solutions 245

where, in general case, there is no closed form analytical solution. We must use, for example,
graphic method or numerical method to approximate the solution. Another possible method is to
develop Taylor polynomial of f  (x), then search for its roots. Note that some functions are very
sensitive to the rounding numerical errors.
By inspection of f (x), we already found that f (x) is defined for x < 0, that the inflection point is
at x = −1/2, that horizontal asymptote is at y = −1, where f (x) < 0 for x > 1. Thus, first good
guess for position of the extreme point could be in the middle of ]0, −1/2[ interval, for example,
x0 = −1/4. After simplification, second order Tylor polynomial around x0 = −1/4 is

f  (x) ≈ T (x) ≈ −8.7x 2 − 5.6325x − 0.8546

∴
√
5.6325 ± 5.63252 − 4 × 8.7 × 0.8546
x0 ≈
−2 × 8.7
∴
x0 ≈ 0.243 (max.)

(the second root of quadratic equation is too far from the initial guess, where T (x) is not correct
anymore).

 
(g) f (x) = 3 x 2 − x 3 = 3 x 2 (1 − x), Fig. 10.20.
Domain: x ∈ R
Asymptotes: (affine): y = 1/3 − x
Derivatives:

x(3x − 2)
f  (x) = −  ; (x = 0)
3 3 x 2 (1 − x)

 2 3 [x 2 (1 − x)]2
f (x) = −
9x 2 (x − 1)2
Critical points:
zeros: (0, 0), (1, 0)
extreme points:  
√3
2 4
min: (0, 0)) max: ,−
3 3
inflection points: (0, 0), (1, 0)

Fig. 10.20
f (x)
Example 10.4-1(g)

1/3
x
0

−2/3

0 2/3 1
246 10 Functions

Note that f  (0) is not defined, nevertheless there is change of sign of f  (x), thus the min. point.

Hint: two limit calculations for the affine asymptote y = ax + b reduce to

√
f (x) 3
x2 − x3 1
a = lim = lim = lim − 1 = −1
3

x→∞ x x→∞ x x→∞ x

∴
 1
b = lim f (x) − ax = lim x 2 − x 3 + x = lim x −1+x
3 3

x→∞ x→∞ x→∞ x

   
3 1
−1+1 
1 x 0
= lim x − 1 + 1 = (∞ 0) = lim =
3

x→∞ x x→∞ 1 0
x
+
3 1
, 0
−1 ,
x -3
1
 −1
= lim
L.H. 3x 2 − 3x = lim − x 2 
x
=
1
x→∞ 1 x→∞ 
0 3
− 2 1
x 3x 2
1− 
x
√
Note that 3
−1 = −1, therefore b = (−1)(−1)(1/3) is positive.

(h) f (x) = x x , Fig. 10.21.

Domain: x > 0
Derivatives:

f  (x) = x x (1 + ln x); (x = 0)
f  (x) = x x (1 + ln x)2 + x x−1
Critical points:
extreme
 point: 
min: 1/e, (1/e)1/e

Fig. 10.21
Example 10.4-1(h) f (x)

x
0
0
 Integrals
11

Important to Know

Basic tabular integrals

;
f (x) f (x) dx

xn n x n−1 + C, (n = −1)
1
ln |x| + C (x = 0)
x
1 ax
eax e C
a
ax
ax +C
ln a
sin x − cos x + C
cos x sin x + C
1
√ arcsin x + C
1 − x2
1
arctan x + C
1 + x2

Basic techniques of integration

; ;
C f (x) dx = C f (x) dx
;   ; ;
f (x) ± g(x) dx = f (x) dx ± g(x) dx
; ;
f (x) g  (x) dx = f (x) g(x) − f  (x) g(x) dx (partial integration)

© Springer Nature Switzerland AG 2021 247

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_11
248 11 Integrals

How to choose the integration method? In general, try the possible methods in the following order:

1. check if the function is found in the list of tabular integrals (i.e. trivial solution);
2. simplify or modify the function’s form by algebra methods, so that it is reduced to the tabulated
forms;
3. one or more subsequent changes of variable that transform a given function to the tabulated forms;
4. partial integration, or
5. combination of change of variable and partial integration techniques.

11.1 Exercises

11.1 Basic Integrals

1. Calculate integrals of the following basic functions:

(a) f (x) = x 3 + 2x − 1 (b) f (x) = x −3 + x
x+1
(c) f (x) = √ (d) f (x) = (x + 1)(2x + 1)
x
(e) f (x) = 3 cos x + sin x (f) f (x) = ex + cos3 x(1 + tan2 x)

11.2 * Integration by Substitution

1. Calculate integrals of the following functions by substitution of variables:


(a) f (x) = x 2 (x 3 + 1)5 (b) f (x) = x x 2 + 1
x3 sin x
(c) f (x) = √ (d) f (x) =
1 − x4 cos4 x
(e) f (x) = tan x (f) f (x) = x cos(x 2 + 2)
(g) f (x) = 1 + tan2 x

11.3 * Integration by Parts

1. Calculate integrals of the following functions by integration by parts:

(a) f (x) = xex (b) f (x) = x 2 e−x (c) f (x) = ln x
2
(d) f (x) = x sin x (e) f (x) = x 2 ln x (f) f (x) = x 3 ex
ln x
(g) f (x) = (x 2 − 1)ex (h) f (x) = ex sin x (i) f (x) =
x2
11.1 Exercises 249

11.4 ** Important Integrals

1. Calculate integrals of the following functions:

;
(a) sin(2x) dx
;
(b) sin2 (x) dx
; 1
(c) f (x)2 dx given that f (x) is a piecewise linear function as defined by its graph, see
0
Fig. 11.1.

Fig. 11.1 Example 11.4-1(c) 1

f (x)
0

−1
0 0.25 0.5 0.75 1
x

11.5 *** Gaussian Integral

1. Evaluate integral
; ∞
e−x dx
2

−∞

11.6 * Integration of Rational Functions

1. Calculate the following rational function by using the method of integration by partial fractions.
When possible, reuse derivations in Example 2.14-1.
x−1 x x+2
(a) f (x) = 2 (b) f (x) = (c) f (x) = 3
x +x (x + 1)(x − 4) x − 2x 2
2x 2 1
(d) f (x) = (e) f (x) =
x4 −1 x3 −1
250 11 Integrals

11.7 * Trigonometric Substitution

1. Calculate integrals of the following functions by using trigonometric substitutions:

 √
x2
(c) f (x) = 25x + 4
2
(a) f (x) = 9 − x 2 (b) f (x) = √
36 − x 2 x 4

11.8 * Definite Integrals

1. In the given closed interval (i.e. including the end points), calculate area under the function.
(a) f (x) = x 3 , x ∈ [2, 3] (b) f (x) = x(1 + x 3 ) x ∈ [−1, 2]
(c) f (x) = cos x x ∈ [−π/4, π/4] (d) f (x) = cos x x ∈ [0, 2π ]
x
(e) f (x) = √ x ∈ [0, 2] (f) f (x) = x 2 cos x x ∈ [0, π ]
1 + 3x 2
2. Calculate area between two functions.
(a) f (x) = x 2 , g(x) = 8 − x 2

3. Given interval x ∈ [5, 6], calculate area under function

x
f (x) =
(x + 1)(x − 4)
4. Using the definite integrals, calculate
(a) area of right-angled triangle, where the two catheti equal a = 2 and b = 1;
(b) area of the circle whose radius equals r = 1; and
(c) volume of cuboid whose sides equal a = 4, b = 3, and c = 2.

5. In the given closed interval (i.e. including the end points), calculate the averages of the following
functions:
(a) f (x) = sin x x ∈ [0, 2π ] (b) f (x) = cos x x ∈ [π/2, 3π/2]
(c) f (x) = 1 + cos x x ∈ [0, 2π ] (d) f (x) = cos2 x x ∈ [0, 2π ]

6. Given closed intervals (i.e. including the end points) and function graph, Fig. 11.2, calculate the
averages
(a) x ∈ [0, 1]
(b) x ∈ [0, 2]
(c) x ∈ [0, 3]
(d) x ∈ [0, 4]
(e) x ∈ [1, 2]
(f) x ∈ [2, 3]
(g) x ∈ [2, 4]
(h) x ∈ [0.5, 3]
(i) x ∈ [1.5, 3.5]
11.1 Exercises 251

Fig. 11.2 Example 11.8-6

f (x)
2

x
0
0 1 2 3 4

7. Given closed intervals (i.e. including the end points) and functions, Fig. 11.3, calculate the
averages of f (x) and g(x) within each interval.
(a) x ∈ [0, 1]
(b) x ∈ [0, 2]
(c) x ∈ [0, 3]
(d) x ∈ [0, 4]
(e) x ∈ [0, 10]
(f) x ∈ [5, 9]
(g) x ∈ [1, 8]
(h) x ∈ [1, 7]
(i) x ∈ [2, 6]

Fig. 11.3 Example 11.8-7 f (x)

2

x
0
g(x)
2

x
0
0 2 4 6 8 10

8. Given closed intervals (i.e. including the end points) and functions, Fig. 11.4, calculate the
averages of f (x) and g(x) within each interval.
(a) x ∈ [0, 1]
(b) x ∈ [0, 2]
(c) x ∈ [0, 3]
(d) x ∈ [0, 4]
252 11 Integrals

(e) x ∈ [0, 10]

(f) x ∈ [5, 9]
(g) x ∈ [1, 8]
(h) x ∈ [1, 7]
(i) x ∈ [2, 6]

Fig. 11.4 Example 11.8-8 f (x)

2

x
0
g(x)
2

x
0
0 2 4 6 8 10

11.9 *** Improper Integral

1. If possible, calculate area under the following functions:

1 1
(a) f (x) = , x ∈ [1, ∞] (b) f (x) = , x ∈ [1, ∞]
x2 x
1
f (x) = √ , x ∈ [−1, 1] f (x) = xe−x , x ∈ [−∞, ∞]
2
(c) 3
(d)
x2
1 1
(e) f (x) = √ , x ∈ [−∞, 0] (f) f (x) = , x ∈ [0, ∞]
1−x 1 + x2

11.10 *** Improper Integral: Discussion

1. The limitation of blind implementation of the Newton–Leibniz formula is illustrated by the

following functions:
1 √
(a) f (x) = , x ∈ [−1, 1] (b) f (x) = 1 − cos 2x, x ∈ [0, 100π ]
x
Solutions 253

Solutions

Exercise 11.1, page 248

1. By using the list of basic integrals, we write

; ; 
x n+1
(a) (x 3 + 2x − 1) dx = xn = +C
n+1
x4 x2 x4
= + 2
− x + C = + x2 − x + C
4 2
4
;
x −2 x2 1 x2
(b) (x −3 + x) dx = + +C =− 2 + +C
−2 2 2x 2
; ; ; ; ;
x+1 x 1
√ dx = √ dx + √ dx = x dx + x − /2 dx
1/2 1
(c)
x x √ x
x 3/2 x 1/2 x x √ √ x 
= + +C =2 +2 x+C =2 x +1 +C
3/2 1/2 3 3
; ; 3
2x 3x 2
(d) (x + 1)(2x + 1) dx = (2x 2 + 3x + 1) dx = + +x+C
3 2
; ; ; 
(e) (3 cos x + sin x) dx = cos x dx = sin x, sin x dx = − cos x
= 3 sin x − cos x + C
; ; ;
 x   3 
(f) e + cos3 x(1 + tan2 x) dx = ex dx + cos x(1 + tan2 x) dx
; 
sin2 x
=e +
x
cos x 1 +
3
dx
cos2 x
⎡ ⎛ ⎞⎤
; 2 2

:1
⎢ ⎜ sinx + cos x ⎟⎥
= ex + ⎣cos3 x ⎝  2 ⎠⎦ dx
cos
 x
;
= ex + cos x dx = ex + sin x + C

Exercise 11.2, page 248

1. Change of variables reduces integral to some of the basic forms.

;  
1
(a) x (x + 1) dx = x + 1 = t ∴ 3x dx = dt ∴ x dx = dt
2 3 5 3 2 2

; 3
1 1 t6 1 3
= t dt =
5
+C = (x + 1) + C
6
3 36 18
;   
1
(b) x x 2 + 1 dx = t = x 2 + 1 ∴ 2x dx = dt ∴ x dx = dt
; ; 2
1 √ 1 1 t 3/2 1 2
= t dt = t dt = = (x + 1)3 + C
1/2

2 2 2
3/2 3
254 11 Integrals

;  
x3 1
(c) √ dx = 1 − x = t ∴ −4x dx = dt ∴ x dx = − dt
4 3 3
1 − x4 ; ; 4
1 dt 1 −1/2 1
=− √ =− t dt = − 1−x +C
4
4 t 4 2
;
sin x  
(d) 4
dx = cos x = t ∴ − sin x dx = dt ∴ sin x dx = − dt
cos x ; ;
dt −4 t −3 1
=− dt = − t dt = − +C = +C
t4 −3 3 cos3 x
; ;
sin x  
(e) tan x dx = dx = cos x = t ∴ − sin x dx = dt
cos x ;
dt
=− = − ln |t| + C = − ln | cos x| + C
t
;  
1
(f) x cos(x 2 + 2) = x 2 + 2 = t ∴ 2x dx = dt ∴ x dx = dt
; 2
1 1
= cos t dt = sin(x + 2) + C2
2 2
; ; ;
sin + cos
2 2
1
(g) 1 + tan2 x dx = 2
dx = dx
⎧ cos cos2 ⎫
⎪ sin x ⎪
⎪
⎨t = tan x = cos x ⎪
⎬
⎪ ⎪
⎩ dt = cos x cos x − sin x(− sin x) dx = cos x + sin x dx = 1 dx ⎪
2
⎪ 2
⎭
2
cos x 2
cos x 2
cos x
∴
;
= dt = t + C = tan x + C

Exercise 11.3, page 248

1. The idea of integration by parts method is to choose one function so that its integral is simpler,
while the other one has simpler derivative.
⎧ ⎫
; ⎨u = x
⎪ ∴ du = dx ⎪
⎬
(a) xe dx =
x ;
⎪
⎩dv = ex dx ∴ v = ex dx = ex ⎪ ⎭
;
= x ex − ex dx = xex − ex + C = ex (x − 1) + C
Solutions 255

;
(b) x 2 e−x dx =
⎧ ⎫
⎨u = x
⎪ ∴ du = 2x dx ⎪
2
⎬
;
⎪
⎩dv = e−x dx ∴ v = e−x dx = {t = −x} = −e−x ⎪ ⎭
; ;
 
= −x 2 e−x − −2xe−x dx = −x 2 e−x + 2 xe−x dx
⎧ ⎫
⎨u = x
⎪ ∴ du = dx ⎪
⎬
;
⎪
⎩dv = e−x dx ∴ v = e−x dx = {t = −x} = −e−x ⎪⎭
 ;
= −x 2 e−x + 2 −xe−x − (−e−x ) dx

= −x 2 e−x − 2xe−x − 2e−x + C

 
= −e−x x 2 + 2x + 2 + C
⎧ ⎫
; ⎪
⎨u = ln x dx ⎪
⎬
∴ du =
(c) ln x dx = x
⎪
⎩dv = dx ⎪
⎭
∴ v=x
;
dx
= x ln x − x
= x ln x − x + C = x(ln x − 1) + C
x

⎧ ⎫
; ⎨u = x
⎪ ∴ du = dx ⎪
⎬
(d) x sin x dx = ;
⎪
⎩dv = sin x dx ∴ v = sin x dx = − cos x ⎪ ⎭
;
= −x cos x + cos x dx = −x cos x + sin x + C
⎧ dx ⎫
⎪ ⎪
; ⎨u = ln x
⎪ ∴ du =
x
⎪
⎬
(e) x 2 ln x dx = ;
⎪
⎪ x3 ⎪⎪
⎩dv = x 2 dx ∴ v = x 2 dx = ⎭
3
; 32
x3 x  dx x3 1 x3 x3 x3
= ln x − = ln x − +C = ln x − +C
3 3 x
1 3 3 3 3 9
; ;
2 2
(f) x 3 ex dx = x 2 x ex dx
⎧ ⎫
⎨u = x
⎪ ∴ du = 2x dx ⎪
2
⎬
;
⎪ 1
∴ v = x ex dx = ex ⎪
2 2 2
⎩dv = x ex dx ⎭
2
 ; 
x2 1 x2
where, t = x ∴ dt = 2x dx ∴
2
x e dx = e
2
; ;
1 2 1 x2 1 2 2
= x 2 ex − e 2
x dx = x 2 ex − x ex dx
2 2
2
ex  2 
2
1 2 x2 1 x2
= x e − e +C = x −1 +C
2 2 2
256 11 Integrals

; ; ;
(g) (x − 1)e dx = x e dx − ex dx
2 x 2 x

; 
 2 
x e dx = ( by parts ) = e x − 2x + 2
2 x x

 
= ex x 2 − 2x + 2 − ex + C
 
= ex x 2 − 2x + 1 + C
⎧ ⎫
⎨u = e
⎪ ∴ du = ex dx ⎪
x
; ⎬
e sin x dx =
x ;
(h) ⎪
   ⎩dv = sin x dx ∴ v = sin x dx = − cos x ⎪
⎭
I ;
= −ex cos x + ex cos x dx
⎧ ⎫
⎨u = e
⎪ ∴ du = ex dx ⎪
x
⎬
;
⎪
⎩dv = cos x dx ∴ v = cos x dx = sin x ⎪
⎭
;
= −e cos x + e sin x −
x x
ex sin x dx
  
I

This integral is an example of a “circular” form, i.e. after applying the integration by parts
technique two times, we must again solve the original integral. However, we write

I = −ex cos x + ex sin x − I ∴ 2I = ex sin x − ex cos x

∴
;
ex
I= ex sin x dx = (sin x − cos x) + C
2

⎧ ⎫
⎪ 1 ⎪
; ⎪
⎨u = ln x ∴ du = dx ⎪
⎬
ln x x
(i) dx = ;
x2 ⎪
⎪ 1⎪
dx = − ⎪
1 1
⎩dv = dx ∴ v= ⎭
x;2 x2 x
ln x 11 ln x 1 1
=− + dx = − − + C = − (ln x + 1) + C
x xx x x x

Exercise 11.4, page 249

1. Three integrals in this exercise are typical forms encountered in engineering.

Solutions 257

(a) Trigonometric functions of multiple angles are solved by change of variable technique.

dt 1
t = 2x ∴ = 2 ∴ dx = dt
dx 2
∴
; ;
1 1 1
sin(2x) dx = sin t dt = − cos t = − cos(2x) + C
2 2 2

(b) Squares of sine/cosine functions are replaced by their trigonometric identities for double
angles, followed by change of variable technique.

1
sin x sin x = [cos(x − x) − cos(x + x)]
2
1 1
sin2 (x) = [cos(0) − cos(2x)] = [1 − cos(2x)]
2 2

Then,
; ; ; ;
1 1 1
sin (x) dx =
2
[1 − cos(2x)] dx = dx − cos(2x) dx
2 2 2

1 1
= x − sin(2x) + C
2 2

(c) Piecewise linear functions are among most often used approximations in engineering. Given
their graphical representation, it is necessary to derive analytical forms for each of the linear
sections separately. The main idea is that the operation of “integration” itself is fundamentally
an operation of addition. Also, a linear function is the easiest function to integrate. Therefore,
knowing a priori the analytical form of each linear piece, the overall integral is reduced to the
sum of simple integrals of linear functions.
A linear section y = ax + b is determined by two points in the plane, that is to say that knowing
coordinates (x, y) of these two points on the line two constants (a, b) are determined with a
simple algebra.
Here we derive three linear sections, as emphasized by different colors in graph and points
(A, B, C, D), see Fig. 11.5.
1. AB: at coordinates of the two end points (x, y), we write y = ax + b as

(0, 0) ∴ 0 = a × 0 + b ⇒ b = 0
1
(0.25, 1) ∴ 1 = a × +b ⇒a =4
4
258 11 Integrals

Therefore, analytical form of AB linear segment is f (x) = 4 x, which is valid in the interval
(0, 0.25). Therefore, in this interval, we calculate

; ; %1 
x 3 %% 4
1 1
4 4 16 1 1
I1 = f (x)2 dx = (4x)2 dx = 16 = −0 =
0 0 3 %0 3 43 12

2. BC: at coordinates of the two end points (x, y), we write y = ax + b as

1
+b
(0.25, 1) ∴ 1 = a ×
4
3
(0.75, −1) ∴ −1 = a × + b
4

The solution of this system of two linear equations is (a, b) = (−4, 2). Therefore, analytical form
of CD linear segment is f (x) = −4 x + 2, which is valid in the interval (0.25, 0.75). Therefore,
in this interval, we calculate
; 3 ; 3 ; 3
4 4 4
I2 = f (x)2 dx = (−4x + 2)2 dx = (16x 2 − 16x + 4) dx
1 1 1
4 4 4
 %3 %3 % 34 
x 3 %% 4 x 2 %% 4 %
= 4 4 % − 4 % + x %%
3 1 2 1 1
4 4 4
 
1 26 1 8 2 13 1
= 16 − 16 +4 = −2 =
3 64 2 16 4 6 6

3. CD: at coordinates of the two end points (x, y), we write y = ax + b as

3
(0.75, −1) ∴ −1 = a × +b
4
(1, 0) ∴ 0 = a × 1 + b

The solution of this system of two linear equations is (a, b) = (4, −4). Therefore, analytical form
of CD linear segment is f (x) = 4 x − 4, which is valid in the interval (0.75, 1). Therefore, in this
interval, we calculate
; 1 ; 1 ; 1
I3 = f (x)2 dx = (4x − 4)2 dx = (16x 2 − 32x + 16) dx
3 3 3
4 4 4
 %1 %1 %1 
x 3 %% x 2 %% %
= 16 % − 2 % + x %%
3 3 2 3 3
4 4 4
 
1 37 7 1 37 1
= 16 − 16 + 16 = −3 =
3 64 16 4 12 12
Solutions 259

(c) (cont.) After finding these three integrals, we return to the original questions to conclude
; 1
1 1 1 1
I= f (x)2 dx = I1 + I2 + I3 = + + =
0 12 6 12 3

Geometrical interpretation is that integral I equals the total area underneath quadratic function
f (x)2 that consists of three distinct regions. These three quadratic sub-functions and their
respective areas are shown by colored sections in Fig. 11.5.

Alternative Solution Due to symmetry of given f (x)2 , area I1 in Fig. 11.5 is identical to areas of
the other three segments, i.e. (0.25, 0.5), (0.5, 0.75), and (0.75, 1.0). Thus, it is sufficient to solve
only the first integral within (0.0, 0.25) interval and keep in mind that it accounts for one-quarter
of the total surface area I ; that is to say,

; ; %1 
x 3 %% 4
1 1
4 4 16 1 1
I1 = f (x)2 dx = (4x)2 dx = 16 = −0 =
0 0 3 %0 3 43 12
∴
1 1
I = 4 I1 = 4 × =
12 3

As a comparison, a simple integral of f (x)

; 1 ; 1 ; 3 ; 1
4 4
A= f (x) = 4x dx + (−4x + 2) dx + (4x − 4) dx = 0
1 3
0 0 4 4

which may also be concluded by graph inspection; the area of triangle above zero (i.e. positive)
equals the area of triangle below zero (i.e. negative), and therefore their sum equals zero. Another
way to say is that average of this function equals zero.

Fig. 11.5 Example 11.4-1(c)

1

I1 I2 I3
f (x)

−1

0 0. 25 0.5 0. 75 1
x
260 11 Integrals

Exercise 11.5, page 249

1. Gaussian integral is one of the most important forms in science and engineering. There are
multiple methods to solve it, one possible method exploits the symmetry of multivariable integral
form by artificially adding second variable y as follows:
; ∞ ; ∞ ; ∞ ; ∞ ; ∞
e−x dx = e−x −y 2
e−x e−y dx dy
2 2 2 2
dx dy =
−∞ −∞ −∞ −∞ −∞
; ∞ ; ∞
e−y dy e−x dx = I 2
2 2
=
−∞ −∞
    
I I

Even though there are two variables involved x and y, the two integrals have the same form,
therefore the same form of their respective solutions. Therefore, we write
; ∞ ; ∞
e−(x +y 2 )
2
I2 = dx dy
−∞ −∞

An illustration of how Cartesian coordinates (x, y) are transformed into polar coordinates (r, θ ) is
shown in Fig. 11.6. We observe the position of point A in Cartesian system and note the relations

x = r cos θ y = r sin θ ∴ x 2 + y 2 = r and dx dy = r dr dθ

The geometrical interpretation shows that the mapping between unity differential surfaces
dx dy = r dr dθ is evident. In order to cover all available surfaces, the new variables are
bound as
 
x ∈ [−∞, ∞] r ∈ [−∞, ∞]
∴
y ∈ [−∞, ∞] θ ∈ [0, 2π ]

We continue with the derivation of Gaussian integral in polar coordinates as

; 2π ; ∞ ; 2π ; ∞ ; ∞
−r 2 −r 2
e−r r dr
2
I =
2
e r dr dθ = dθ e r dr = 2π
0 0 0 0 0
 ; 
1 2
e−r r dr = − e−r
2
t = −r 2 ∴ dt = −2r dr ∴
2
 % 1
1 2 %∞  −∞
*0
 
= 2
π − e  −
e−r % = −π  e
7
0
=π
2
0

∴
; ∞ √
e−x dx =
2
I= π
−∞
Solutions 261

Fig. 11.6 Example 11.5-1 f (x)

y + dy
dy
dr rd q
y
dx q + dq
r
r sin q
x
0 q
r cos q
0 x x + dx r r + dr

Exercise 11.6, page 249

1. Rational functions are decomposed by the partial fraction transformations.

;  
x−1 1 2
= partial fraction: − +
x2 + x x x+1
(a) ; ;
1 2
=− dx + dx = − ln |x| + 2 ln |x + 1| + C
x x+1
;  
x 1 4
= partial fraction: +
(x + 1)(x − 4) 5(x + 1) 5(x − 4)
; ;
1 4
(b) = dx + dx
5(x + 1) 5(x − 4)
1 4
= ln |x + 1| + ln |x − 4| + C
5 5
;  
x+2 1 1 1
= partial fraction: − − +
x 3 − 2x 2 x x2 x−2
; ; ;
1 1 1
(c) = − dx − dx + dx
x x2 x−2
1
= − ln |x| + + ln |x − 2| + C
x
;  
2x 2 1 1 1
= partial fraction: 2 − +
x4 − 1 x + 1 2(x + 1) 2(x − 1)
(d)
1 1
= arctan x − ln |x + 1| + ln |x − 1| + C
2 2
;  
1 1 x+2
= partial fraction: −
x3 − 1 3(x − 1) 3(x 2 + x + 1)
; ;
1 1 1 x+2
(e) = dx − dx
3 x−1 3 x +x+1
2
; ;
1 1 x 2 1
= ln |x − 1| − dx − dx
3 3 x +x+1
2 3 x +x+1
2
262 11 Integrals

(e) (cont.) The last two integrals we can solve are as follows:
;
x
dx
x2 +x+1
  2  2  
1 1 1 1 2 3
= x +x+1=x +2 x+
2 2
− +1= x+ +
2 2 2 2 4
;  
x 1 2t − 1
=  dx = t = x + ∴ dt = dx and x =
1 2 3 2 2
x+ +
2 4
; ; ;
2(2t − 1) 2t 1
= dt = 2 dt − 2 dt
4t 2 + 3 4t 2 + 3 4t 2 + 3
 
z = 4t 2 + 3 ∴ dz = 4 × 2t dt
;
1 1
= ln(4t 2 + 3) − 2 dt
2 4t 2 + 3
⎧ √ ⎫
⎪
⎪
1
=
1 1
= √
2t
∴ =
3 ⎪
⎪
⎪ 2
⎪  , n dn dt ⎪
⎪
⎨ 4t + 3
⎪ 3 2t 2 3 2 ⎪
⎬
√ +1
⎪ 3 ⎪
⎪
⎪ ; ⎪
⎪
⎪
⎪ 1 ⎪
⎪
⎩ so that = arctan n ⎭
n +1
2

1 1 2t
= ln(4t 2 + 3) − √ arctan √
2 3 3

1 1 2x + 1
= ln(4x 2 + 4x + 4) − √ arctan √
2 3 3

And, similarly,
; 
1 2 2x + 1
dx = √ arctan √
x +x+1
2
3 3

so that we write
;
1 1
= ln |x − 1|
x3 − 1 3

1 1 2x + 1
− ln(4x + 4x + 4) + √ arctan
2
√
6 3 3 3

4 2x + 1
− √ arctan √ +C
3 3 3

1 1 1 2x + 1
= ln |x − 1| − ln(4x + 4x + 4) − √ arctan
2
√ +C
3 6 3 3 3
Solutions 263

Exercise 11.7, page 250

1. Form of integrals that are solved by trigonometric substitutions.

; 
 
(a) 9 − x 2 dx = x = 3 sin θ ∴ dx = 3 cos θ dθ and x 2 = 9 sin2 θ
; 
= 9 − 9 sin2 θ 3 cos θ dθ
; 
= 3 1 − sin2 θ 3 cos θ dθ
; ;
=9 cos θ cos θ dθ = 9 cos2 θ dθ
 
1 + cos 2θ
cos2 θ =
2
; ; ;
9 9
= (1 + cos 2θ) dθ = dθ + cos 2θ dθ
2 2
9 1  
= θ + sin 2θ + C = sin 2θ = 2 sin θ cos θ
2 2
9
= [θ + sin θ cos θ ] + C
⎧ 2
x x⎫
⎪
⎪ = 3 sin θ ∴ sin θ = ∴ θ = arcsin ⎪ ⎪
⎨x 3 3⎬
√
⎪
⎪  9 − x2 ⎪
⎪
⎩3 cos θ = 9 − x 2 ∴ cos θ = ⎭
√ 3
9 x 9
x 9 − x 2
= arcsin + +C
2 3 2 3
3

9 x x
= arcsin + 9 − x2 + C
2 3 2
;
x2  
(b) √ dx = x = 6 sin θ ∴ dx = 6 cos θ dθ and x 2 = 36 sin2 θ
36 − x ;
2
36 sin2 θ
=  6 cos θ dθ
36 − 36 sin2 θ
; ;
36 sin2 θ 
= 

6 cos θ dθ = 36 sin2 θ dθ
cos
6
 θ
 
1 − cos 2θ
sin θ =
2
2
; ; ;
= 18 (1 − cos 2θ) dθ = 18 dθ − 18 cos 2θ dθ
264 11 Integrals

;
 
= 18θ − 18 cos 2θ dθ = y = 2θ ∴ dy = 2 dθ
 ; 
1 1
= 18 θ − cos y dy = 18 θ − sin 2θ
2 2
 
sin 2θ = 2 sin θ cos θ
= 18 (θ − sin θ cos θ ) + C
It is necessary to reintroduce the original variable x. From the trigonometric substitution it
follows.
x x 
x = 6 sin θ ∴ sin θ = ∴ θ = arcsin
6 6
√
 36 − x 2
36 − x 2 = 6 cos θ ∴ cos θ =
6

Therefore,
;
x2
√ dx = 18 (θ − sin θ cos θ ) + C
36 − x 2
x  √
x 36 − x 2
= 18 arcsin − 1
18 +C
6 6
6
2
x  x 
= 18 arcsin − 36 − x 2 + C
6 2

⎧  ⎫
⎪
⎪ u2 + k 2 substitution is u = k tan θ ⎪
⎪
⎪
⎪ ⎪
⎪
; √ ⎪
⎪ ⎪
⎪
25x 2 + 4 ⎨ 2 dθ ⎬
(c) dx = ∴ 5x = 2 tan θ ∴ dx =
x4 ⎪ 5 cos2 θ ⎪
⎪
⎪ ⎪
⎪
⎪
⎪  ⎪
⎪
⎪
⎩ 
4 2 ⎪
⎭
∴ 25x 2 + 4 =  25 tan2 θ + 4 =
; 

25 cos θ
2 2 dθ
=  4 2
cos θ 2 tan θ 5 cos θ
5
; ;
125 cos4 θ dθ 125 cos θ
= 
 4  2 = dθ
4
cos θ sin θ cos θ 4 sin4 θ
 
t = sin θ ∴ dt = cos θ dθ
125 1
=− +C
12 sin3 θ
Solutions 265

The original variable is reinserted as follows:

sin θ 5x 5x
5x = 2 tan θ ∴ tan θ = = ∴ sin θ = cos θ
cos θ 2 2
 2 2
25x 2 + 4 = ∴ cos θ = √
cos θ 25x 2 + 4
∴
5x 2

sin θ = √
2
25x 2 + 4

So that the final solution is

; √ √ 3
25x 2 + 4 125 1  
125 25x 2 + 4
dx = − +C =− +C
x4 12 sin3 θ 12 5
x
√
( 25x 2 + 4)3
=− +C
12x 3

Exercise 11.8, page 250

1. Area under the curve A is calculated as the definite integral bound of the given interval. Note that
when calculating finite integrals, the integration constants are cancelled.
(a) This is tabular integral,
; 3 ; 
x n+1
A= x 3 dx = x dx =
n
2 n
%
x 4 %%
3
1 4  1 65
= % = 3 − 24 = (81 − 16) =
4 2 4 4 4
The sign of area A under the curve is always positive, see Fig. 11.7.

Fig. 11.7 Example 11.8-1(a)

27 f (x)

18

9
A

0 x
0 1 2 3 4
266 11 Integrals

(b) This is sum of two tabular integrals,

; 2 ; 2 ; 2 %2 %2
x 2 %% x 5 %%
A= x(1 + x ) dx =
3
xdx + x dx =
4
+
−1 −1 −1 2 %−1 5 %−1
1 2 1 3 33 81
= [2 − (−1)2 ] + [25 − (−1)5 ] = + =
2 5 2 5 10
Note that surface area is “signed area addition”, i.e. in the absolute sense the small negative area
in interval [−1, 0] is subtracted from the large positive area in interval [0, 1], see Fig. 11.8.

Fig. 11.8 Example 11.8-1(b)

18 f (x)

+
0 x
−1 0 1 2

(c) This is tabular integral (Fig. 11.9),

; π/4 %π/4
% π −π  
cos x dx = sin x %% = sin − sin = sin(−x) = − sin(x)
−π/4 −π/4 4 4
√
π 2 √
= 2 sin = 2
= 2
4 2

Fig. 11.9 Example 11.8-1(c)
f (x)
1

x
0

−p /2 0 p /2
Solutions 267

(d) This is tabular integral,

; 2π %2π
%
cos x dx = sin x %% = sin(2π ) − sin(0) = 0
0 0

Note that the positive and negative areas of sine form function are perfectly matched over one
period, and thus the sum is zero, see Fig. 11.10.

Fig. 11.10
f (x)
Example 11.8-1(d) 1

+ +
0 x

−1
0 p 2p

(e) This integral is solved by using the integration by substitution technique (change of
variables), ⎧ ⎫
; 2 ⎪
⎨ dt ⎪
⎬
x t = 1 + 3x 2
∴ dt = 6x dx ∴ x dx =
√ dx = 6
0 1 + 3x 2 ⎪
⎩x = 0 ⇒ t = 1, and x = 2 ⇒ t = 13 ⎪
⎭
; 13  
1 dt √ 1 −n
= √ = n
x =x , n =x
m m/n

1 6 t x
; 13 %
1 −1 1 t %%
1/2 13
1 √ √
= t /2 dt = % = [ 13 − 1]
6 1 6
3 1/2 1 3
Fig. 11.11
Example 11.8-1(e) f (x)

0.5

0 x
0 1 2

Note that surface area equals an irrational number smaller than one, see Fig. 11.11.
268 11 Integrals

(f) This integral is solved by using the partial integration technique two times.
; π ; ; 
x 2 cos x dx = u dv = u v − v du
0
⎧ ⎫
⎪ du ⎪
⎪
⎨u = x ⇒ dx = 2x ∴ du = 2x dx
2 ⎪
⎬
= ; ;
⎪
⎪ ⎪
⎩dv = cos x dx ⇒ dv = cos x dx ∴ v = sin x ⎪
⎭
;
= x 2 sin x − 2 x sin x dx
⎧ ⎫
⎪
⎪ u = x ⇒ du = dx ⎪
⎪
⎪
⎪ ; ; ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ dv = sin x dx ⇒ dv = sin x dx ∴ v = − cos x ⎪
⎬
=
⎪
⎪ ∴ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ ; ; ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎩ x sin x dx = −x cos x + cos x dx = −x cos x + sin x ⎪
⎭
π
= x 2 sin x − 2(−x cos x + sin x)
0
π
= (x 2 − 2) sin x + 2x cos x)
0

π
sin
= (π 2 − 2) : 0 − (0 − 2) :

sin
0 : −1 − 2(0)
cosπ
0 + 2π  
:0
 cos 0

= −2π
The negative result is illustrated in Fig. 11.12, where it is clearly visible that the negative signed
area under the curve dominates.

Fig. 11.12
Example 11.8-1(f)
f (x)
0 x
+

−1

0 p /2 p

2. In order to calculate the area bound by two functions, e.g. f (x) and g(x), it is necessary
to determine the applicable interval. In this case, we can visualize f (x) “cutting out” g(x)
and removing the overlapping area, see Fig. 11.13 (left), which leaves area that represents the
Solutions 269

difference between the two, Fig. 11.13 (right). The interval boundaries are found at the cross-over
points A and B, i.e. when

f (x) = g(x) ∴ x 2 = 8 − x 2
2x 2 = 8
x 2 = 4 ∴ x = ±2

Therefore, the “cutting out” operation is a simple difference as

; 2 ; 2 ; 2 ; 2
A= [g(x) − f (x)] dx = [8 − x 2 − x 2 ] dx = 8 dx − 2 x 2 dx
−2 −2 −2 −2
%2 %2
% x 3 %% 2 64
%
= 8x % − 2 % = 8(2 − (−2)) − (23 − (−2)3 ) =
−2 3 −2 3 3

Fig. 11.13 Example 11.8-2

3. Given rational function

x
f (x) =
(x + 1)(x − 4)

and the interval of interest, we use the partial fraction decomposition technique, see Exam-
ple 11.6-1(b).

x A B A(x − 4) + B(x + 1) Ax − 4A + Bx + B
= + = =
(x + 1)(x − 4) x+1 x−4 (x + 1)(x − 4) (x + 1)(x − 4)
(A + B)x − 4A + B
=
(x + 1)(x − 4)
∴
A + B = 1 and − 4A + B = 0 ∴ B = 1 − A and B = 4A
∴
270 11 Integrals

1 4
A= et B =
5 5

Introduce two variables: t = x + 1 and r = x − 4; therefore, their respective intervals are:

t ∈ (6, 7) and r ∈ (1, 2):

; 6 ; 6 ; ;
x 1 6 dx 4 6 dx
f (x) dx = dx = +
5 x − 3x − 4 5 5 x+1 5 5 x−4
2
5
; 7 ; 2 %7 %2
1 dt 4 dr 1 % 4 %
= + = ln(t)% + ln(r)%%
%
5 6 t 5 1 r 5 6 5 1

1 4 0
= (ln 7 − ln 6) + ln 
ln 2 −  *
1
5 5
∴
(the answer may be in any of the following forms)

1 1 7 × 24 1 56 56
= (ln 7 − ln 6 + 4 ln 2) = ln = ln = ln
5

5 5 6 5 3 3
= 0.58534...

Graph illustrating the solution is in Fig. 11.14. Note that by visual inspection of the graph, we can
confirm that the calculated surface area is indeed approximately 0.6 or so.

Fig. 11.14 Example 11.8-3 1

f (x)

0.5

0
5 6
x
Solutions 271

4. The well known high school-level geometry equations are derived by using calculus.
(a) A right-angled triangle is easily mapped into the coordinating system where the two catheti
a = 2 and b = 1 are assumed to be the horizontal and vertical coordinates, Fig. 11.15. In this
respect, obviously, hypotenuse is a linear function whose algebraic form is found by knowing that
the line must cross two points whose coordinates are (0, 0) and (2, 1). We write two equations as

f (x) = ax + b
(y, x) = (0, 0) : 0 = a × 0 + b ∴ b = 0
1
(y, x) = (2, 1) : 1 = a × 2 ∴ a =
2
1
∴ f (x) = x, x ∈ [0, 2]
2

Therefore, the area of triangular surface is calculated as

; ; %2
2 2
1 1 x 2 %% 1
A= f (x) dx = x dx = = (4 − 0) = 1
1 1 2 2 2 %1 4

Of course, the high school formula that gives A = ab/2 = 2 × 1/2 = 1 is a simple summary of
the calculus method.

Fig. 11.15
f (x)
Example 11.8-4(a) 1
1 x
2
)=
f (x

x
0
0 2

(b) One possible technique to calculate the circle area is as follows. From algebraic equation for
circle, given r = 1, we write

x2 + y2 = r 2 ∴ x2 + y2 = 1 ∴ y = 1 − x 2 = f (x), x ∈ [0, 1], y ∈ [0, 1]

The plot of f (x), see Fig. 11.16, shows a surface that is equivalent to one-quarter of the full circle
area. Thus, at the end, the result of the definite integral of f (x), x ∈ [0, 1] must be multiplied by
four.
272 11 Integrals

Fig. 11.16
Example 11.8-4(b) f (x)
1

x
0

0 1

This integral is solved by using the integration by substitution technique (change of variables)
two times. The idea is that in a unity circle, the horizontal projection equals x = r cos θ = cos θ,
because r = 1.
; 1 ; 1 
A4 = f (x) dx = 1 − x 2 dx
0 0
⎧ dx ⎫
⎪
⎨x = cos θ ∴ = − sin θ ∴ dx = − sin θ dθ ⎪
⎬
= dθ
⎪
⎩x = 0 ⇒ θ = π , x = 1 ⇒ θ = 0 ⎪
⎭
2
; 0
= 1 − cos2 θ (− sin θ) dθ
π/2

.  /
= sin2 x + cos2 x = 1 ∴ sin x = 1 − cos2 x
; 0 ; π/2  
1
= − sin2 θ dθ = sin2 θ dθ = sin2 x = (1 − cos(2x))
π/2 0 2
; π/2 ; π/2 ; π/2
1 1 1
= (1 − cos(2θ)) dθ = dθ − cos(2θ) dθ
2 0 2 0 2 0

we introduce second change of variables and continue as

⎧ dt ⎫
⎪
⎨t = 2θ ∴ dt = 2dθ ∴ dθ = ⎪
⎬ %π ;
2 1 %% /2 1 π
dt
A4 = = θ% − cos t
⎩θ = 0 ∴ t = 0, θ = π ∴ t = π ⎪
⎪ ⎭ 2 0 2 0 2
2
%π
1 π 1 % π 1   0 π
sin π: −  :

0
= −0 sin t %% = −  sin0 =
2 2 4 0 4 4 4
π
∴ A = 4A4 = 4A = π
4A
Solutions 273

The high school-level formula summarizes the above calculus result by stating A = r 2 π =
(r = 1) = π .

(c) Calculation of volumes is done by “integral of an integral” or multiple integral, in other

words, one integration for each variable. Visually, a surface is created by adding infinitely many
lines in parallel next to each other, where the respective length of each line equals f (x) (i.e. first
dimension) and the lines are added within the given range (i.e. second dimension).
For example, the lowest horizontal surface, see Fig. 11.17, is created by aligning y-directional
lines (f (x)) within range of x ∈ [0, 4], where each line has the length equal to three. In other
words, the surface area A1 is found as
; ; %4
4 4 %
A1 = f (x) dx = 3 dx = 3 x %% = 3(4 − 0) = 12
0 0 0

which is the result that we know from the high school formula for rectangle surface A1 = ab =
3 × 4 = 12. Visually, this surface is equivalent to one page of a book whose volume we want to
calculate. That is to say, we add pages on top of each other in z-direction. Thickness of surface is
infinitely small, that is to say it is equal to dz. We have to simply “pile them up” between zero and
two. Mathematically, this addition of surfaces is another integral in z-direction where z ∈ [0, 2].
This integral (i.e. addition) of surfaces is written as
; %2
2  % 
V = A1 dz = A1 = 12 = const. = 12 z %% = 12(2 − 0) = 24
0 0

which is a well known result for cuboid volume V = abc = 4 × 3 × 2 = 24.

This process of double integration is formally written in one line as
; 2 ; 4 ; 2 ; 4 ; 2 ; 4
V = f (x) dx dz = 3 dx dz = 3 dz dx
0 0 0 0 0 0
; %4 ; %2
2 % 2 %
=3 dz x %% = 3 4 dz = 12 z %% = 12 × 2 = 24
0 0 0 0

Note that the order of addition is not relevant, and thus we choose most convenient one. In
addition, this trivial example illustrates the principle and way to reason the volume calculations
in 3D space. Nevertheless, the same reasoning and techniques apply to higher order spaces and
non-trivial objects.
Fig. 11.17
Example 11.8-4(c)
z
2

(0, 0, 0) 4 x
A1
3
f (x) = 3
y
274 11 Integrals

5. One possible geometric interpretation of “average” may be by comparing (signed) area under the
function with the area of a rectangle in the same interval. Surface area under function is calculated
with the finite integral. Surface area of rectangle A equals its length a (equal to the interval)
multiplied by its height h. Therefore, the height of the rectangle is calculated by dividing its area
by its length. If the two surface areas happen to be equal, then the height h of the rectangle is
called the “average” of the function.
For example, see Fig. 11.18 (left), by inspection we find the area under piecewise linear function
f (x) in x ∈ [0, 4] interval to be equal to 0 + 2 + 0 + 2 = 4 units. At the same time, rectangular
surface in Fig. 11.18 (right) in x ∈ [0, 4] interval is A = a × h = 4 × 1 = 4. Thus, in this special
case, we write
; 4
A 1
h= = f (x) dx
a 4 0

In this case, we say that h is the average value of f (x) and we use syntax f (x) to convey that
information. In general case, i.e. any continuous function f (x) in x ∈ [a, b], we write the formal
definition of average as
; b
1
f (x) = f (x) dx
b−a
       a  
height length surface area

Fig. 11.18 Example 11.8-5 f (x) f (x)

2 2

average,  f (x) = 1
x x
0 0
0 1 2 3 4 0 1 2 3 4

(a) Average of a periodic function as calculated over one period is

; %2π
1 2π
1 % 1
sin x = sin x dx = (− cos x) %% = (−(−1) − 1) = 0
2π 0 2π 0 2π

To interpret result that sin x = 0, x ∈ [0, 2π ], we look at Fig. 11.19. Within one period of sine
function, positive and negative surface areas are perfectly matched and thus cancelled to zero.
That is to say, the height of the equivalent rectangle equals zero. The same conclusion is valid for
cos x = 0, x ∈ [0, 2π ], Fig. 11.19 (right).
Solutions 275

Fig. 11.19 f (x)

f (x)
Example 11.8-5(a) 1 1

+ + +
0 x 0 x

−1 −1
0 p 2p 0 p 2p

(b) Average of sine and cos functions over interval, for example from π/2 to 3p/2 (i.e. half-
period long, but only over negative surface section), is
; %3π/2
1
3π/2
1 % 1 2
cos x = cos x dx = sin x %% = (−1 − 1) = −
π π/2 π π/2 π π

which is not equal to zero. However, see Fig. 11.19 (right), we find that the average cos x = 0 if
x ∈ [0, π ], which is also a half-period long but over both positive and negative surface sections.

(c) Average over one period of a sine function (“AC”) that is added to a constant (“DC”) is found
as follows:

*0
1
; 2π
1
; 2π
1
; 2π 
f (x) = (1 + cos x) dx = dx +  
cos x dx
2π 0 2π 0 2π 0
%2π
1 % 1
= x %% = (2π − 0) = 1
2π 0 2π

This is an important case to notice, and the graphical interpretation is shown in Fig. 11.20.

Fig. 11.20
2 f (x) 2 f (x)
Example 11.8-5(c)

1 1

x x
0 0
0 p 2p 0 p 2p
276 11 Integrals

(d) Integral that is solved by the change of variable technique in addition to using an important
trigonometric identity transformation.
; 2π  
1 1
f (x) = cos2 x dx =
cos2 x = (1 + cos(2x))
2π0 2
; 2π %2π ; 2π
1 1 1 % 1 1
= (1 + cos(2x)) dx = x %% + cos(2x) dx
2π 0 2 4π 0 2π 0 2
⎧ ⎫
⎪
⎨t = 2x ∴ dt = 2 dx ∴ dx = dt ⎪
⎬
= 2
⎩x = 0 ∴ t = 0 and x = 2π ∴ t = 4π ⎪
⎪ ⎭
; 4π
 :0
1 1 1 1 : 0)
= (2π − 0) + cos t dt = + (
sin(4π ) − 
sin(0)
4π 4π 0 2 4π
1
=
2

Due to symmetrical regions of this function, similar to the illustration in Fig. 11.20 (right) again,
we can see how the area under f (x) is “rearranged” into rectangle whose height equals one-half,
Fig. 11.21.

Fig. 11.21
f (x)
Example 11.8-5(d) 1

0 x
0 S 2S

6. By inspection of Fig. 11.2, we calculate surface area divided by the interval as

(a) 0/(1 − 0) = 0 (b) 2/(2 − 0) = 1 (c) 2/(3 − 0) = 2/3
(d) 4/(4 − 0) = 1 (e) 2/(2 − 1) = 2 (f) 0/(3 − 2) = 0
(g) 2/(4 − 2) = 1 (h) 2/(3 − 0.5) = 4/5 (i) 2/(3.5 − 1.5) = 1

7. By inspection of Fig. 11.3, we calculate surface area divided by the interval for the top and bottom
functions to find
(a) f  = 0, g = 0 (b) f  = 1, g = 0
(c) f  = 3/2, g = 3/2 (d) f  = 1, g = 1
(e) f  = 1, g = 4/5 (f) f  = 1, g = 1
(g) f  = 8/7, g = 8/7 (h) f  = 6/7, g = 6/7
(i) f  = 1, g = 1
Solutions 277

8. By inspection of Fig. 11.4, we calculate surface area divided by the interval for the top and bottom
functions to find
(a) f  = 0, g = 2 (b) f  = 1, g = 1
(c) f  = 2/3, g = 4/3 (d) f  = 1/2, g = 3/2
(e) f  = 2/5, g = 8/5 (f) f  = 1/2, g = 3/2
(g) f  = 4/7, g = 10/7 (h) f  = 1/3, g = 5/3
(i) f  = 0, g = 2

Exercise 11.9, page 252

1. In case of unbound functions, it is necessary to find if they converge or diverge. The idea is to see
if the function’s limit is finite or infinite. Definite integrals give the surface area under function,
and thus if the function reaches infinite value, the surface area is also infinite (i.e. diverge). It is
not always obvious to determine the function’s behavior by simple inspection.
; a  %  
1 1 %%a 1 1
(a) lim dx = lim − = lim − − −
a→∞ 1 x 2 a→∞ x %1 a→∞ a 1
= (0 − (−1)) = 1
; a  %a
1 %
(b) lim dx = lim ln |x|%% = lim (ln a − ln 1)
a→∞ 1 x a→∞ a→∞
1
= ∞ − 0 = ∞ ∴ diverges

The difference between functions 1/x and 1/x 2 is illustrated in Fig. 11.22. Although, by
inspection, both surface areas seem to converge, it is clearly visible that their respective integrals
behave very differently: ln x is divergent and 1/x is convergent for x → ∞.

Fig. 11.22
Example 11.9-1(a)(b) 2 1/x
1/x2
ln x
−1/x

0
x

0 2 4 6
278 11 Integrals

(c) This function is not defined for x = 0, because f (0) = ∞. Therefore, we evaluate its
integral as two improper integrals, one approaching zero from the left side and one approaching
zero from the right side.
; a ; 1
1 1
lim √ dx + lim √ dx
a→0− −1 3 x 2 a→0+ a 3
x2
; a ; 1
−2/3
= lim x dx + lim x −2/3 dx
a→0− −1 a→0+ a
% %
√ %a √ %1
= 3 lim 3 x %% + 3 lim 3 x %%
a→0− −1 a→0+ a
√ √  √ √ 
= 3 lim a − −1 + 3 lim 1− 3 a
3 3 3

a→0− a→0+
√   √ 
=3 0+1 +3 1− 0
3 3

=6
In this case, even though there is vertical asymptote within the given interval x ∈ [−1, 1], the
surface area converges to 3 on each side, Fig. 11.23.

Fig. 11.23
3
Example 11.9-1(c) f (x)

0 f (x)
f (x)dx x
−1 0 1

; %a  2 
2%
a
1 1
xe−x dx = lim − e−x %% = − lim e−a − e−(−a)
2 2
(d) lim
a→∞ −a a→∞ 2 −a 2 a→∞
1
= − (0 − 0) = 0
2
Solutions 279

; %0
0
1 √ %
(e) lim √ dx = lim −2 1 − x %
a→−∞ a 1−x a→−∞
√ a

√
= −2 lim 1−0− 1−a
a→−∞

= −2(1 − ∞) = ∞ ∴ diverges
Even though, by a quick look inspection, it may appear that area under f (x) converges, its integral
diverges, Fig. 11.24.

Fig. 11.24
Example 11.9-1(e) f (x )
x
0

−2

f (x)
−4 
f (x)dx

−4 −2 0

; a
1
(f) lim , dx = lim arctan x|a0 = lim (arctan a − arctan 0)
a→∞ 0 1 + x2 a→∞ a→∞
π π
= −0=
2 2

Exercise 11.10, page 252

1. Knowing that definite integrals give the total signed surface area within the given interval,
sometimes we can intuitively deduce the correct result.
(a) Inverse function has vertical asymptote for x = 0; thus, it is logical to calculate limit a → 0
after applying the Newton–Leibniz formula; that is to say,
; ; ; %a %1
−1
1 a
1 1
1 % %
dx = lim dx + lim dx = lim ln |a|% + lim ln |a|%%
%
1 x a→0− −1 x a→0+ a x a→0− −1 a→0+ a

= lim (ln |a| − ln | − 1| + ln |1| − ln |a|) = −∞ − 0 + 0 − ∞

a→0

∴
(undermined, incorrect conclusion)
Mathematical software that blindly implements the Newton–Leibniz formula reports this result.
However, the inverse function is odd, that is to say the origin-symetric, Fig. 11.25. One should
conclude that the total sum is zero. Indeed, the evaluation of undermined result should be
modified as
280 11 Integrals

; −1
1 :0

| −1| + 
ln 
dx = lim (ln |a| −  ln * 0 − ln |a|)

|1|
1 x a→0

|a|

= lim ln  = lim ln 1 = 0
a→0 |a|
 a→0
∴
(correct)

Fig. 11.25
Example 11.10-1(a) f (x )

0 x

f (x)

f (x)dx

−1 0 1

(b) Periodic functions may present a challenge as well. Direct implementation of the Newton–
Leibniz formula may lead into wrong conclusion.
; .
100π √ x/
1 − cos 2x dx = 1 − cos x = 2 sin2
0 2
; % %
√ 100π √ %; 100π %
= 2 | sin x| dx = 2 %% sin x dx %%
0 0
% %100π %%
√ %% % % √
= 2 % cos x %% % = 2 |cos(50 × 2π ) − cos 0|
% 0 %
√
= 2 |1 − 1| = 0
However, by inspection, area under f (x) is obviously non-zero and positive ( | sin x| ≥ 0), and
it is periodic with the period T = π , see Fig. 11.26. One way to make correct conclusion is to
calculate the surface area within one period.
Solutions 281

; % %
π √ √ %; π %
1 − cos 2x dx = 2 %% sin x dx %%
0 0
% %π %
√ % % % √
= 2 % cos x %% %% = 2 |cos(π ) − cos 0|
%
0
√ √
= 2 | − 1 − 1| = 2 2
√
Therefore, the surface area within interval x ∈ [0, 100π ] equals 200 2.

Fig. 11.26
Example 11.10-1(b) f (x)
1

0 x
0 S
 Multivariable Functions
12

Important to Know

Multivariable functions are in the form f (x, y, z, . . . ).

Limits:

lim f (x, y, z, . . . )
x→x0 ,
y→y0 ,
z→z0 ,
...

exists if all variable limits converge to the same value; otherwise, the limit does not exist. In its basic
form, for example, the following limits converge to the same value (note, the limit is calculated along
one variable at the time, while all the others are kept constant):

lim f (x, y, z, . . . ) = a
x→x0 ,
y0 ,
z0 ,
...

lim f (x, y, z, . . . ) = a
x0 ,
y→y0 ,
z,
...

lim
x ,
f (x, y, z, . . . ) = a
0
y0 ,
z→z0 ,
...

···

Note that this condition is necessary but not sufficient to prove the existence of the global limit.

Partial Derivatives: these are calculated for each variable relative to all the others. For example, in the
case of a two-variable function f (x, y), we can use the following equivalent syntax form notations to
specifically write the order of second partial derivatives as

   ∂f ∂f ∂ 2f
fx x = fxx = =
∂x ∂x ∂x 2

© Springer Nature Switzerland AG 2021 283

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_12
284 12 Multivariable Functions


   ∂f ∂f ∂ 2f
fx y = fxy = =
∂y ∂x ∂y∂x

   ∂f ∂f ∂ 2f
fy x = fyx = =
∂x ∂y ∂x∂y

   ∂f ∂f ∂ 2f
fy y = fyy = =
∂y ∂y ∂y 2

Integrals: the volume of multidimensional space is calculated as integral along each variable (the
integral sum is a commutative operation). For example, in the case of three-dimensional space, we
systematically reduce the number of variables as
;;;
V3D = f (x, y, z) dx dy dz
x,y,z
;; ;
= dy dz f (x, y, z) dx
y,z
x  
f (y, z)
; ;
= dz f (y, z) dy
z y
  
f (z)
;
= f (z) dz
z

where, in the each integration step along given variable, the reminding variables are treated as
constants.

12.1 Exercises

12.1 *** Domain of Multivariable Functions

1. Sketch plot of the following surfaces:

(a) f (x, y) = x 2 + y 2 (b) f (x, y) = x 2
(c) r 2 = x 2 + y 2 + z2 (d) z2 = x 2 + y 2

2. Determine domains of the following functions:

√
(a) f (x, y) = x + y. (b) f (x, y) = ln(−x − y).
 √ 
(c) g(x, y) = 9 − x 2 − y 2 . (d) h(x, y) = 1 − x 2 + y 2 − 1.
12.1 Exercises 285

12.2 *** Limits of Multivariable Functions

1. Calculate the following limits:

lim (x 2 y 3 − x 3 y 2 + 3x + 2y) x2 − y2
(a) (b) lim
(x,y)→(1,2) (x,y)→(0,0) x 2 + y 2

xy 3x 2 y
(c) lim (d) lim
(x,y)→(0,0) x 2 + y2 + y2
(x,y)→(0,0) x 2

12.3 *** Derivatives of Multivariable Functions

1. (a) Derive partial derivatives of f (x, y) = 4−x 2 −2y 2 and then calculate f  (1, 1) and f  (1, 1).
x y
x
(b) Derive partial derivatives of f (x, y) = sin .
1+y
(c) Given x 3 + y 3 + z3 + 6xyz = 1, derive ∂z
∂x
and ∂z
∂y
.
   
(d) Given function f (x, y) = x 3 + x 2 y 3 − 2y 2 , derive fxx , fxy , fyx , fyy .

2. Derive partial derivatives of the following composite functions:

(a) Derive partial derivatives of z = ex sin y, where x = st 2 and x 2 t.

(b) Given u = ln (x − a)2 + (y − b)2 , where a and b are constants, prove the following
equation:

∂ 2u ∂ 2u
+ 2 =0
∂x 2 ∂y

12.4 *** Integrals of Multivariable Functions

1. ; 3 ; 2
(a) Calculate the following integral: x 2 y dx dy
0 1
;; 
(b) Calculate the following integral: 1 − x 2 dA where R is bound of the two variables
R
forming surface A, given that |y| ≤ 2.
286 12 Multivariable Functions

Solutions

Exercise 12.1, page 284

1. Two-variable functions define surfaces whose elevation is measured along z-axis.

(a) Starting with a point at (x, y) = (0, 0), (b) Function f (x, y) = x 2 does not depend
function f (x, y) = x 2 + y 2 generates a circle on y, and thus when looking into y-axis there
whose radius is larger and larger for each sub- is only quadratic function z = x 2 visible.
sequent (x, y) pair and its value is assigned to However, this quadratic function is found at any
z-axis, Fig. 12.1. position along y ∈ [−∞, ∞], Fig. 12.2.

x y

Fig. 12.2 Example 12.1-1(b)

Fig. 12.1 Example 12.1-1(a)
Solutions 287

(c) As the next expansion of 2D circle equa- (d) Two-sided conus function z2 = x 2 + y 2
tion, function r 2 = x 2 +y 2 +z2 defines a sphere also starts with a single point at (x, y) = (0, 0)
whose radius equals r, Fig. 12.3. and then progressively moves in both directions
of z-axis, while the radius of circles changes
linearly, Fig. 12.4.

Fig. 12.3 Example 12.1-1(c)

Fig. 12.4 Example 12.1-1(d)

2. Domain of a multivariable function is limited by the limitations of each variable.

√
(a) Function f (x, y) = x + y is defined for x ∈ R and y ≥ 0.
(b) Function f (x, y) = ln(−x − y) is limited by ln t, and thus −x − y > 0 ∴ x + y < 0.


(c) Given g(x, y) = 9 − x 2 − y 2 , we conclude
 % 
%
D = (x, y) %% 9 − x 2 − y 2 ≥ 0
 % 
% 2
%
∴ (x, y) % x + y ≤ 9
2

Therefore, in x, y plane, g(x, y) is contained

 within a circle whose radius is r = 3. At the same
time, z is limited only to z ≥ 0 and 9 − x 2 − y 2 ≤ 3 (i.e. when (x, y) = (0, 0)), that is to say,
0 ≤ z ≤ 3 or, in other words, the upper half of a sphere.
288 12 Multivariable Functions

(d) Function
 
h(x, y) = 1 − x 2 + y 2 − 1
is limited by two square roots, Fig. 12.5, and thus we write

1 − x 2 ≥ 0 ∴ x 2 ≤ 1 ∴ |x| ≤ 1
y 2 − 1 ≥ 0 ∴ y 2 ≥ 1 ∴ |y| ≥ 1

Fig. 12.5 Example 12.1-2(d)

Exercise 12.2, page 285

1. In the first iteration, limits of a multivariable function are calculated along each variable, in the
given examples along (x, y). However, in order to be defined (i.e. the surface is continuous in any
direction), all limits along any arbitrary directions must be equal.

(a) lim (x 2 y 3 − x 3 y 2 + 3x + 2y) = 12 23 − 13 22 + 3(1) + 2(2) = 11

(x,y)→(1,2)

x2 − y2 0−0
(b) lim = ∴ does not exist
(x,y)→(0,0) x 2 + y 2 0+0
We can confirm this conclusion by the two limits along x-axis and y-axis as

x 2 − 02 x2
lim = =1
(x,y)→(x,0) x 2 + 02 x2
02 − y 2 −y
2
lim = = −1
(x,y)→(0,y) 02 + y 2 2
y

These two limits are not equal, and thus we conclude that in this case lim(x,y)→(x,0) f (x, y) does
not exist.
Solutions 289

(c) However, even if the two limits along x- and y-axes are equal, it is necessary but not
sufficient condition. For example,

xy x (0)
lim = 2 =0
(x,y)→(x,0) x2 +y 2 x +0
xy (0) y
lim = =0
(x,y)→(0,y) x2 + y2 0 + y2
however, limit along y = x is

xy xx x2 1
lim = 2 = =
(x,y)→(x,x) x 2 +y 2 x +x 2
2x 2 2

therefore, even though the first the two limits are equal, they are different from the third, and thus
the limit does not exist.

(d) This limit appears to not exist, because


3x 2 y 0 3x 2 y 0
lim = however, lim = 2 =0
(x,y)→(0,0) x 2 + y 2 0 (x,y)→(x,0) x 2 + y 2 x
3x 2 y 0
lim = 2 =0
(x,y)→(0,y) x 2 + y 2 y
3x 3 3x
lim = lim =0
(x,y)→(x,x) 2x 2 (x,y)→(0,0) 2

Even though the three limits are equal, the question is how to prove that this limit exists along any
arbitrary line? We can use “the squeeze theorem” (sometimes known as “the two police officers
and burglar” or “the sandwich theorem”). The idea is to determine two extreme boundaries of a
given function and then to bring these two boundaries together; the function itself must always
stay in between the two extremes. In the given analogy, if a burglar is squeezed between two
officers, and if the two officers are walking towards prison building, the burglar is forced to end
up in the prison.

Since we are looking for lim(x,y)→(0,0) f (x), we express the distance between the origin point
(0, 0) and any point of f (x) as
% % % %
% 3x 2 y % % 3x 2 y % 3x 2 |y|
% − %=% %
% x2 + y2 0 % % x2 + y2 % = x2 + y2 (12.1)

where the last absolute value expression is valid because number “3”, x 2 , and y 2 are always
positive, and thus they are equal to their respective absolute values. Consequently, the lower side
limit of (12.1) is zero.
290 12 Multivariable Functions

Accordingly, because x 2 ≤ x 2 + y 2 is always true, the upper side limit is

≤1
x2 x 2
>
x2 ≤ x2 + y2 ∴ ≤ 1 ∴  3 |y| ≤ 3 |y|
x2 + y2 x2 + y2

In conclusion, two extremes of (12.1) are on the lower side it is limited by “0” and on the upper
side by 3|y|, that is to say,

3x 2 y
0≤ ≤ 3 |y|
x2 + y2
∴
3x 2 y
lim 0≤ lim ≤ lim 3 |y| (12.2)
(x,y)→(0,0) (x,y)→(0,0) x 2 + y 2 (x,y)→(0,0)

Now, the left and right side inequalities in (12.2) are heading towards the origin, i.e. (x, y) →
(0, 0); then, by the virtue of its position between the two extremes, the distance is “squeezed”
between two limits and must also converge to zero, Fig. 12.6.
In summary, the limit exists in any direction as

3x 2 y
lim =0
(x,y)→(0,0) x 2 + y 2

Fig. 12.6 Example 12.1-1(d)

Exercise 12.3, page 285

1. Partial derivatives are calculated as follows:

(a) Calculate derivates relative to each variable and simply substitute given coordinates.

∂
fx = (4 − x 2 − 2y 2 ) = −2x ∴ fx (1, 1) = −2
∂x
∂
fy = (4 − x 2 − 2y 2 ) = −4y ∴ fx (1, 1) = −4
∂y
Solutions 291

(b) Derivatives of composite fractions are found as


∂ x x ∂ x 1 x
fx = sin = cos = cos
∂x 1+y 1 + y ∂x 1+y 1+y 1+y

∂ x x ∂ x x x
fy = sin = cos =− cos
∂y 1+y 1 + y ∂y 1+y (1 + y)2 1+y

(c) Implicit partial derivation is done by holding all variables constant except the derivative
variable. Note the derivatives of a two functions product.

∂z  3  ∂z ∂z
x + y 3 + z3 + 6xyz = 1 ∴ 3x 2 + 3z2 + 6yz + 6xy =0
∂x    ∂x ∂x
f (x)f (z)

∂z  2 
∴ 3
z + 6
2xy = −3
x 2 − 6
2yz
∂x
∂z x 2 + 2yz
∴ =− 2
∂x z + 2xy

∂z  3  ∂z ∂z
x + y 3 + z3 + 6xyz = 1 ∴ 3y 2 + 3z2 + 6xz + 6xy =0
∂y    ∂y ∂y
f (y)f (z)

∂z y 2 + 2xz
∴ =− 2
∂y z + 2xy

(d) Second derivatives may be with the same variable as the first derivative or mixed.
 
fx x 3 + x 2 y 3 − 2y 2 = 3x 2 + 2xy 3
 
fy x 3 + x 2 y 3 − 2y 2 = 3x 2 y 2 − 4y

 ∂fx  ∂f 
fxx = = 6x + 2y 3 ; fxy = x = 6xy 2
∂x ∂y


∂fy 
∂fy
fyx = = 6xy 2 ; fyy = = 6x 2 y − 4
∂x ∂y

2. Relative to each variable, derivatives of composite and parametric functions follow the same rules
and techniques as for single variable functions.
292 12 Multivariable Functions

(a) Given parametric form of function z = ex sin y, where x = st 2 and x 2 t, we find its
derivatives as follows:

∂z ∂z ∂x ∂z ∂y
= + = ex (sin y) t 2 + ex (cos y) 2st
∂s ∂x ∂s ∂y ∂s
2 2
= t 2 est (sin s 2 t) + 2st est (cos s 2 t)
∂z ∂z ∂x ∂z ∂y
= + = ex (sin y) 2st + ex (cos y) s 2
∂t ∂x ∂t ∂y ∂t
2 2
= 2st est (sin s 2 t) + s 2 est (cos s 2 t)

(b) Given

u = ln (x − a)2 + (y − b)2

we find

∂u 1 ∂ 
= (x − a)2 + (y − b)2
∂x (x − a)2 + (y − b)2 ∂x
1 1
= 
(x − a)2 + (y − b)2 2 (x − a)2 + (y − b)2
∂ # $
× (x − a)2 + (y − b)2
∂x
1 1
=  [2(x − a)]
(x − a)2 + (y − b)2 2 (x − a)2 + (y − b)2
x−a
=
(x − a)2 + (y − b)2
∴
∂ 2u −(x − a)2 + (y − b)2
= # $2
∂x 2 (x − a)2 + (y − b)2

Similarly,

∂u y−a ∂ 2u (x − a)2 − (y − b)2

= ∴ = # $2
∂y (x − a)2 + (y − b)2 ∂y 2 (x − a)2 + (y − b)2

Therefore,

∂ 2u ∂ 2u −(x − a)2 + (y − b)2 (x − a)2 − (y − b)2

+ = # $ 2
+ # $2 = 0
∂x 2 ∂x 2 (x − a)2 + (y − b)2 (x − a)2 + (y − b)2
Solutions 293

Exercise 12.4, page 285

1. Fundamentally, integration is an addition, and thus it is commutative operation and the order of
integration variables can be changed as necessary.
; 3; 2 ; 3 ; 2 ; 3 %2 ;
y 2 %% 3 3 2
(a) x 2 y dx dy = x 2 dx y dy = x 2 dx = x dx
0 1 0 1 0 2 %1 2 0
%3
3 x 3 %% 3
27 27
= = =
2 3 %0 2 3
2

(b) Double integral delivers surface area bound by the integral limits. Thus, given
;; 
1 − x 2 dA (12.3)
R

and interval |y| ≤ 2, it is necessary to determine the bounds as well as the shape√
of this surface.
Operation of square root imposes limit 1−x 2 ≥ 0 ∴ |x| ≤ 1. In addition, z = 1 − x 2 defines
a positive semicircle whose radius is r = 1. Because

−2 ≤ y ≤ 2

than (12.3) defines surface of a half-cylinder whose y-direction length is l = 4, see Fig. 12.7. The
length of semicircle equals s = r 2 π/2 = π/2, and therefore rectangular area equals

π
A=s×l = × 4 = 2π
2
∴
;; 
1 − x 2 dA = 2π
R

Fig. 12.7 Example 12.1-1(b)

z

x −2
−2 −11
1 2 y
 Complex Functions in Engineering
and Science 13

Important to Know

Decibel unit: it is used as a relative measure G between two quantities A1 and A2

A2
GdB = 10 log
def

A1

“dBm” unit: it is used as an absolute measure G of quantity A2 normalized to 10−3 (i.e. “milli” on
the engineering scale)

A2
GdBm = 10 log
def

10−3

Decibel scale: (note the number of zeros in the ration number x and the value of its respective
function log x)

Ratio Calculation [dB] Ratio Calculation [dB]

.. .. .. .. .. ..
. . . . . .
1/1000 10 log(1/1000) = 10 × (−3) −30 1/8 10 log(1/8) = 10 × (−0.9) −9
1/100 10 log(1/100) = 10 × (−2) −20 1/4 10 log(1/4) = 10 × (−0.6) −6
1/10 10 log(1/10) = 10 × (−1) −10 1/2 10 log(1/2) = 10 × (−0.3) −3
1 10 log(1/1) = 10 × (0) 0 1 10 log(1/1) = 10 × (0) 0
10 10 log(10) = 10 × (1) 10 2 10 log(2) = 10 × (0.3) 3
100 10 log(100) = 10 × (2) 20 4 10 log(4) = 10 × (0.6) 6
1000 10 log(1000) = 10 × (3) 30 8 10 log(8) = 10 × (0.9) 9
.. .. .. .. .. ..
. . . . . .

Basic transfer function forms: in engineering, instead of using letter i as the complex number, letter
j is used (the “i” is already used to indicate AC current). In this book, we do not distinguish between
the two.

H1 (j x) = a0 (a0 = const. ∈ ) (13.1)

© Springer Nature Switzerland AG 2021 295

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_13
296 13 Complex Functions in Engineering and Science

x
H2 (j x) = j (j 2 = −1, x0 = const. ∈ ) (13.2)
x0
x
H3 (j x) = 1 + j (13.3)
x0
1
H4 (j x) = x (13.4)
1+j
x0

13.1 Exercises

13.1 ** Basic Forms of Complex Functions z(x)

1. For each of z(x) functions, derive H (x) = 20 log |z(x)| and P (x) = z(x) functions in log(x)
scale. Then, calculate the following limits:

lim H (x) lim H (x) lim H (x) and

x→+∞ x→0 x→x0

lim P (x) lim P (x) lim P (x)

x→+∞ x→0 x→x0

where x0 is the number in each of the respective x/x0 fractions. Finally, calculate H (x = 0.1x0 ),
H (x = 10x0 ), P (x = 0.1x0 ), and P (x = 10x0 ).
x x x
(a) z(x) = j (b) z(x) = j (c) z(x) = 1 + j
2 10 2
x x x
(d) z(x) = 1 + j (e) z(x) = 1 − j (f) z(x) = 1 − j
10 2 10
1 1 1
(g) z(x) = x (h) z(x) = x (i) z(x) = x
1+j 1+j 1−j
2 10 10
1
(j) z(x) = x
1−j
2

13.2 *** Piecewise Linear Approximation of z(x)

1. For each z(x) function in Example 13.1-1, using log(x) scale, show the piecewise linear
approximation graphs of H (x) = 20 log |z(x)| and P (x) = z(x).

13.3 **** Complex Functions: Case Study

1. Show piecewise linear approximation graphs of H (x) and P (x) functions for the following
complex function:
Solutions 297

2 + jx
z(x) = 20,000 (13.5)
220 j x + 4000 − x 2

Solutions

Exercise 13.1, page 296

1. After using (13.1)–(13.4), we write

(a) % =
Given z(x) %j /2, then x0 = 2 and
x
% x% x
|z(x)| = %j % = ∴
2 2
x 
H (x) = 20 log |z(x)| = 20 log = 20 log(x) − 20 log(2)
2
lim H (x) = lim 20 log(x) − 20 log(2) = −∞
x→0 x→0

lim H (x) = lim 20 log(x) − 20 log(2) = 0

x→2 x→2

lim H (x) = lim 20 log(x) − 20 log(2) = +∞

x→+∞ x→+∞
and, phase function P (x) is found bydefinition as
(z(x)) x/2 π
P (x) = arctan = arctan = arctan(+∞) = = const.
(z(x)) 0 2
therefore, all limits of P (x) = π/2, i.e. not a function of x.
We calculate H (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times greater
than x0 = 2 as  
0.2 1
H (x = 0.2) = 20 log = 20 log = −20
2 10

20
H (x = 20) = 20 log = 20 log (10) = +20
2

x
(b) Given z(x) = j , then x0 = 10 and
% x % 10
% % x
|z(x)| = %j % = ∴
10 10
x
H (x) = 20 log |z(x)| = 20 log = 20 log(x) − 
20  : 20
log(10)
10
lim H (x) = lim 20 log(x) − 20 = +∞
x→+∞ x→+∞

lim H (x) = lim 20 log(x) − 20 = 0

x→10 x→10

lim H (x) = lim 20 log(x) − 20 = −∞

x→0 x→0
and, phase function P (x) is found bydefinition as
(z(x)) x/10 π
P (x) = arctan = arctan = arctan(+∞) = = const.
(z(x)) 0 2
therefore, all limits of P (x) = /2, i.e. not a function of x.
π
298 13 Complex Functions in Engineering and Science

We calculate H (x) at points x = 1, i.e. ten times smaller, and x = 100, i.e. ten times greater than
x0 = 10 as 
1
H (x = 1) = 20 log = −20
10

100

H (x = 100) = 20 log = +20

10

x
(c) Given z(x) = 1 + j , then x0 = 2 and
2
% x %%  x 2
%
|z(x)| = %1 + j % = 1 + 2 ∴
2 2
 x 2
H (x) = 20 log |z(x)| = 20 log 1 +
2
! "
 x 2
lim H (x) = 20 log lim 1 + = 20 log(1) = 0
x→0 x→0 2
! "
 x 2 √
lim H (x) = 20 log lim 1 + = 20 log 2 = 3
x→2 x→2 2
! " ! "
 x 2  x 2
lim H (x) = 20 log lim 1 + ≈ 20 log lim
xx0 xx0 2 xx0 2
x 
≈ 20 log lim ≈ 20 log |x|
xx0 2
and, phase function P (x) is found bydefinition as
(z(x)) x/2 x 
P (x) = arctan = arctan = arctan
(z(x)) 1 2
x 
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 2
x  
1 π
lim P (x) = lim arctan = arctan = = 45◦
x→2 x→2 2 1 4
x  π
lim P (x) = lim arctan = arctan(∞) = = 90◦
x→+∞ x→+∞ 2 2
We calculate H (x) and P (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times
greater than x0 = 2 as
Solutions 299

*
 2
0.2
H (x = 0.2) = 20 log 1 + ≈ 20 log(1) = 0
2
*
 2
20
H (x = 20) = 20 log 1 + ≈ 20 log (10) = +20
2

0.2
P (x = 0.2) = arctan = arctan(0.1) = 5.7◦ ≈ 0
2

20 π
P (x = 20) = arctan = arctan(10) = 84.3◦ ≈ −90◦ =
2 2

x
(d) Given z(x) = 1 + j , then x0 = 10 and
10
% x %%  x 2
%
|z(x)| = %1 + j % = 12 + ∴
10 10
 x 2
H (x) = 20 log |z(x)| = 20 log 1 +
10
! "
 x 2
lim H (x) = 20 log lim 1 + = 20 log(1) = 0
x→0 x→0 10
! "
 x 2 √
lim H (x) = 20 log lim 1 + = 20 log 2 = 3
x→10 x→10 10
! " ! "
 x 2  x 2
lim H (x) = 20 log lim 1 + ≈ 20 log lim
xx0 xx0 10 xx0 10
x
≈ 20 log lim ≈ 20 log |x|
xx0 10

and, phase function P (x) is found bydefinition as

(z(x)) x/10 x
P (x) = arctan = arctan = arctan
(z(x)) 1 10
x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 10
x 
1 π
lim P (x) = lim arctan = arctan = = 45◦
x→10 x→10 10 1 4
x  π
lim P (x) = lim arctan = arctan(∞) = = 90◦
x→+∞ x→+∞ 10 2
300 13 Complex Functions in Engineering and Science

We calculate H (x) and P (x) at points x = 1, i.e. ten times smaller, and x = 100, i.e. ten times
greater than x0 = 10 as*

0.1 2
H (x = 1) = 20 log 1 + ≈ 20 log(1) = 0
10
*

100 2
H (x = 100) = 20 log 1 + ≈ 20 log (10) = +20
10

1
P (x = 1) = arctan = arctan(0.1) = 5.7◦ ≈ 0
10

100 π
P (x = 100) = arctan = arctan(10) = 84.3◦ ≈ 90◦ =
10 2

x
(e) Given z(x) = 1 − j , then x0 = 2 and
2
% x %%  x 2
%
|z(x)| = %1 − j % = 1 + 2 ∴
2 2
 x 2
H (x) = 20 log |z(x)| = 20 log 1 +
2
! "
 x 2
lim H (x) = 20 log lim 1 + = 20 log(1) = 0
x→0 x→0 2
! "
 x 2 √
lim H (x) = 20 log lim 1 + = 20 log 2 = 3
x→2 x→2 2
! " ! "
 x 2  x 2
lim H (x) = 20 log lim 1 + ≈ 20 log lim
xx0 xx0 2 xx0 2
x 
≈ 20 log lim ≈ 20 log |x|
xx0 2
and, phase function P (x) is found bydefinition as 
(z(x)) −x/2 −x
P (x) = arctan = arctan = arctan
(z(x)) 1 2

−x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 2
 
−x −1 π
lim P (x) = lim arctan = arctan = − = −45◦
x→2 x→2 2 1 4

−x π
lim P (x) = lim arctan = arctan(−∞) = − = −90◦
x→+∞ x→+∞ 2 2
Solutions 301

We calculate H (x) and P (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times
greater than x0 = 2 as *

0.2 2
H (x = 0.2) = 20 log 1 + ≈ 20 log(1) = 0
2
*

20 2
H (x = 20) = 20 log 1 + ≈ 20 log (10) = +20
2

−0.2
P (x = 0.2) = arctan = arctan(−0.1) = −5.7◦ ≈ 0
2

−20 π
P (x = 20) = arctan = arctan(−10) = −84.3◦ ≈ −90◦ = −
2 2

x
(f) Given z(x) = 1 − j , then x0 = 10 and
10
% x %%  x 2
%
|z(x)| = %1 − j % = 1 + ∴
10 10
 x 2
H (x) = 20 log |z(x)| = 20 log 1 +
10
! "
 x 2
lim H (x) = 20 log lim 1 + = 20 log(1) = 0
x→0 x→0 10
! "
 x 2 √
lim H (x) = 20 log lim 1 + = 20 log 2 = 3
x→10 x→10 10
! " ! "
 x 2  x 2
lim H (x) = 20 log lim 1 + ≈ 20 log lim
xx0 xx0 10 xx0 10
x
≈ 20 log lim ≈ 20 log |x|
xx0 10

and, phase function P (x) is found bydefinition as 

(z(x)) −x/10 −x
P (x) = arctan = arctan = arctan
(z(x)) 1 10

−x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 10
 
−x −1 π
lim P (x) = lim arctan = arctan = − = −45◦
x→10 x→10 10 1 4

−x π
lim P (x) = lim arctan = arctan(−∞) = − = −90◦
x→+∞ x→+∞ 10 2
302 13 Complex Functions in Engineering and Science

We calculate H (x) and P (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times
greater than x0 = 2 as *

1 2
H (x = 1) = 20 log 1 + ≈ 20 log(1) = 0
10
*

100 2
H (x = 100) = 20 log 1 + ≈ 20 log (10) = +20
10

−1
P (x = 1) = arctan = arctan(−0.1) = −5.7◦ ≈ 0
10

−100 π
P (x = 100) = arctan = arctan(−10) = −84.3◦ ≈ −90◦ = −
10 2

1
(g) Given z(x) = x , then x0 = 2 and
% % 1+j2
% %
% 1 %
% % 1
|z(x)| = % %=  x 2 ∴
%1 + j x %
% % 1+
2 2
 x 2
 :0

H (x) = 20 log |z(x)| =  20
log(1) − 20 log 1 +
2
! "
 x 2
lim H (x) = −20 log lim 1 + = −20 log(1) = 0
x→0 x→0 2
! "
 x 2 √
lim H (x) = −20 log lim 1 + = −20 log 2 = −3
x→2 x→2 2
! "
 x 2
lim H (x) = −20 log lim 1+ ≈ −20 log |x|
x→+∞ x→+∞ 2
and, phase function P (x) is found by definition, after explicitly deriving the real and imaginary
parts of z(x), as
1 1 − j x/2 1 x/2
z(x) = = −j ∴
1 + j x/2 1 − j x/2 1 + (x/2)2 1 + (x/2)2
1 x/2
(z(x)) =  and (z(x)) = − 
1 + (x/2)2 1 + (x/2)2
Solutions 303

which again is reduced to  

(z(x)) −x/2 −x
P (x) = arctan = arctan = arctan
(z(x)) 1 2

−x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 2

−x π
lim P (x) = lim arctan = arctan(−1) = − = −45◦
x→2 x→2 2 4

−x π
lim P (x) = lim arctan = arctan(−∞) = − = −90◦
x→+∞ x→+∞ 2 2
We calculate H (x) and P (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times
greater than x0 = 2 as *

0.2 2
H (x = 0.2) = −20 log 1 + ≈ −20 log(1) = 0
2
*

20 2
H (x = 20) = −20 log 1 + ≈ −20 log (10) = −20
2

−0.2
P (x = 0.2) = arctan = arctan(−0.1) = −5.7◦ ≈ 0
2

−20 π
P (x = 20) = arctan = arctan(−10) = −84.3◦ ≈ −90◦ = −
2 2

1
(h) Given z(x) = x , then x0 = 10 and
% % 1 + j 10
% %
% %
% 1 % 1
|z(x)| = % %=  x 2 ∴
%1 + j x %
% % 1 +
2
10 10
 x 2
 :0

H (x) = 20 log |z(x)| = 
20
log(1) − 20 log 1 +
10
! "
 x 2
lim H (x) = −20 log lim 1 + = −20 log(1) = 0
x→0 x→0 10
! "
 x 2 √
lim H (x) = −20 log lim 1 + = −20 log 2 = −3
x→10 x→10 10
! "
 x 2
lim H (x) = −20 log lim 1 + ≈ −20 log |x|
xx0 xx0 10
304 13 Complex Functions in Engineering and Science

and, phase function P (x) is found by definition, after explicitly deriving the real and imaginary
parts of z(x), as

1 1 − j x/10 1 x/10
z(x) = = −j ∴
1 + j x/10 1 − j x/10 1 + (x/10)2 1 + (x/10)2

1 x/10
(z(x)) =  and (z(x)) = − 
1 + (x/10)2
1 + (x/10)2
which again is reduced to
 
(z(x)) −x/10 −x
P (x) = arctan = arctan = arctan therefore
(z(x)) 1 10

−x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 10

−x π
lim P (x) = lim arctan = arctan(−1) = − = −45◦
x→10 x→10 10 4

−x π
lim P (x) = lim arctan = arctan(−∞) = − = −90◦
x→+∞ x→+∞ 10 2
We calculate H (x) and P (x) at points x = 1, i.e. ten times smaller, and x = 100, i.e. ten times
greater than x0 = 10 as
*

1 2
H (x = 1) = −20 log 1 + ≈ −20 log(1) = 0
10

*
 2
100
H (x = 100) = −20 log 1 + ≈ −20 log (10) = −20
10


−1
P (x = 1) = arctan = arctan(−0.1) = −5.7◦ ≈ 0
10

−100 π
P (x = 100) = arctan = arctan(−10) = −84.3◦ ≈ −90◦ = −
10 2
Solutions 305

1
(i) Given z(x) = x , then x0 = 10 and
% % 1 − j 10
% %
% %
% 1 % 1
|z(x)| = % %=  x 2 ∴
%1 − j x %
% % 1 +
2
10 10

 x 2
 :0

H (x) = 
20
log(1) − 20 log 1 +
10
! "
 x 2
lim H (x) = −20 log lim 1 + = −20 log(1) = 0
x→0 x→0 10
! "
 x 2 √
lim H (x) = −20 log lim 1 + = −20 log 2 = −3
x→10 x→10 10
! "
 x 2
lim H (x) = −20 log lim 1 + ≈ −20 log |x|
xx0 xx0 10
and, phase function P (x) is found by definition, after explicitly deriving the real and imaginary
parts of z(x), as
1 1 + j x/10 1 x/10
z(x) = = +j ∴
1 − j x/10 1 + j x/10 1 + (x/10)2 1 + (x/10)2
1 x/10
(z(x)) =  and (z(x)) = 
1 + (x/10)2 1 + (x/10)2
which again is reduced
 to
x/10 x 
P (x) = arctan = arctan therefore
1 10
x
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 10
x π
lim P (x) = lim arctan = arctan(1) = = 45◦
x→10 x→10 10 4
x  π
lim P (x) = lim arctan = arctan(+∞) = = 90◦
x→+∞ x→+∞ 10 2
We calculate H (x) and P (x) at points x = 1, i.e. ten times smaller, and x = 100, i.e. ten times
greater than x0 = 10 as *

1 2
H (x = 1) = −20 log 1 + ≈ −20 log(1) = 0
10
*

100 2
H (x = 100) = −20 log 1 + ≈ −20 log (10) = −20
10

1
P (x = 1) = arctan = arctan(0.1) = 5.7◦ ≈ 0
10

100 π
P (x = 100) = arctan = arctan(10) = 84.3◦ ≈ 90◦ =
10 2
306 13 Complex Functions in Engineering and Science

1
(j) Given z(x) = , then x0 = 2 and
% % 1 − j x2
% %
% 1 %
% % 1
|z(x)| = % %=  x 2 ∴
%1 − j x %
% % 1 +
2
2 2
 x 2
 :0

H (x) =  20
log(1) − 20 log 1 +
2
! "
 x 2
lim H (x) = −20 log lim 1 + = −20 log(1) = 0
x→0 x→0 2
! "
 x 2 √
lim H (x) = −20 log lim 1 + = −20 log 2 = −3
x→2 x→2 2
! "
 x 2
lim H (x) = −20 log lim 1 + ≈ −20 log |x|
xx0 xx0 2

and, phase function P (x) is found by definition, after explicitly deriving the real and imaginary
parts of z(x), as
1 1 + j x/2 1 x/2
z(x) = = +j ∴
1 − j /2 1 + j /2
x x
1 + (x/2)2
1 + (x/2)2
1 x/2
(z(x)) =  and (z(x)) = 
1 + (x/2) 2
1 + (x/2)2
which again is reduced
 to
x/2 x 
P (x) = arctan = arctan therefore
1 2
x 
lim P (x) = lim arctan = arctan(0) = 0◦
x→0 x→0 2
x  π
lim P (x) = lim arctan = arctan(1) = = 45◦
x→2 x→2 2 4
x  π
lim P (x) = lim arctan = arctan(+∞) = = 90◦
x→+∞ x→+∞ 2 2
We calculate H (x) and P (x) at points x = 0.2, i.e. ten times smaller, and x = 20, i.e. ten times
greater than x0 = 2 as *

0.2 2
H (x = 0.2) = −20 log 1 + ≈ −20 log(1) = 0
2
*

20 2
H (x = 20) = −20 log 1 + ≈ −20 log (10) = −20
2

0.2
P (x = 0.2) = arctan = arctan(0.1) = 5.7◦ ≈ 0
2

20 π
P (x = 20) = arctan = arctan(10) = 84.3◦ ≈ 90◦ =
2 2
Solutions 307

Exercise 13.2, page 296

1. The following graphs are commonly known as the Bode plots:

(a) For z(x) = j x/2, Example 13.1-1(a), we (b) For z(x) = j x/10, Example 13.1-1(b), we
found that when x = 2, the value of H (x = found that when x = 10, the value of H (x =
2) = 0, and at the same time, angle is constant 10) = 0, and at the same time, angle is constant
P (x) = +π/2. In addition, every tenfold P (x) = +π/2. In addition, every tenfold
increase x results in +20 increase of H (x) (see increase x results in +20 increase of H (x) (see
Fig. 13.1). Fig. 13.2).
20 20
+20dB/dec +20/dec

20 log H(x)
20 log H(x)

0 0

−20 −20

phase (degree)
phase (degree)

90 90
0 0
−90 −90

0.2 2 20 0.1 10 100

x x

Fig. 13.1 Example 13.2-1(a) Fig. 13.2 Example 13.2-1(b)

(c) For z(x) = 1 + j x/2, Example 13.1-1(c), (d) For z(x) = 1 + j x/10, Example 13.1-
we found that when x = 2, the value of H (x = 1(d), we found that when x = 10, the value of
2) = 3, and at the same time, angle is P (x = H (x = 10) = 3, and at the same time, angle
2) = +π/4. In addition, angle limits are 0◦ and is P (x = 10) = +π/4. In addition, angle
+90◦ , while every tenfold increase x results in limits are 0◦ and +90◦ , while every tenfold
+20 increase of H (x). After accounting for all increase x results in +20 increase of H (x).
limits of lim H (x) (see Fig. 13.3). After accounting for all limits of lim H (x) (see
Fig. 13.4).
+20 +20dB/dec
20 log H(x)

+20 +20dB/dec
20 log H(x)

+3dB
0 +3dB
0
phase (degree)

+90
phase (degree)

+45 +90
0 +45
0
0.02 0.2 2 20 200
x
0.1 1 10 100 1000
x
Fig. 13.3 Example 13.2-1(c)
Fig. 13.4 Example 13.2-1(d)
308 13 Complex Functions in Engineering and Science

(e) For z(x) = 1 − j x/2, Example 13.1-1(e), (f) For z(x) = 1 − j x/10, Example 13.1-1(f),
we found that when x = 2, the value of H (x = we found that when x = 10, the value of
2) = 3, and at the same time, angle is P (x = H (x = 10) = 3, and at the same time, angle
2) = −π/4. In addition, angle limits are 0◦ and is P (x = 10) = −π/4. In addition, angle
−90◦ , while every tenfold increase x results in limits are 0◦ and −90◦ , while every tenfold
+20 increase of H (x). After accounting for all increase x results in +20 increase of H (x).
limits of lim H (x) (see Fig. 13.5). After accounting for all limits of lim H (x) (see
Fig. 13.6).
+20 +20dB/dec
20 log H(x)

+20 +20dB/dec

20 log H(x)
+3dB
0 +3dB
0
phase (degree)

phase (degree)
−45 0
−90 −45
−90
0.02 0.2 2 20 200
x
0.1 1 10 100 1000
x
Fig. 13.5 Example 13.2-1(e)
Fig. 13.6 Example 13.2-1(f)

1 1
(g) For z(x) = , Example 13.1-1(g), (h) For z(x) = x , Example 13.1-1(h),
1 + j x2 1 + j 10
we found that when x = 2, the value of we found that when x = 10, the value of
H (x = 2) = −3, and at the same time, angle H (x = 10) = −3, and at the same time, angle
is P (x = 2) = −π/4. In addition, angle is P (x = 10) = −π/4. In addition, angle
limits are 0◦ and −90◦ , while every tenfold limits are 0◦ and −90◦ , while every tenfold
increase x results in −20 increase of H (x). increase x results in −20 increase of H (x).
After accounting for all limits of lim H (x) (see After accounting for all limits of lim H (x) (see
Fig. 13.7). Fig. 13.8).
20 log H(x)

0
20 log H(x)

0
−3dB −20/dec −3dB −20/dec
−20 −20
phase (degree)
phase (degree)

0 0
−45 −45
−90 −90

0.02 0.2 2 20 200 0.1 1 10 100 1000

x x

Fig. 13.7 Example 13.2-1(g) Fig. 13.8 Example 13.2-1(h)

Solutions 309

1 1
(i) For z(x) =x , Example 13.1-1(i), (j) For z(x) = , Example 13.1-1(j),
1 − j 10 1 − j x2
we found that when x = 10, the value of we found that when x = 2, the value of
H (x = 10) = −3, and at the same time, H (x = 2) = −3, and at the same time,
angle is P (x = 10) = π/4. In addition, angle angle is P (x = 2) = π/4. In addition, angle
limits are 0◦ and +90◦ , while every tenfold limits are 0◦ and +90◦ , while every tenfold
increase x results in −20 increase of H (x). increase x results in −20 increase of H (x).
After accounting for all limits of lim H (x) (see After accounting for all limits of lim H (x) (see
Fig. 13.9). Fig. 13.10).

20 log H(x)
0
20 log H(x)

0
−3dB −20/dec −3dB −20dB/dec
−20 −20
phase (degree)

phase (degree)
+90 +90
+45 +45
0 0

0.1 1 10 100 1000 0.02 0.2 2 20 200

x x

Fig. 13.9 Example 13.2-1(i) Fig. 13.10 Example 13.2-1(j)

It is important to note that, as long as the form of analytical function z(x) is same where the
only change is the value of x0 , then all respective graphs have the same shapes. That is to say,
the graphs are only shifted to x = x0 . This property greatly simplifies the analysis of more
complicated z(x) functions. General idea is to factorize complicated z(x) functions into these
basic simpler “building blocks” and then “assemble” the final response by adding these basic
functions.

Exercise 13.3, page 296

1. This is the second order function; thus, in order to find out if roots of its denominator are reel or
complex, it is necessary to factorize its denominator.

2 + jx 2 + jx
H (j ω) = 2 000 = 2 000 2 2
220j x + 4 000 − x 2 j x + 220j x + 4 000
2 + jx 2 + jx
= 2 000 = 2 000
j 2x2 + 20j x + 200j x + 4 000 j x (20 + j x) + 200 (20 + j x)
2 + jx
= 2 000
(20 + j x)(200 + j x)
 x
2
1 + j
  2
=
2 000 x   x 
 1+j

20 200 1 + j

20 200
310 13 Complex Functions in Engineering and Science

x
1+j
= 2
x x 
1+j 1+j
20 200

This factorized form is suitable for conversion into the sum of the basic form simply by rewriting
it in its logarithmic form,

x ⎡ ⎤
1+j
⎢ 2 ⎥
20 log z(x) = 20 log ⎣  x x ⎦
1+j 1+j
20 200
 x  x  x 
= +20 log 1 + j −20 log 1 + j −20 log 1 + j
  2   20    200 
1 x0 =2 2 x0 =20 3 x0 =200

Obviously, this example is a second order function whose poles are real, and thus it is possible to
decompose it into the basic first order functions. Each summation term in the logarithmic form is
in effect one of the already studied basic forms, as annotated.
Similarly, the phase plot is created as the sum of the corresponding linear terms,

x x x
z(x) = + arctan − arctan − arctan
2
       20  200
1 x0 =2 2 x0 =20 3 x0 =200

Once the gain and phase logarithmic forms are factorized, first we plot each of the summing terms
1,2 and 3 separately, see Fig. 13.11(left). Then, the summing operation is done by simply adding
the linear sections of all three terms to produce the Body plots for gain and phase, see Fig. 13.11
(right).

Fig. 13.11 Example 13.3-1 40

20
20 log H(x)

0dB/dec
20 log H(x)

1
20
+17 +20dB/dec −20dB/dec
0 3 +3
0
2
−20
−20
90 90 0◦ /dec
phase (degree)

phase (degree)

1
45 45 +45◦ /dec
−90◦ /dec
0 0
3
−45 −45
2
−90 −90
0.02 0.2 2 20 200 2 000 20 000 0.02 0.2 2 20 200 2 000 20 000
x x
 Differential Equations
14

Important to Know

First order differential equations take the form of

dy
f (x, y, y  ) = 0 where y  ≡
dx

and the common types are as follows:

1. equations with separable variable take the form of

f (x)
y =
g(x)

2. homogeneous equations take the form of

y 
y = f
x

3. linear equations take the form of

y  + f (x) y = g(x)

Orthogonality between two functions, for example, at a given point (x, y) relation f (x) ⊥ g(x) is
satisfied for
 
 1
F (x, y, yf ) and G x, y, − 
yf

where yf is derivative of f (x) and (−1/yf ) is derivative of g(x), that is to say, tangent, at a given
point (x, y).

© Springer Nature Switzerland AG 2021 311

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_14
312 14 Differential Equations

Homogeneous linear differential equations with the constant coefficients take the form of

a1 y  + a0 y = 0 (1st order)
a2 y  + a1 y  + a0 y = 0 (2nd order)
a3 y  + a2 y  + a1 y  + a0 y = 0 (2nd order)
···

Characteristic polynomial Pn (r) is written as

a1 r + a0 = 0 (1st order)
a2 r 2 + a1 r + a0 = 0 (2nd order)
a3 r 3 + a2 r 2 + a1 r + a0 = 0 (2nd order)
···

Solutions of homogeneous linear differential equations with the constant coefficients are written after
calculating the roots of Pn (r) = 0, where n indicates nth order polynomial, and thus there are n roots
r1 , r2 , · · · , rn .

1 : (r1 = r2 · · · = rn ) ∴ y(x) = C1 exp(r1 x) + C2 exp(r2 x) + · · · + Cn exp(rn x)

2 : (r = r1 = r2 = ri>2 ) ∴ y(x) = C1 exp(r x) + C2 x exp(r x) + · · · + Cn exp(rn x)
= exp(r x) [C1 + C2 x] + · · · + Cn exp(rn x)

where, in order to derive more compact y(x) function, the complex conjugate roots may be further
transformed by algebra techniques.

14.1 Exercises

14.1 * Separable Variables

1. Solve equations:
(a) y  + y = a, (a ∈ R) (b) y − 2xy  = 1

2. Solve equations:
√ √
(a) x dy = y dx (b) x(1 + y 2 ) = yy 
14.1 Exercises 313

14.2 ** First Order Homogenous Equations

1. Solve equations:
y2 + x2 y2 − x2
(a) y = (b) y =
xy 2xy
2. Solve equations:
(a) (x + y) dx − x dy = 0 (b) xy 2 dy = (x 3 + y 3 ) dx

14.3 * Linear Equations

1. Solve equations:
(a) y  − 4xy = x (b) y  − 2y − 3 = 0
2y
y  + 2xy = 2xe−x y −
2
(c) (d) − (x + 1)3 = 0
x+1

14.4 * Linear Equations with Constant Coefficients

1. Solve equations:
(a) y  − 9y = 0 (b) y  − 2y  + y = 0
(c) y  + 4y  + 4y = 0 (d) y  + 6y  + 25y = 0

14.5 *** Engineering Examples

1. In a series RL circuit, see Fig. 14.1, at moment t = 0, current i(0) = 0. Derive expression for
i(t) (t ≥ 0), i.e. when the switch is closed, and sketch its time domain graph.

Fig. 14.1 Example 14.5-1 vR

E iL
R
L vL

2. At moment t = 0, a free falling ball whose mass is m and the initial velocity (a vector quantity)
#»
v (0) = 0 is dropped, Fig. 14.2. Assuming the air resistance to be proportional to velocity as k #»
v,
derive expression for the ball’s speed (a scalar quantity) v(t) as a function of time.
314 14 Differential Equations

Fig. 14.2 Example 14.5-2

kv

v
mg

3. Given an RLC circuit, Fig. 14.3, where the switch is closed at moment t = 0. Assuming the initial
capacitor charge is q(0) = q0 , derive the expressions for i(t) and v(t).

Fig. 14.3 Example 14.5-3 vR

E R
L vL
i(t) C

vC

Solutions

Exercise 14.1, page 312

1. Separation of variables technique is used to solve differential equations whose form can be written
as

dy f (x)
y = = ∴ g(y) dy = f (x) dx
dx g(y)

Therefore, the solution y(x) is found by integration of both sides of the equation.
(a) Separation of variables technique
y  − y = a, (a ∈ R)
; ;
dy dy dy dy
=a−y ∴ = dx ∴ = − dx ∴ =− dx
dx a−y y−a y−a
 
= y − a = t ∴ dy = dt
;
dt
∴ = −x + C1 ∴ ln |t| = −x + C1 ∴ ln |y − a| = −x + C1
t
 
∴ |y − a| = e−x+C1 = e−x eC1 = C = ±eC1
y(x) = a + Ce−x
Solutions 315

(b) Separation of variables technique

y − 2xy  = 1
dy dy 1 dx
y = 2xy  + 1 ∴ y − 1 = 2xy  = 2x ∴ =
dx y−1 2 x
∴
; ;
dy 1 dx   1
= = y − 1 = t ∴ dy = dt ∴ ln |y − 1| = ln |x| + ln C1
y−1 2 x 2
1  
ln |y − 1| = ln(C1 |x|) = ln(C1 |x|) /2 = ln x, (x > 0) ∴ ln |x| = ln x
1

2
∴
   
y−1= C1 |x| = C1 |x| = C |x|

y(x) = 1 + C |x|

2. Separation of variables technique

; ;
√ √ dy dx √ √
(a) x dy = y dx ∴ √ = √ ∴ y− x=C
y x
(b) Constant C is arbitrary, so it may as well be ln(C) or eC if it helps to write a more compact
solution.
; ;
y dy y
x(1 + y 2 ) = yy  ∴ x = ∴ x dx = dy
1 + y dx
2 1 + y2
 
1 + y 2 = t ∴ 2y dy = dt
; ;
1 dt
∴ x dx =
2 t
x2 1 
= ln(1 + y 2 ) + ln C = ln[C(1 + y 2 )]
2
2

or,

y = C1 e x 2 − 1

where C1 = 1/C.

Exercise 14.2, page 313

1. Differential equations that can be written in form

y 
y = f
x

are known as “homogeneous” and are solved by the following change of variables technique:

y  
= t ∴ y = tx = [f (x)g(x)] = f  g + f g 
t
316 14 Differential Equations

∴
dy dt
dy = t dx + x dt ⇒ =t +x
dx dx

followed by the method of variable separation.

(a) Reorganize the equation into the homogeneous form, then change the variables and integrate.
⎫
x2 + y2 dy y x 1
y = ∴ = + =t+ ⎪ ⎪
⎬
xy dx x y t 1 dt
tC + = tC + x
dy dt ⎪
⎪ t dx
=t +x ⎭
dx dx
; ;
1 dt dx t2  
=x ∴ = t dt ∴ ln |x| + ln C1 = ; C = ln C1
t dx x 2
 y 
t = 2 ln(C|x|) ∴ = 2 ln(C|x|)
x

∴ y = x 2 ln(C|x|)

(b) Homogeneous differential equation with parametric form solution.

 ⎫
y2 − x2 dy y x 1 1 ⎪
y = ∴ = − = t− ⎪
⎬
2xy dx 2x 2y 2 t
dy dt ⎪
⎪
=t +x ⎭
dx dx

dt 1 1 t2 − 1
t +x = t− =
dx 2 t 2t
dt t2 − 1 t 2 − 1 − 2t 2 t2 + 1
x = −t = =−
dx 2t 2t 2t
; ;
dx 2t
∴ =− dt
x t2 + 1
;
2t dt
∴ ln |x| = −
t2 + 1
2 
= t + 1 = r ∴ 2t dt = dr, (t 2 + 1 = 0)
;
dr
=− = − ln |r| = − ln[t 2 + 1] + C1 (C = ln C1 )
r
t2 + 1 C
ln |x| = ln ∴ x= 2 , and
C t +1
Ct
y=t ∴ y=
t2 +1
Solutions 317

At this stage, the solution is parametric, i.e. y = y(t) and x = x(t). The parametric variable t
may be eliminated as follows:
⎫
C2 C2 C ⎪
x2 = ∴ (t 2 + 1)2 = ∴ t2 = − 1⎪
⎪
⎬
(t + 1)2
2 x2 x
∴
C2 t 2 ⎪
⎪
y 2
= 2 ⎪
⎭
(t + 1)2

C2 (C/x − 1)
 (C − x)/x

y2 = = = x(C − x) = Cx − x 2
C/x
 2 2
1/x 2
∴

y 2 + x 2 − C x = 0 or, y = Cx − x 2

2. Homogenous differential equations are solved by using y/x = t change of variables.

(a) After the initial transformations, we can change the variables.

(x + y) dx − x dy = 0
%
%
x dx + y dx = x dy %% /x dx

y dy
1+ =
x dx
 
y dy
=t ∴ y =xt ∴ = y = t + x t 
x dx
∴
; ;
dx
1 + tC = tC + x t  ∴ dt = ∴ t = ln C x
x
∴
y = x ln C x
318 14 Differential Equations

(b) After the initial transformations, we can change the variables.

 2
dy x3 + y3 x y
xy dy = (x + y ) dx ∴
2 3 3
= = +
dx xy 2 y x
 
y =tx
∴
1
tC + x t  = 2 + tC
t
∴
; ; 
dx t3
t 2 dt = ∴ = ln C |x| ∴ y = x 3 3 ln C |x|
x 3

Exercise 14.3, page 313

1. Equations in the form of “linear non-homogeneous differential equation”

y  + f (x) y = g(x)

may be solved with the following steps:

1. Calculate the “integration factor” μ as

<
μ=e f (x) dx

2. Assume the relationship

d
(y μ) = μg(x)
dx

3. Integrate and solve per y.

Solutions 319

(a) y  − 4xy = x ∴ f (x) = −4x and g(x) = x

; ;
x2
= −2x 2 ∴ μ = e−2x
2
f (x) dx = −4 x dx = −4
2

;   ;
d  −2x 2  −2x 2 −2x 2
y = x e−2x dx
2
e y =xe ∴ d e
dx
;
e−2x y = x e−2x dx
2 2

 
1
−2x = t ∴ −4x dx = dt ∴ x dx = − dt
2
4
;
1 1 1
et dt = − et = − e−2x + C
2
=−
4 4 4
∴
1 2
y = − + Ce2x
4

(b) We declare f (x) = −2 and g(x) = 3 so that

; ;
y  − 2y − 3 = 0 ∴ y  − 2y = 3 ∴ f (x) = −2 dx = −2x

∴
d  −2x 
μ = e−2x ∴ ye = 3 e−2x
dx
∴
; ; %
  3 −2x %
d ye −2x
=3 e −2x
dx ∴ y e −2x
=− e + C %% /e−2x
2
∴
3
y = − + C e2x
2
320 14 Differential Equations

(c) Given y  + 2xy = 2xe−x , declare

2

f (x) = 2x and g(x) = 2x e−x

2

so that
; ;
2
f (x) dx = 2 x dx = x 2 ∴ μ = ex

It follows that
;
ye x2
=2
2
ex x  2 dx = x 2 + C ∴ y = e−x 2 x 2 + C 
e−x

(d) Declare

2x
f (x) = and g(x) = (x + 1)3
(x + 1)

so that

2y 2
y − − (x + 1)3 = 0 ∴ y  − y = (x + 1)3
x+1 x+1
∴
; ;
2 1
f (x) dx = − dx = −2 ln |x + 1| = ln
x+1 (x + 1)2
∴

1 1
μ=HH @
exp ln =
(x + 1)2 (x + 1)2
∴
;  ;
A d y 1 1  
= (x + 1)3 dx = x + 1 = t, dx = dt
AA A (x + 1)2 (x
 
+ 1)2

(x + 1)2 (x + 1)4
= + C1 ∴ y = + C1 (x + 1)2
2 2
Solutions 321

Or, if we calculate the last integral without the change of variables technique, the only
difference would be in the integration constant C, i.e.
; ; ;
x2
(x + 1) dx = x dx + dx = +x+C
2
∴

x2
y = (x + 1) 2
+x+C
2

however, if C = C1 + 1/2, which is also a valid integration constant, then the two forms of result
are equivalent.

Exercise 14.4, page 313

1. Equations in the form of “linear differential equation with constant coefficients” as

an y (n) + an−1 y (n−1) + · · · + a2 y  + a1 y  + a0 y = 0

are, in general, solved with the following steps:

1. Write the equivalent “characteristic polynomial”

Pn (r) = an r n + an−1 r n−1 + · · · + a2 r 2 + a1 r + a0 = 0

where the derivative orders in differential equation are replaced with the same order powers.
2. Calculate the roots of nth order polynomial Pn (r) = 0, i.e. (r1 , r2 , . . . , rn ).
3. Depending upon whether all the roots ri are unique, or multiple but real, or if there are complex
roots, the solution is found in the following form:
  
y(x) = f x n , Ci eri x (n = 0, 1, 2 . . . )
i

where the term x n depends on multiplicity of the P (r) roots, i.e. if a root is unique, then n = 0, if
the root is a “double root”, then there are x 0 , x 1 terms, if it is a triple root, then there are x 0 , x 1 , x 2
terms, etc.
(a) All roots of P (r) are real and unique

y  − 9y = 0 ∴ P2 (r) = r 2 − 9 = (r − 3)(r + 3) = 0
∴ r1 = 3, r2 = −3 (r1 = r2 )
322 14 Differential Equations

Therefore, we write the solution in the form

y(x) = C1 er1 x + C2 er2 x = C1 e3x + C2 e−3x

Verification:

y(x) = C1 e3x + C2 e−3x

y  (x) = 3C1 e3x − 3C2 e−3x
y  (x) = 9C1 e3x + 9C2 e−3x

Therefore,

y  − 9y = 0
 −3x
  
9C1 e3x + 9C2 e − 9 C1 e3x + C2 e−3x = 0
     
9C
 1e
3x
+9C2e
−3x
−9C1e
3x
9C
− 2e
−3x
=0
0=0 

(b) Not all real roots of P (r) are unique, there is one double root

y  − 2y  + y = 0 ∴ ∴ P2 (r) = r 2 − 2r + 1 = (r − 1)(r − 1) = 0
∴ r1 = 1, r2 = 1 (r1 = r2 = r)

Therefore, we write the solution in the form

y(x) = C1 erx + x C2 erx = C1 ex + x C2 ex = ex (C1 + x C2 )

Otherwise, the solution terms are not independent.

Verification:

y(x) = C1 ex + xC2 ex = ex (C1 + x C2 )

 
[f g] = f  g + f g 
y  (x) = C1 ex + C2 ex + xC2 ex = ex (C1 + C2 (x + 1))
y  (x) = C1 ex + C2 ex + C2 ex + xC2 ex = ex (C1 + C2 (x + 2))
Solutions 323

Therefore,

y  − 2y  + y = 0 = 0
# x $ # $ # $
e (C1 + C2 (x + 2)) − 2 ex (C1 + C2 (x + 1)) + ex (C1 + x C2 ) = 0 ÷ex
xC2 + H
C1 +  2C1 − 
2CH2 −  2 − H
 2CH2 + 
C1 +  2 = 0
 2xC xC
0=0 

(c) Not all real roots of P (r) are unique, there is one double root

y  + 4y  + 4y = 0 ∴ P2 (r) = r 2 + 4r + 4 = (r + 2)(r + 2) = 0
∴ r1 = −2, r2 = −2 (r1 = r2 = r)

Therefore, we write the solution in the form

y(x) = C1 erx + x C2 erx = C1 e−2x + x C2 e−2x = e−2x (C1 + x C2 )

which can be confirmed by the verification.

(d) Roots of P (r) are complex and therefore conjugated

y  + 6y  + 25y = 0 ∴ P2 (r) = r 2 + 6r + 25
√
−6 ± 36 − 4 × 25
r1,2 = = −3 ± 4i
2
∴
r1 = −3 + 4i
r2 = −3 − 4i (r1 = r2∗ )

In the case of unique roots, we write the solution of form

y(x) = C1 er1 x + C2 er2 x = C1 e(−3+4i)x + C2 e(−3−4i)x

# $
= e−3x C1 e4xi + C2 e−4xi
 ix 
e = cos x + i sin x
= e−3x [C1 (cos 4x + i sin 4x) + C2 (cos(−4x) + i sin(−4x))]
 
cos(−x) = cos x, sin(−x) = − sin x
= e−3x [C1 (cos 4x + i sin 4x) + C2 (cos 4x − i sin 4x)]
= e−3x [(C1 + C2 ) cos 4x + i(C1 − C2 ) sin 4x]
 
C1 + C2 = D1 , i(C1 − C2 ) = D2
324 14 Differential Equations

= e−3x (D1 cos 4x + D2 sin 4x) [D1,2 ∈ C]

As an illustration, assuming a simple case of D1 = D2 = 1, the plot of y(x) module is shown

in Fig. 14.4. For x < 0, the amplitude of y(x) shows oscillatory behaviour, while for x ≥ 0 it is
damped by the e−3x term.

Fig. 14.4 Example 14.4-1

y(x)
1

0 x

−1 y(x)
e−3x
−1 0 1 2

Exercise 14.5, page 313

1. This classic example is analysed by the first order non-homogeneous differential equation with
constant coefficients. In general, first we solve the homogenous part, then the non-homogeneous
part, and then based on the initial conditions, we solve the complete equation as the sum of two
partial solutions.

The circuit equation setup: KVL equation of the circuit loop is derived as

di
E = vR + vL = i R + L
dt
∴
di di R E
L +Ri = E ∴ + i=
dt dt L L

Because E, R, L =const., this equation is in the form of first order non-homogeneous differential
equation with constant coefficients. The initial condition is i(0) = 0.

Homogenous part solution, ih (t):

R R R
i + i = 0 ∴ P (r) = r + = 0 ∴ r = − ∴ ih (t) = C e−(R/L) t
L L L

Non-homogenous part solution, ip (t):

⎫
R E ⎪

i + i = ⎪
⎬
L L ip = 0 R E E
∴ 0 + C2 = ∴ C2 =
E ⎪ p
⎪ i = C L L R
= const.⎭
2
L
Solutions 325

Complete solution, i(t) = ih (t) + ip (t):

⎫
E⎪
=Ce −(R/L) t
+ ⎬
 :
i(t) 1 E E
R e−(R/L)
i(0) = C  0
+ =0 ∴ C=−
⎪
⎭ R R
i(0) = 0

Therefore, the complete solution is in form

E E E 
i(t) = − e−(R/L) t = 1 − e−(R/L) t
R R R
∴
E
lim i(t) =
t→∞ R

It is important to note that at moment τ = L/R, the amplitude of current reached the level

E  E  E
i(τ ) = 1 − e−(R/L) (L/R) = 1 − e−1 ≈ 0.632
R R R
∴
E
i(τ ) = 0.632 = 0.632 i(∞)
R

It is also important to note that at moment 5τ = 5(L/R), the amplitude of current reached the
level

E  E  E
i(5τ ) = 1 − e−(R/L) 5 (L/R) = 1 − e−5 ≈ 0.632
R R R
∴
E
i(5τ ) = 0.993 = 0.993 i(∞)
R

Variable τ is commonly known as “time constant” and is a very important “unit of measure” in
engineering and science, Fig. 14.5.

Fig. 14.5 Example 14.5-1 i(t)

E
R
0.632

t
0
0 W 2W 3W 4W 5W
326 14 Differential Equations

2. This classic example is also analysed by the same type of the first order non-homogeneous
differential equation with constant coefficients. In general, first we solve the homogenous part,
then the non-homogeneous part, and then based on the initial conditions, we solve the complete
equation as the sum of two partial solutions.

The force equation setup: in accordance with Newton’s laws of motion, we write

dv
F = ma = m = mg − kv
dt
∴
k
mv  + kv = mg ∴ v  + v=g
m

Because k, m, g =const., this equation is in the form of first order non-homogeneous differential
equation with constant coefficients. The initial condition is v(0) = 0.

Homogenous part solution, vh (t):

k k k
v + v = 0 ∴ P (r) = r + =0 ∴ r=− ∴ vh (t) = C e−(k/m) t
m m m

Non-homogenous part solution, vp (t):

⎫
k ⎪
v + v = g⎬ v = 0

k mg
m p
∴ 0+ C2 = g ∴ C2 =
⎪ vp = C2
g = const. ⎭
m k

Complete solution, v(t) = vh (t) + vp (t):

mg 
= C e−(k/m) t +
 :
v(t) 1 mg mg
e−(k/m)
k
v(0) = C  0
+ =0 ∴ C=−
v(0) =0 k k

Therefore, the complete solution is in form

mg mg −(k/m) t mg  
i(t) = − e = 1 − e−(k/m) t
k k k
∴
mg
lim i(t) =
t→∞ k

In other words, there is “terminal velocity” that a free falling body can achieve in the resistive
media (e.g. air). In addition, exactly the same type of equations describes very different physical
processes, for example, the velocity of falling body and current in RL circuit.
3. This example is based on the second order homogeneous differential equation with constant
coefficients.
The circuit’s KVL equation setup:
Solutions 327

;
di 1 d
vL + vR + vC = E ∴ L + R i + i dt = E
dt C dt
∴
1 R 1
Li  + R i  + i = 0 ∴ i  + i  + i=0
C L LC

Characteristic equation:

R 1
P (r) = r 2 + + =0
L LC
∴
4L
R2 −
R C
r1,2 =− ±
2L 2L
* *
 2  2
R R 4
L R R 1
=− ± − =− ± −
2L 2L 4
L2 C 2L 2L LC

Case 1: R 2 > 4L/C: both roots are real, unique, and negative. They are negative because
⎧ 
 ⎨−R + R 2 − 4L/C <0
R 2 − 4L/C < R 2 ∴ R 2 − 4L/C < R ∴
⎩−R − R 2 − 4L/C <0

and thus, the amplitude in this case is exponentially falling in time, see Fig. 14.6.

i(t) = C1 er1 t + C2 er2 t

Case 2: R 2 = 4L/C: both roots are real, negative, and equal; that is to say,

−R
r1 = r2 = r = ∴ i(t) = C1 er t + t C2 er t = er t (C1 + t C2 )
2L

In this case, for (t > 0), amplitude quickly reaches maximum, and then it exponentially tends to
zero, Fig. 14.6.
Case 3: R 2 < 4L/C: both roots are complex (conjugate) and unique; that is to say,

−R 1 R2 −R 1 R2
r1 = +i − = α + iω r2 = −i − = α − iω
2L LC 4L2 2L LC 4L2
∴
i(t) = C1 er1 t + C2 er2 t = e−α t (D1 cos ω t + D2 sin ω t)
328 14 Differential Equations

In this case, the response is damped oscillatory, Fig. 14.6. However, in the extreme case of no damping,
i.e. α = 0, the oscillatory behaviour continues indefinitely.

Fig. 14.6 Example 14.5-3

y(x) 1
2
1 3

0 x

−1

0 1 2
 Part II
Mathematics for Signal Processing
 Series
15

Important to Know

Arithmetic progression: It is a sequence of numbers where the difference d between any two
subsequent terms is constant. That being case, given the first term a1 , a general term an is found
as

an = a1 + (n − 1) d

The sum of first nterms in arithmetic progression:


n
n
Sn = ak = (a1 + an )
k=1
2

Geometric progression: It is a sequence of numbers where the ratio r between any two subsequent
terms is constant. That being case, given the first term a1 , a general term an is found as

a1 , a2 = r a1 , a3 = r a2 , · · ·
∴
an = a1 r n−1

The sum of geometric progression: if a1 = a; (a = 0) then the sum of first n terms is found as


n 
n
Sn = ak = a r k−1 = a + ar + ar 2 + · · · + ar n−1
  
k=1 k=1 n terms

We consider the following cases:

1. r = 1: Assuming a = 0, the sum is calculated simply as

© Springer Nature Switzerland AG 2021 331

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_15
332 15 Series


n→∞ 
n→∞
*1
 
n→∞
a r n−1 = 
1n−1
a = a = a + a + a + ··· → ∞
  
n=1 n=1 n=1
infinitely terms

2. r = 1: The sum may be found as follows:


+
Sn = a + 
ar ar
ar2 + · · · +  n−1
∴

+
r Sn = 
ar ar
ar2 + · · · +  n−1
+ ar n
∴
a (1 − r n )
(Sn − r Sn ) = a − ar n ⇒ Sn =
1−r

Convergence of the geometric sum: Depending upon value of |r| = 0, the infinite sum may be
calculated in the following two cases:
(a) |r| ≥ 1: In this case, each term of the geometric series is greater and greater; therefore,
∞
 a (1 − r n ) ∞ −anr n−1
a r n−1 = lim = lim →∞
L.H.
(=)
n=1
n→∞ 1−r ∞ n→∞ −1

that is to say, for |r| ≥ 1 the sum of geometric series is divergent.

(b) |r| < 1: In this case, we recall that powers of numbers inferior to one tend to zero very fast,
thus

∞ 0
 r
a( 1−>
n
) a
ar n−1
= lim =
n=1
n→∞ 1−r 1−r

that is to say, for |r| < 1 the sum of geometric series is convergent.

15.1 Exercises

15.1 * Basic Sums

1. Write the first few terms of the following series:

.√ /∞ . nπ /∞
(a) S = n−3 (b) S = cos
n=3 6 n≥0
2. Calculate the following sums:

4 
4 
3
(a) i (b) 2 (c) i2
i=1 i=1 i=1


5

4
1 3
n−1
(d) 2j (e) (f)
j =0 k=1
k n=1
n2 + 3
15.1 Exercises 333

3. Calculate the following sums:


n 
n 
n 
n
(a) 1 (b) i (c) i2 (d) i(4i − 3)
i=1 i=1 i=1 i=1

4. Write the following series as sum of its general term:

1 2 3
(a) S = 23 + 33 + · · · + n3 (b) S= + + + ···
2 3 4

15.2 *** Limits of Sums

1. Calculate the following limits:

∞
! "
 3 i 2
(a) lim +1
n→∞
i=1
n n

15.3 *** Geometric Series

1. Write the following series as sum of its general term:

S = a + ar + ar 2 + ar 3 + · · · + ar n + · · ·

then calculate the sum if

(a) r=1 (b) r=1 (c) |r| < 1

2. Calculate the following sums:

∞

10 20 40
(a) S =5− + − + ··· (b) S= 22n 31−n
3 9 27 n=1
∞

(c) S= x n ; |x| < 1
n=1
334 15 Series

Solutions

Exercise 15.1, page 332

1. Given values for n and the general term, for infinite series we evaluate the general term for several
subsequent n.
.√ /∞ ∞
 √ √ √ √
(a) S= n−3 = n − 3 = 0 + 1 + 2 + 3 + ··· + n − 3 + ···
n=3
n=3
. √
nπ /∞
∞
 nπ 3 1 nπ
(b) S = cos = cos =1+ + + 0 + · · · + cos + ···
6 n=0
n≥0
6 2 2 6

2. For limited series (i.e. with less than infinite number of terms), it is possible to add all terms.

4
(a) i = 1 + 2 + 3 + 4 = 10
i=1


4
(b) 2=2+2+2+2=8
i=1


3
(c) i 2 = 12 + 22 + 32 = 1 + 4 + 9 = 14
i=1


5
(d) 2j = 20 + 21 + 22 + 23 + 24 + 25 = 1 + 2 + 3 + 8 + 16 + 32 = 63
j =0


4
1 1 1 1 1 25
(e) = + + + =
k=1
k 1 2 3 4 12

3
n−1 0 1 2 13
(f) = + + =
n=1
n +3
2 3 7 12 42

3. Summing the first n terms results in a general algebraic expression.


n

(a) S= 1 = 1 + 1 + 1 + ··· + 1 = n
  
i=1 n×1


n
n(n + 1)
(b) S= i = 1 + 2 + 3 + ··· + n =
i=1
2
The legend has it that this classic problem of the sum of an arithmetic sequence was solved by
Gauss when he was still in the elementary school. His proof is based on understanding that the
addition is a commutative operation (meaning that order does not matter), thus the same sum
can be written in both ascending and descending orders, then term by term addition of the two
equations gives
Solutions 335

S = 1 +2 +3 + · · · +(n − 1) +n
S = n +(n − 1) + · · · +3 +2 +1
∴
2S = (n + 1) + (n + 1) + (n + 1) + · · · + (n + 1) + (n + 1)
  
n×(n+1)

Therefore, we write

n(n + 1)
2S = n(n + 1) ∴ S =
2

This result is often used and assumed to be well known.

(c) The sum of first n squares is also often used result.


n
n(n + 1)(2n + 1)
S= i 2 = 12 + 22 + 32 + · · · + (n − 1)2 + n2 = (15.1)
i=1
6

Derivation of sum (15.1) is based on the mathematical induction technique. The main idea is split
into two steps: First we prove that equality is valid for a certain number, often for n = 0 or n = 1
(but, any number n = N could be used as the starting point). In the second step, we make the
claim that if the equality is valid for a given n = k case, then it must be also valid for the next
term, i.e. n = k + 1.

Specifically,

1. Case: n = 1 we must show that the following equality is correct:


1
n(n + 1)(2n + 1)
S1 = i 2 = 12 = 1 =
i=1
6

Indeed, the last formula gives

1 × (1 + 1)(2 × 1 + 1) 1×2×3 6
S1 = = = =1
6 6 6

which simply proves that (15.1) is correct for n = 1.

336 15 Series

2. Induction: In general case when n = k, we say that if

k(k + 1)(2k + 1)
Sk = then,
6
Sk+1 = 12 + 22 + · · · + k 2 +(k + 1)2
  
Sk

k(k + 1)(2k + 1) 6
= Sk + (k + 1)2 = + (k + 1)2
6 6
k(k + 1)(2k + 1) + 6(k + 1)2 k(2k + 1) + 6(k + 1)
= = (k + 1)
6 6
2k 2 + 7k + 6 2k 2 + 4k + 3k + 6
= (k + 1) = (k + 1)
6 6
(k + 1) (k + 2)(2k + 3)
=
6
(k + 1) [(k + 1) + 1] [2(k + 1) + 1]
= (15.2)
6

where the form of (15.1) is identical to (15.2) after each n is replaced by n + 1, which is by
definition Sk+1 sum. Thus, after proving the first step, i.e. n = 1, then n = 2 is also valid, then
n = 3, etc., similar to the domino effect after the first block falls down.

(d) Knowing the sums of first n integers and sum of their respective squares, we write


n 
n 
n 
n 
n
i(4i − 3) = 4i 2 − 3i = 4 i2 − 3 i
i=1 i=1 i=1 i=1 i=1

n(n + 1)(2n + 1) n(n + 1)

=4 −3
6 2
4n(n + 1)(2n + 1) − 9n(n + 1) 8n2 − 5n
= = (n + 1)
6 6
n(n + 1)(8n − 5)
=
6

4. The series counter can be adjusted to start at any number.


n
(a) S = 23 + 33 + · · · + n3 = k3
k=2

but also,

n−1
S = 2 + 3 + ··· + n =
3 3
(j + 1)3
3

j =1
Solutions 337

or,

n−2
S = 2 + 3 + ··· + n =
3 3 3
(m + 2)3
m=0

 n ∞
1 2 3
(b) S= + + + ··· =
2 3 4 n=1
n+1

Exercise 15.2, page 333

1. Note which variable is the sum counter, and which one is the limit counter.
(a) Depending upon which operation is in progress, one or the other variable behaves as a
constant.
∞
! " ∞ ∞
 3 i 2  3 i2 3  3 2 3
lim + 1 = lim 2
+ = lim i +
n→∞
i=1
n n n→∞
i=1
n n n n→∞
i=1
n3 n
! ∞ ∞
"
3  2 3 
= lim i + 1
n→∞ n3 n i=1
i=1

= lim
+ 1)(2n + 1) + 3 n
3
n(n


n→∞ n3 2 6
2 n


1 n + 1 2n + 1
= lim +3
n→∞ 2 n n
!     "

1+ n n
2+ n
1 1
1 n
= lim +3
n→∞ 2 n n



⎡ ⎤
0 0
 
⎢1 1 1 ⎥
= lim ⎢
⎣ 1+  2+  + 3⎥
⎦
n→∞ 2 n n

1
= (1 + 0)(2 + 0) + 3
2
=4

Exercise 15.3, page 333

1. Geometric series
∞

S = a + ar + ar + ar + · · · + ar + · · · =
2 3 n
a r n−1 (a = 0)
n=1

may reach the following sums Sn for the first n terms:

338 15 Series

(a) Case: r = 1, where (a = 0)

∞
 ∞

Sn = a 1n−1 = a → ±∞; (a = 0)
n=1 n=1

In this case, therefore, the sum diverges.

(b) Case: r = 1.

Sn = a + ar + ar 2 + ar 3 + · · · + ar n−1 (15.3)
∴
r Sn = ar + ar 2 + ar 3 + · · · + ar n (15.4)

Consequently, the difference of (15.3) and (15.4)

  
+
Sn − r Sn = a − 
ar −
ar ar2 +  ar
ar2 − · · · −  n−1
+ ar n−1 − ar n
= a − ar n
∴
a(1 − r n )
Sn (1 − r) = a − ar n ∴ Sn =
1−r

(c) Case: |r| < 1 is most important because

0
r
a(1 − >
n
)
lim Sn = lim
n→∞ n→∞ 1−r
∴
∞
 a
a r n−1 = ; |r| < 1
n=1
1−r

2. First, we verify if given series is indeed geometric series, i.e. if the ratio of any two subsequent
terms is constant.
Solutions 339

(a) Given,

10 20
10 20 40 −
S =5− + − + ··· ∴ 3 = 9 = · · · = −2
3 9 27 5 10 3
−
3

i.e. the ratio of any two subsequent terms equals −2/3. Therefore,
 0  1  2
10 20 40 2 2 2
S =5− + − + ··· = 5 − +5 − +5 − + ···
3 9 27 3 3 3

which is to say that

∞  % %
 2 n−1 % 2% 2
S= ar n−1
=5 − %
a = 5, |r| = %− %% = < 1
n=1
3 3 3

∴
a 5 5

S= =  = =3
1−r 1 − − 32 5

3

(b) We find that this is geometric series as

∞
 ∞ ∞ ∞
4n 4 × 4n−1
S= 22n 31−n = (22 )n 3−(n−1) = n−1
=
n=1 n=1 n=1
3 n=1
3n−1

∞  n−1
4 4
= 4 ∴ |r| = >1
n=1
3 3

∴
S→∞

(c) This is geometric series as

∞
 ∞
 1
S= x n ; |x| < 1 ∴ = 1 xn =
n=1 n=1
1−x
 Special Functions
16

Important to Know

Delta function: It is defined as

δ(t) = 0; (t = 0)
; ∞
δ(t) dt = 1
−∞

Properties of Delta function: As the consequence of δ(t) non-zero value at only one point, i.e. t = 0,
at any other point t = 0 all products of x(t) δ(t) equal zero (which may be either delayed or advanced
relative to t = 0 by some constant t0 ). For that reason, Dirac function is also referred to as the
“sampling function.”

δ(−t) = δ(t)
x(t) δ(t − t0 ) = x(t0 ) δ(t − t0 )
; +∞
x(t) δ(t) dt = x(0)
−∞
; +∞
x(t) δ(t − t0 ) dt = x(t0 )
−∞
; +∞
x(τ ) δ(t − τ )dτ = x(t)
−∞
; +∞
x(t + t0 ) δ(t) dt = x(t0 )
−∞
; 
t2
x(t0 ); t1 < t0 < t2
x(t) δ(t − t0 ) dt =
t1 0; otherwise

© Springer Nature Switzerland AG 2021 341

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_16
342 16 Special Functions

16.1 Exercises

16.1 * Special Functions

1. Sketch a graph of

x(t) = es t where, s = σ + j ω and σ, ω ∈ R

in the following cases:

(a) σ = 0, ω = 0 (b) σ > 0, ω = 0 (c) σ = 0, ω = 0
(d) σ > 0, ω = 0 (e) σ = 0, ω = 0 (f) σ < 0, ω = 0

2. Sketch a graph and give definition of Dirac delta (sometimes called “the unit impulse”) function
δ(t).
3. Sketch a graph and give definition of Dirac comb (sometimes known as “sampling”) function
XT (t).
4. Sketch a graph and give definition of “step” (sometimes called “switch”) function u(t).
5. Sketch a graph and give definition of “sign” function sign(t).
6. Sketch a graph and give definition of “ramp” function r(t).
7. Sketch a graph and give definition of “rectangular” function Π(t).
8. Sketch a graph and give definition of “triangular” (sometimes called “hat” or “tent”) function
Λ(t).
9. Sketch a graph and give definition of “cardinal sine” function sinc(t).

16.2 ** Special Function, Interrelations

1. Problems with δ(t) function.

; ∞ ; ∞
e−at δ(t + 5) dt, (a > 0)
2
(a) sin(2.3π t) δ(t − 1) dt (b)
−∞ −∞
; ∞ ; ∞
e−at δ(t + 1) dt, (a > 0)
2
(c) (d) x(t) δ(t − a) dt, a ∈ R
0 −∞

2. Solve,
; t
(a) x(t) = u(τ ) dt, t > 0 (b) x(t) = u (t)
−∞

(c) u(t) as function of sign (t) (d) Π(t) as function of δ(t)

(e) Π(t) as function of u(t) (f) Π(t) as function of sign (t)
Solutions 343

Solutions

Exercise 16.1, page 342

1. Exponential function

x(t) = es t where, s = σ + j ω and σ, ω ∈ R

may be transformed as follows:

⎡ ⎤
⎢ ⎥
x(t) = e(σ +j ω )t = eσ t ej ω t = eσ t ⎣ cos(ω t) + j sin(ω t) ⎦
     
(x(t)) (x(t))

and, its complex conjugate version

x ∗ (t) = e(σ −j ω )t = eσ t e−j ω t = eσ t [cos(ω t) − j sin(ω t)]

That is, by using the Euler’s formula, complex exponent function is transformed into product of
a real exponential function and trigonometric functions. In other words, the complex exponential
function generates both cosine (along real axis) and sine (along perpendicular imaginary axis)
functions, Fig. 16.1.
Fig. 16.1 Example 16.1-1 (x(t))
t))
(x(
 x(t)

In the following three cases, we find.

(a) σ = 0, ω = 0: Therefore, degenerates into a constant.

x(t) = e(σ +j ω )t = e0 = const.

(b) σ > 0, ω = 0: Therefore, becomes a real exponential function.

x(t) = e(σ +j ω )t = eσ t
344 16 Special Functions

(c) σ = 0, ω = 0: Therefore, becomes a complex exponential function that can be used to

generate real sine/cosine functions.

x(t) = e(σ +j ω )t = ej ω t = [cos(ω t) + j sin(ω t)]

In addition, the sum and difference of complex exponents give

x(t) + x ∗ (t) = ej ω t + e−j ω t

= [cos(ω t) + j sin(ω t)] + [cos(ω t) − j sin(ω t)]
= 2 cos(ω t)
∴
ej ω t + e−j ω t
cos(ω t) =
2
x(t) − x ∗ (t) = ej ω t − e−j ω t
= [cos(ω t) + j sin(ω t)] − [cos(ω t) − j sin(ω t)]
= 2j sin(ω t)
∴
ej ω t − e−j ω t
sin(ω t) =
2j

(d) σ > 0, ω = 0: Generates cosine function along the real axis, Fig. 16.2, as well as sine
function in the complex plane, both with progressively increasing amplitudes.

Fig. 16.2 Example 16.1-1(d)

f (t)

t
0

es t cos (w t + j)
es t , (s > 0)
0
Solutions 345

(e) σ = 0, ω = 0: Generates cosine function along the real axis, Fig. 16.3, as well as sine
function in the complex plane, both with constant amplitudes.

Fig. 16.3 Example 16.1-1(e)

f (t) es t cos (w t + j )
es t , (s = 0)
1

0 t

−1

(f) σ < 0, ω = 0: Generates cosine function along the real axis, Fig. 16.4, as well as sine
function in the complex plane, both with progressively decreasing amplitudes.

Fig. 16.4 Example 16.1-1(f)

f (t) es t cos (w t + j)
es t , (s < 0)

t
0

2. Dirac delta function δ(t), Fig. 16.5, is defined by the means of limited integral as
; ∞
δ(t) dt = 1
−∞

δ(t) = 0, t = 0
346 16 Special Functions

Main properties of δ(t) are:

δ(t) = δ(−t)
x(t)δ(t − t0 ) = x(t0 )δ(t − t0 )
; ∞
x(t)δ(t) dt = x(0)
−∞
; ∞
x(t)δ(t − t0 ) dt = x(t0 )
−∞
; ∞
x(t + t0 )δ(t) dt = x(t0 )
−∞
; 
b
x(t0 ), t0 ∈ (a, b)
x(t)δ(t − t0 ) dt =
a 0, otherwise

Fig. 16.5 Example 16.1-2

d (t)

t
0

3. Dirac comb (sometimes known as “sampling”) function XT (t) is created by a series of

equidistant Dirac functions, Fig. 16.6.
∞

XT (t) = δ(t − kT )
k=−∞

where, T is sampling time, thus sampling frequency is f = 1/T .

Fig. 16.6 Example 16.1-3
T (t)

0
t
−2T −T 0 T 2T
Solutions 347

4. Step function u(t), Fig. 16.7, is defined as


0, t < 0
u(t) =
1, t ≥ 0

Fig. 16.7 Example 16.1-4

u(t)

t
0

5. Sketch a graph and give definition of “sign” function sign(t), Fig. 16.8.

⎧
⎪
⎨−1, t < 0
sign (t) = 0, t = 0
⎪
⎩
1, t ≥ 0

or, alternatively,

t |t|
sign (t) = = , ∀x = 0
|t| t

Fig. 16.8 Example 16.1-5 sign(t)

1

t
0

−1
0

6. Ramp function r(t), Fig. 16.9, is defined as follows:

348 16 Special Functions


0, t < 0
r(x) =
t, t ≥ 0

Fig. 16.9 Example 16.1-6

r(t)

t
0

0 1

7. Rectangular function Π(t), Fig. 16.10, is defined as follows:

⎧
⎪ T T
 ⎪
⎨1, − 2 ≥ t ≥ 2
t
Π =
T ⎪
⎪
⎩
0, otherwise

Fig. 16.10 Example 16.1-7 

Π t
T

t
0
T

− T2 0 T
2
Solutions 349

8. Triangle function Λ(t), Fig. 16.11, is defined as follows:

⎧
⎪ t T
⎪
⎪ + 1, − ≤ t ≤ 0
⎪
⎪ T /2 2
⎪
⎪
 ⎪
⎪
t ⎨
Λ = 1− t , 0≤t ≤ T
T ⎪
⎪
⎪
⎪ T /2 2
⎪
⎪
⎪
⎪
⎪
⎩0, otherwise

Fig. 16.11 Example 16.1-8 

Λ t
T

t
0
T

− T2 0 T
2

9. Cardinal sine function sinc (t) is defined as follows:

sin t sin(π t)
sinc (t) = or, sinc (t) =
t πt

where the last definition is normalized to sine period T , Fig. 16.12 (left). Equally important is
|sinc (t) | version, Fig. 16.12 (right).

sinc (t) | sinc (t) |

1 1

t
0
0
t

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3

Fig. 16.12 Example 16.1-9

350 16 Special Functions

Exercise 16.2, page 342

1. Product of a function x(t) and δ(t − t0 ) equals to zero, except on the point where δ(t) is located,
i.e. t = t0 . That is to say, while solving the definite integral, in effect we “search” the location
of δ(t) function. Consequently, only at that point function x(t0 ) keeps its value, in any other case
δ(t) x(t) = 0, t = t0 . Therefore, we write as follows.
(a) Given,
; ∞
sin(2.3π t) δ(t − 1) dt
−∞

we search the location of δ(t − 1). By definition, it is found at the position when its argument
equals zero, i.e. when t − 1 = 0 ∴ t = 1. Consequently, the product sin(2.3π t) δ(t − 1) is not
equal to zero only for t = 1. Thus, we write,
; ∞ ; ∞
sin(2.3π t) δ(t − 1) dt = sin(2.3π × 1) δ(t − 1) dt
−∞ −∞   
const.
; ∞ :1


= sin(2.3π )  δ(t−1) dt
−∞
= sin(2.3π ) ≈ 0.809

This integration is illustrated in Fig. 16.13, which shows the position of δ(t − 1) function relative
to sin(2.3π t). The multiplication operation is done before the integration and, obviously, only
one point i.e. sin(2.3π ) is not multiplied by zero. This example illustrates the idea of “sampling”
(or, saving) certain points of a given function by using Dirac function.

Fig. 16.13
f (t)
Example 16.2-1(a)
1 f (1)

0 t

−1
sin (2.3w t)
d (t − 1)
0 1 2

(b) First, we find position of δ(t + 5), i.e. when t + 5 = 0 ∴ t = −5, which is within the
bonds of definite integral, then we write
Solutions 351

; ∞ ; ∞
−at 2
e−a(−5) δ(t + 5) dt
2
e δ(t + 5) dt, (a > 0) =
−∞ −∞
; ∞ :1

= e−25a  5)dt
+ = e−25a
 
δ(t
−∞

(c) However, if δ(t) is not found within the bonds of definite integral, then all relevant products
equal zero because δ(t) = 0, see Fig. 16.14.
That is to say,
; ∞
e−at δ(t + 1) dt, (a > 0)
2

0
 
= (t = −1) ∈ / [0, ∞]
; ∞
:0

e−at δ(t+1) dt
2
=
0
; ∞
e−at × 0 dt = 0
2
=
0

Fig. 16.14
f (t) 2
Example 16.2-1(c) e−at
1 d (t + 1)

0 t

−2 −1 0 1 2

(d) Considering that δ(t − a) is certainly found within interval t ∈ [−∞, ∞], then in general
case we write
; ∞ ; ∞
:1

x(t) δ(t − a) dt = δ(t−a) dt = x(a)
x(a) 
−∞ −∞
352 16 Special Functions

2. (a) Step function equals to zero for negative t, and it is constant (i.e. equals one) for t ≥ 0, thus
we write (using temporarily variable τ )

0
; t ; 0 ; t ; 0 >
 ; t
u(τ ) dτ = u(τ ) dτ + u(τ ) dτ = 
0 dτ + 1 dτ
−∞ −∞

0 −∞ 0
; %t
t %
= dτ = τ %% = t
0 0

In conclusion,
; 
t
t; t ≥ 0
u(τ ) dτ = = r(t)
−∞ 0; t < 0

which is logical, considering that by definite integral we calculate surface of the rectangular area
(whose height equals to one) and its length linearly increases as t increases.

(b) We can deduce derivative u (t) by observing graphical interpretation, Fig. 16.15.
While keeping in mind that derivative of a constant equals to zero, and that in general case it
shows the function’s rapidity of change over given interval, we conclude that if the interval equals
to zero, the change must be infinite (due to division by zero). By definition, that infinite change
in amplitude over zero interval describes Dirac function.

d u(t)
= δ(t)
dt

Fig. 16.15
u(t) const.
Example 16.2-2(b) 1

const. t
0
Δx = 0
d u(t)
dt = d (t)

u (t) = 0 u (t) = 0 t
0
Δx = 0
0
Solutions 353

(c) Sign(t) function can be derived from u(t) by simple transformations as illustrated in
Fig. 16.16.
In the first step u(t) is multiplied by factor two, which doubles its amplitude for t > 0. At this
stage we note that sign (t) function has the same form, but appears shifted down by step of one.
Thus,

sign (t) = 2u(t) − 1

Fig. 16.16
2u(t)
Example 16.2-2(c) 2
−1 2
1
t
0

2u(t) − 1 = sign(t)
1
2 t
0

−1
0

(d) Following the same reasoning as in Example 16.2-2(b), we deduce.

Relation between Π(t) and δ(t) is illustrated in Fig. 16.17. There are two rapid changes at the
edges of Π(t) that generate Dirac functions at t = −T /2 and t = T /2 moments. We note that the
leading edge of Π(t) jumps from zero to one, thus the derivative is positive. However, the trailing
edge jumps from one to zero, thus its associated derivative is negative. In summary,
 
dΠ T T
=δ t+ −δ t −
dt 2 2

Fig. 16.17
Example 16.2-2(d)
354 16 Special Functions

(e) Simple combinations of basic functions can be used to generate another function. One
possible relation between Π(t) and u(t) is illustrated in Fig. 16.18. For example, one step function
can be advanced and one delayed in time. The difference between these two u(t) functions
produces Π(t) as

  
t T T
Π =u t+ −u t −
T 2 2

which is simply deduced by calculating the difference within each piecewise interval. For
example, for t ∈ [−T /2, T /2],

u(t + T /2) = 1; u(t − T /2) = 0

∴
u(t + T /2) − u(t − T /2) = 1 − 0 = 1

Fig. 16.18
Example 16.2-2(e) 1 u(t + T /2)
t
0
u(t − T /2)
t
0

Π t
T
1
t
0
T 0 T
−
2 2

(f) One possible relation between Π(t) and sign (t) is illustrated in Fig. 16.19. On the left side,
it is shown how to create two sign (t) functions: one advanced by τ/2, and one first delayed by
τ/2 then inverted sign. In the next step illustrated on top of Fig. 16.19 (right), the two sign (t)
functions are scaled by factor of 1/2 so that their sum adds up to one. The addition of these two
functions produces Π(t/τ ), as summarized by equation
 
t 1 τ 1  τ
Π = sign t + − sign t −
τ 2 2 2 2
Solutions 355

Fig. 16.19 1 sign(t + W2 ) 1 W 1

1 − 2 sign(t − 2 ) 2 sign(t + W2 )
Example 16.2-2(f) t
0 2
−1 t
0
1 sign(t − W2 ) 1
−
t 2
0 
−1 Π t
W
1
1
t
0 0
t
W
−1 −sign(t − 2 )
−W/2 0 W/2 −W/2 0 W/2
 Convolution Integral
17

Important to Know

Convolution integral: It is defined as

; ∞
z(t) = x(t) ∗ y(t) = x(τ ) y(t − τ ) dτ (17.1)
−∞

where, τ is the provisional integration variable. That is to say, at any given instance of time t, the
products of the two functions x(τ ) and y(t − τ ) are integrated relative to the parametric variable τ .

17.1 Exercises

17.1 * Transformations of Continuous Functions

1. Given x(t) function in Fig. 17.1, show the following graphs:

(a) Delayed: x(t − 2)
(b) Compressed: x(2t)
(c) Expanded: x(0.2t)
(d) Time inverted: x(−t)

Fig. 17.1 Example 17.1-1 x(t)

1

t
0

2 1 0 1 2 3

© Springer Nature Switzerland AG 2021 357

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_17
358 17 Convolution Integral

2. Given x(t) function in Fig. 17.2, show graph of y(t) = x(2t + 3) .

Fig. 17.2 Example 17.1-2 x(2t)

1
t
0

1
2 0 2

3. Given x(t) function in Fig. 17.3, show graph of y(t) = x(− 2t − 4) .

Fig. 17.3 Example 17.1-3

2 x(t)
1
t
0
1
−1 0 1 2

4. Given x(t) function in Fig. 17.4, decompose x(t) into a sum of one even xp (t) and one odd xi (t)
function so that x(t) = xp (t) + xi (t), then show their respective graphs.

Fig. 17.4 Example 17.1-4 x(t)

1

t
0

3 2 1 0 1 2 3

5. Show graphical representation of the following function:

x(t) = 3u(t) + r(t) − r(t − 1) − 5u(t − 2)

6. Show the analytical form of function x(t) given in Fig. 17.5:

Fig. 17.5 Example 17.1-6 x(t)

2

1
t
0
2 0 2

7. Show the analytical form of function x(t) given in Fig. 17.6:

17.1 Exercises 359

Fig. 17.6 Example 17.1-7 x(t)

2

1
t
0
2 0 2 4 6

17.2 ** Convolution, Piecewise Linear Functions

1. Calculate y(t) = x(t) ∗ h(t), where x(t) and h(t) are given in Fig. 17.7.

x(t) h(t)
1 1

t t
0 0

1 0 1 1 0 3

Fig. 17.7 Example 17.2-1

2. Calculate z(t) = x(t) ∗ y(t), where x(t) and y(t) are given in Fig. 17.8.

x(t) y(t)
1 1

t t
0 0

0 0 1 3

Fig. 17.8 Example 17.2-2

3. Calculate y(t) = u(t) ∗ u(t).

4. Solve convolution y(t) = x(t) ∗ h(t), given

x(t) = [u(t + 1) − u(t − 1)] t

h(t) = [u(t + 2) − u(t)]
t  t 
5. Calculate x(t) = Λ 2
∗Λ 2
.
360 17 Convolution Integral

17.3 * Convolution, Continuous Functions

1. Solve convolution z(t) = x(t) ∗ y(t), where x(t) = t 3 u(t), and y(t) = t 2 u(t).
2. Solve convolution w(t) = x(t) ∗ v(t), given
 
1 − t; 0≤t ≤1 e−t ; t ≥0
x(t) = and, v(t) =
0; otherwise 0; otherwise

17.4 ** Energy and Power of Continuous Functions

1. Given x(t) functions and calculate their respective energies and powers.
(a) x(t) = A (b) x(t) = sin(t) (c) x(t) = e−t u(t)
(d) x(t) = A ej ω 0 t (e) x(t) = u(t) (f) x(t) = 2Π(t − 1)

Solutions

Exercise 17.1, page 357

1. In order to carry on the required transformations, visually, it helps to pick some of the
typical points in the given function, for example x(−1), x(0), x(2), and follow their respective
movements.
(a) A function is delayed or advanced when a constant is added to the initial argument, for
example if

x(0) → x(t − 2)

where “−2” is the shifting constant. New coordinate positions are found by equalizing the two
arguments, e.g.

0=t −2 ∴ t =2

Similarly, we find transformed positions of the other typical points, and conclude that this
transformation shifts the original function in the positive direction by two steps, Fig. 17.9.
Solutions 361

Fig. 17.9 Example 17.1-1(a)

x(t − 2)
1

t
0

0 1 2 3 4

(b) Similarly, for compressed function x(2t) we write,

x(0) : 2t = 0 ∴ t = 0
x(2) : 2t = 2 ∴ t = 1
2
x(−1) : 2t = −1 ∴ t = −
2
···

here, the origin point does not move. It is evident that when the function’s argument is multiplied
by a constant greater that one, visually, the overall effect is “compression” of the original form,
Fig. 17.10.
Fig. 17.10
Example 17.1-1(b) x(2t)
1

t
0

1 0 1 2 3

(c) However, the function’s argument is multiplied by a constant lesser than one; visually the
overall effect is “expansion” of the original form. In the case of x(0.2t), Fig. 17.11, we find that

x(0) : 0.2t = 0 ∴ t = 0
x(2) : 0.2t = 2 ∴ t = 10
x(−1) : 0.2t = −1 ∴ t = −5
362 17 Convolution Integral

Fig. 17.11
Example 17.1-1(c) x(0.2t)
1

t
0

5 0 5 10

(d) Time inverted function x(−t) is simply revolved around the vertical axes, Fig. 17.12, where

x(0) : − t = 0 ∴ t = 0
x(1) : − t = 1 ∴ t = −1
···

Fig. 17.12
Example 17.1-1(d) x(−t)
1

t
0

3 2 1 0 1 2

2. When multiple transformations are required, there are multiple ways to execute them.
(a) Method 1:
Transformation y(t) = x(2t + 3) consists of one compression (i.e. 2t) and one delay (i.e. +3). If
the delay is applied before the compression, we find

delay compression
x(t) −−−−−→ x(t + 3) −−−−−−−−→ x(2t + 3)

In the first step, the delay x(t + 3), we write

x(0) : t + 3 = 0 ∴ t = −3

which is shown in Fig. 17.13.

Fig. 17.13
Example 17.1-2(a) 1 x(t + 3)

t
0

−1
6 3 0 3
Solutions 363

In the second transformation, the compression, we keep the delay and while the time is
compressed i.e. x(2t) so that we write

x(0) : 2t = 0 ∴ t = 0
x(1) : 2t = 1 ∴ t = 1/2
x(−3) : 2t = −3 ∴ t = −3/2
···

as shown in Fig. 17.14.

Fig. 17.14
Example 17.1-2(a) 1 x(2t + 3)

t
0

−1
4 2 0 2

(b) Method 2:
Alternatively, the two transformations in y(t) = x(2t + 3) may be executed so that the
compression is applied before the delay, thus we find
 
compression delay 3
x(t) −−−−−−−−→ x(2t) −−−−−→ x(2t + 3) = x 2 t +
2

That is to say, application of compression x(2t) results in the function form as in Fig. 17.15

Fig. 17.15
Example 17.1-2(b) 1 x(2t)

t
0

−1
2 0 2

Second transformation, the delay of (t + 3/2) shifts the function horizontally, resulting in the
same result as in Fig. 17.14.

3. By focusing, for example, to point x(0) we then compare when the new argument also equals zero.
Given x(t) function, one way to derive the total transformation may be to apply delay x(t − 4)
first, followed by time inversion x(−t) along with the compression factor of 0.5t:
Delayed function is shown in Fig. 17.16,
364 17 Convolution Integral

Fig. 17.16 Example 17.1-3

2 x(t − 4)
1
t
0
−1
2 4 6

while the time inversion with the scaling factor we find as, for example for point

t
x(4) : − = 4 ∴ t = −8
2

so that the total transformation is as in Fig. 17.17,

Fig. 17.17 Example 17.1-3 

2 x − 2t − 4
1
t
0
−1
 12  10 8 6

Using some other order of transformations does not change the final result.
4. Given x(t) function, its accompanying even and odd functions are found as

x(t) + x(−t)
xp (t) = (17.2)
2
x(t) − x(−t)
xi (t) = (17.3)
2

We verify that

x(t) + x(−t) x(t) − x(−t)

xp (t) + xi (t) = +
2 2
x(t) x(t) x(−t) x(−t)
= + + −
2 2 2 2
= x(t)

We calculate even and odd functions (17.2) and (17.3) by using graphical technique, i.e. by
carrying on calculations within each individual interval.
In order to derive even function xp (t), as defined by (17.2) first, we derive x(−t) by mirroring
x(t) around the vertical axis, Fig. 17.18 (middle). Then, additions are within each piecewise linear
interval. For example, we add x(t) + x(−t) as

x(t) + x(−t) 1+0 1

t = 0− : = =
2 2 2
Solutions 365

x(t) + x(−t) 0+1 1

t = 0+ : = =
2 2 2

that is to say, on both sides of t = 0 we find xp (0) = 1/2. Similarly, we calculate xp (t)
at other typical points of t ∈ [−3, 3], for example xp (±2), xp (±1), and create graph in
Fig. 17.18 (bottom), which is the literal implementation of (17.2).

Fig. 17.18 Example 17.1-4

x(t)
1

t
0

x(−t)
1

t
0

x p (t)
1

t
0

2 1 0 1 2

By using the same technique, we derive odd function xi (t), as defined by (17.3) by calculating
the difference at the key points. For example, at t = 0 we calculate one minus one and divide by
two, thus xi (0) = 0, Fig. 17.19.

Fig. 17.19 Example 17.1-4 x p (t)

0.5
t
0

−0.5
−2 −1 0 1 2

5. Function x(t) is synthesized by adding the basic functions u(t) and r(t) as

x(t) = 3u(t) + r(t) − r(t − 1) − 5u(t − 2)

Graphical representation of each term is shown in Fig. 17.20, while the sum is highlighted in
the background. The addition is done by simply adding values of the four functions at typical
coordinates. For example, for t = 0+ , by definition of basic functions we write
366 17 Convolution Integral

3u(t) = 3u(0) = 3(1) = 3

r(t) = r(0) = 0
−r(t − 1) = −r(0 − 1) = −(0) = 0
−5u(t − 2) = −5u(0 − 2) = −5(0) = 0

thus, x(0) = 3 + 0 + 0 + 0 = 3. Similar additions are repeated for t ∈ [1, 2, 3].

Fig. 17.20 Example 17.1-5

5 x(t) r(t)

3u(t)
t
0

−5u(t − 2)
−r(t − 1)
−5
−2 0 2 4

6. By inspection of function x(t) given in Fig. 17.5, we write the following sum:

x(t) = 2u(t + 2) − u(t + 1) − r(t − 1) + r(t − 2)

which is easily verified either by calculating the sums at various points in time, or by graphical
method as in Example 17.1-5.
7. By inspection of function x(t) given in Fig. 17.6, we write the following sum:

x(t) = u(t) + r(t − 1) − 2r(t − 2) + r(t − 3) + u(t − 4) − 2u(t − 5)

which is easily verified either by calculating the sums at various points in time, or by graphical
method as in Example 17.1-5. Note how x(t) is formed in t ∈ [1, 3] interval by correctly adding
“r(t − 1) − 2r(t − 2) + r(t − 3)” ramp functions.

Exercise 17.2, page 359

1. Convolution integral (17.1) contains the product of x(τ ) and h(t − τ ) functions. We derive these
two functions by using graphical method. In sequence of transformations, we convert x(t) →
x(τ ), as well as h(t) → h(τ ) → y(−τ ) → h(t − τ ) function, as in Fig. 17.21.
(a) In the first step, Fig. 17.21a, b, the change of variable t into provisional integration variable
τ does not change the original forms of neither x(t) nor h(t) functions, thus only transformations
of h(t) function are illustrated.
Solutions 367

(b) In the second step, h(−τ ) is the time inverted version of function h(τ ), i.e. is mirrored
around the vertical axis, Fig. 17.21c. Note that the origin τ = 0 reference is still known and,
visually, the form of h(t) stays unchanged.

(c) In the third step, h(−τ ) is shifted to an arbitrary position t thus h(t − τ ), Fig. 17.21d.
Consequently, depending upon relative position of t in reference to the origin point τ = 0 the
following cases are possible:

Fig. 17.21 Example 17.2-1

h(t) a) h(−t ) c)
1 1

t t
0 0

−10 3 −3 0 1
h(t ) b) h(t − t ) d)
1 1

t t
0 0

− 10 3 t− 3 t+ 1

Case 1 : (t + 1) ≤ −1:
The leading edge of h(t − τ ) still does not surpass (τ = −1) point, that is to say h(t − τ ) does
not overlap x(τ ) as long as

t + 1 ≤ −1 ∴ t ≤ −2

Fig. 17.22
Example 17.2-1 (case 1) h(t − t ) x(t )
1

t
0

t −3 t + 1 −1 0 1

That being the case, it is irrelevant where exactly (t + 1) is relative to the trailing edge of x(τ ),
see Fig. 17.22. What is important is that under this condition, due to either h(t − τ ) or x(τ ) equals
zero, their product is therefore

x(τ ) h(t − τ ) = 0, ∀τ
∴
368 17 Convolution Integral

; ∞ ; ∞
y(t) = x(τ ) h(t − τ ) dτ = 0 dτ = 0 (t ≤ −2)
−∞ −∞

Case 2 : (t + 1) > −1 and (t + 1) ≤ 1:

The leading edge of h(t − τ ) is found within τ ∈ [−1, 1] interval determined by x(τ ); however,
the trailing edge of h(t − τ ) is still outside of this interval, i.e.

t + 1 > −1 ∴ t > −2 and

t +1≤1 ∴ t ≤0

Fig. 17.23
Example 17.2-1 (case 2) h(t − t ) x(t )
1

t
0
t +1
t −3 −1 0 1

Under this condition, see Fig. 17.23, the overlapping surface is found in the interval τ ∈ [−1, t +
1] where h(t − τ ) = 1 and x(τ ) = 1; therefore, the convolution integral (i.e. the overlapping
surface area) is found to be
; ; %t+1
∞ t+1 %
y(t) = x(τ ) h(t − τ ) dτ = (1) (1) dτ = τ %% = t + 1 − (−1)
−∞ −1 −1

= t + 2 (−2 ≤ t ≤ 0)

Case 3 : (t + 1) > 1 and (t − 3) ≤ −1:

The leading edge of h(t − τ ) is found at (t + 1 > 1) while its trailing edge is still (t − 3 ≤ −1),
i.e.

t + 1 > 1 ∴ t > 0 and

t − 3 ≤ −1 ∴ t ≤ 2

Fig. 17.24
Example 17.2-1 (case 3) h(t − t ) x(t )
1

t
0
t −3 t +1
−1 0 1
Solutions 369

Within these boundaries of t, see Fig. 17.24, the overlapping surface is found in the interval τ ∈
[−1, +1] where h(t − τ ) = 1 and x(τ ) = 1, in other words it is maximal overlapping surface
possible whose rectangular area equals two,
; ; %1
∞ 1 %
y(t) = x(τ ) h(t − τ ) dτ = (1) (1) dτ = τ %%
−∞ −1 −1

= 2 (0 ≤ t ≤ 2)

Case 4 : (t − 3) ≥ −1 and (t − 3) ≤ 1:
The trailing edge of h(t − τ ) is found within τ ∈ [−1, 1] interval determined by x(τ ), i.e.

t − 3 ≥ −1 ∴ t ≤ 2 and
t −3<1 ∴ t <4

Fig. 17.25
Example 17.2-1 (case 4) x(t ) h(t − t )
1

t
0
t −3 t +1
−1 0 1

Within these boundaries of t, see Fig. 17.25, the overlapping surface is found in the interval τ ∈
[t − 3, +1] where h(t − τ ) = 1 and x(τ ) = 1, i.e.
; ; %1
∞ 1 %
y(t) = x(τ ) h(t − τ ) dτ = (1) (1) dτ = τ %%
−∞ t−3 t−3

= 4 − t (2 ≤ t ≤ 4)

Case 5 : (t − 3) ≥ 1:
The trailing edge of h(t − τ ) is found at (t − 3 ≥ 1), see Fig. 17.26, i.e.

t −3≥1 ∴ t ≥4

Fig. 17.26
Example 17.2-1 (case 5) x(t ) h(t − t )
1

t
0
t −3 t +1
−1 0 1
370 17 Convolution Integral

That being the case, it is irrelevant where exactly (t − 3) is relative to the leading edge of x(τ ),
see Fig. 17.26. What is important is that within these boundaries of t, due to either h(t − τ ) or
x(τ ) equals zero, their product is therefore

x(τ ) h(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
y(t) = x(τ ) h(t − τ ) dτ = 0 dτ = 0 (t ≥ 4)
−∞ −∞

Summary: The total convolution integral y(t) shows the progression of the overlapping surface
area between x(τ ) and h(t − τ ), Fig. 17.27, as
⎧
⎪
⎪ 0 t ≤ −2
⎪
⎪
⎪
⎪
⎨2 + t −2 ≤ t ≤ 0
y(t) = 2 0≤t ≤2
⎪
⎪
⎪
⎪4 − t 2≤t ≤4
⎪
⎪
⎩
0 t ≥4

Fig. 17.27 Example 17.2-1

2 y(t)

1
t
0
−2 0 2 4

2. Convolution integral (17.1) contains the product of x(τ ) and y(t − τ ) functions. We derive these
two functions by using graphical method. In sequence of transformations, we convert u(t) →
u(τ ), as well as y(t) → y(τ ) → y(−τ ) → y(t − τ ) function, as in Fig. 17.28.
(a) After the first step, Fig. 17.28a, b, the change of variable t into temporarily integration
variable τ does not change the initial form of neither x(t) nor y(t) functions, thus only y(τ )
function is illustrated.

(b) In the second step, in order to create y(−τ ), function y(τ ) is simply mirrored around the
vertical axis, Fig. 17.28c. Note that the origin τ = 0 point is still known, and the form of y(t)
visually stays unchanged.

(c) In the third step, y(−τ ) is shifted to an arbitrary position t, thus y(t − τ ) in Fig. 17.28d
shows only the positions of rising (found at −3 + t) and falling (found at −1 + t) edges relative
to τ . Consequently, depending upon relative position of t in reference to the origin point τ = 0
the following three cases are possible:
Solutions 371

Fig. 17.28 Example 17.2-2

y (t ) y( − t )
1 1

t t
0 0
a) c)
0 1 3 −3 −1 0
y(t ) y (t − t )
1 1

t t
0 0
b) d)
0 1 3 −3 + t −1 + t

Case 1 : (t − 1) ≤ 0:
The leading edge of y(t − τ ) is still in the negative side, i.e.

t −1≤0 ∴ t ≤1

Fig. 17.29
Example 17.2-2(case 1) y(t − t ) x(t )
1

t
0
t −3 t −1
0

That being the case, it is irrelevant where exactly t is relative to the origin, see Fig. 17.29. What is
important is that within these boundaries of t, because always either of the two function equals
zero, their product is therefore

x(τ ) y(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
z(t) = x(τ ) y(t − τ ) dτ = 0 dτ
−∞ −∞

= 0 (t ≤ 1)

Case 2 : (t − 1) > 0 and (t − 3) ≤ 0:

The leading edge of y(t − τ ) is shifted into the positive side; however, the trailing edge is still in
the negative side, i.e.
372 17 Convolution Integral

t − 1 > 0 ∴ t > 1 and

t −3≤0 ∴ t ≤3

Fig. 17.30
Example 17.2-2(case 2) y(t − t ) x(t )
1

t
0
t −3 t −1
0

Within these boundaries of t, see Fig. 17.30, the overlapping surface is found in the interval τ ∈
[0, t −1] where y(t −τ ) = 1 and x(τ ) = 1; therefore, the convolution integral (i.e. the overlapping
surface area) results in
; ; %t−1
∞ t−1 %
z(t) = x(τ ) y(t − τ ) dτ = (1) (1) dτ = τ %%
−∞ 0 0

= t − 1 (1 ≤ t ≤ 3)

Case 3, (t − 3) > 0:
The trailing edge of y(t − τ ) is shifted into the positive side, see Fig. 17.31, i.e.

t −3>0 ∴ t >3

Fig. 17.31
Example 17.2-2(case 3) x(t ) y(t − t )
1

t
0
t −3 t −1
0

Therefore, the overlapping surface area is at its maximum (i.e. the full area of Π (t) function, and
henceforth stays constant)
; ; %t−1
∞ t−1 %
z(t) = x(τ ) y(t − τ ) dτ = (1) (1) dτ = τ %% = (t − 1) − (t − 3)
−∞ t−3 t−3

= 2 (t > 3)

Summary:
The total convolution integral shows progression of the overlapping, Fig. 17.32, as
Solutions 373

⎧
⎪
⎨0 t ≤1
z(t) = t − 1 1≤t ≤3
⎪
⎩
2 t >3

Fig. 17.32 Example 17.2-2

2 z(t)

t
0
0 1 3

3. By applying the same technique as in the previous examples, we derive u(τ ) and u(t −τ ) functions
by using graphical method, see Fig. 17.33.
After the first step, Fig. 17.33a, b, the change of variable t into provisional integration variable
τ does not affect the initial form of u(t). Then, u(τ ) is simply mirrored around the vertical axis
then shifted horizontally by t, Fig. 17.33c, d.

Fig. 17.33 Example 17.2-3

a) u(t) c) u(−t )
1 1

t t
0 0

0 0
b) u(t ) d) u(t − t )
1 1

t t
0 0

0 t
There are two distinct cases to solve.
Case 1 : t ≤ 0:
The leading edge of u(t − τ ) is still in the negative side, i.e.

t ≤1

It is irrelevant where exactly t is relative to the origin, see Fig. 17.34. What is important is
that under this condition, because always either of the two function equals zero, their product
is therefore

u(τ ) u(t − τ ) = 0, ∀τ
∴
374 17 Convolution Integral

; ∞ ; ∞
z(t) = u(τ ) u(t − τ ) dτ = 0 dτ
−∞ −∞

= 0 (t ≤ 0)

Fig. 17.34
Example 17.2-3 (case 1) u(t − t ) u(t )
1

t
0

t 0

Case 2 : t > 0:
The leading edge of u(t − τ ) is shifted into the positive side, Fig. 17.35, i.e.

t >1

thus,
; ; %t
∞ t %
z(t) = u(τ ) u(t − τ ) dτ = (1) (1) dτ = τ %%
−∞ 0 0

= t (t > 0)

(Surface of the overlapping rectangular area is calculated as height (equals one) times its length
(equals t).)

Fig. 17.35
Example 17.2-3 (case 2) u(t − t ) u(t )
1

t
0

0 t

Summary:
The total convolution integral is therefore r(t) function, see Fig. 17.36, or written in equally
elegant product form of linear and step functions as

0 t ≤0
z(t) = = r(t) or, z(t) = t u(t)
t t >0
Solutions 375

Fig. 17.36
u(t)
Example 17.2-3 (summary) 1
z(t) = r(t) = t u(t)
t
0
−1
t
0 1

4. We synthesize h(t) = [u(t + 2) − u(t)] as the sum of u(t) functions as illustrated in Fig. 17.37.

Fig. 17.37 Example 17.2-4 u(t + 2)

1
t
0
−1
−u(t)
−2 1
u(t + 2) − u(t)
1
t
0
−1

−2 0
h(t − t )
1
t
0

−1
t t +2

While the syntheses of x(t) = [u(t + 1) − u(t − 1)] t as in the following steps, see Fig. 17.38.
376 17 Convolution Integral

Fig. 17.38 Example 17.2-4 u(t + 1)

x(t)
1
t
0
−1
−u(t − 1)
−1 1
u(t + 1) − u(t − 1)
1
t
0
−1
t
−1 1
[u(t + 1) − u(t − 1)]t
1
t
0
−1

−1 0 1

There are four distinct cases to solve.

Case 1 : t + 2 ≤ −1:
The leading edge of h(t − τ ) is still far away from τ = −1, that is to say

t + 2 ≤ −1 ∴ t ≤ −3

Fig. 17.39
h(t − t ) x(t )
Example 17.2-4 (case 1) 1
t
0
−1

t t +2 −1 0 1

That being the case, it is irrelevant where exactly t is relative to the origin, see Fig. 17.39. What
is important is that under this condition, because always either of the two function equals zero,
their product is therefore

x(τ ) h(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
y(t) = x(τ ) h(t − τ ) dτ = 0 dτ
−∞ −∞
Solutions 377

= 0 (t ≤ −3)

Case 2 : t + 2 > −1 and t + 2 ≤ 1

The leading edge of h(t − τ ) is found within the interval τ ∈ [−1, 1], that is to say

t + 2 > −1 ∴ t > −3 and

t + 2 ≤ 1 ∴ t ≤ −1

Fig. 17.40
h(t − t ) x(t )
Example 17.2-4 (case 2) 1
t
0
−1
t +2
t −1 0 1

Within these boundaries of t, see Fig. 17.40, the overlapping surface is found in the interval τ ∈
[−1, t + 2] where h(t − τ ) = 1 and x(τ ) = t, consequently
; ; %
∞ t+2
1 2 %%t+2
y(t) = x(τ ) h(t − τ ) dτ = (τ ) (1) dτ = τ %
−∞ −1 2 −1

1# $ 12 
= (t + 2)2 − (−1)2 = t + 4t + 3
2 2
1
= (t + 1)(t + 3), (−3 ≤ t ≤ −1)
2

Case 3 : t > −1 and t ≤ 1

The trailing edge of h(t − τ ) is found within the interval τ ∈ [−1, 1], that is to say

t > −1 and t ≤ 1

Fig. 17.41
x(t ) h(t − t )
Example 17.2-4 (case 3) 1
t
0
−1

−1 t 0 1 t +2

Within these boundaries of t, see Fig. 17.41, the overlapping surface is found in the interval τ ∈
[t, 1] where h(t − τ ) = 1 and x(τ ) = t, consequently
378 17 Convolution Integral

; ; %
∞ 1
1 2 %%1
y(t) = x(τ ) h(t − τ ) dτ = (τ ) (1) dτ = τ %
−∞ t 2 t
1  
= (1 − t 2 ) = (a 2 − b2 ) = (a − b)(a + b)
2
1
= − (t − 1)(t + 1), (−1 ≤ t ≤ 1)
2

Case 4 : t > 1
The trailing edge of h(t − τ ) is found after τ = 1, that is to say

t >1

That being the case, it is irrelevant where exactly t is relative to the origin, see Fig. 17.42. What
is important is that under this condition, because always either of the two function equals zero,
their product is therefore

x(τ ) h(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
y(t) = x(τ ) h(t − τ ) dτ = 0 dτ
−∞ −∞

= 0 (t > 1)

Fig. 17.42
x(t ) h(t − t )
Example 17.2-4 (case 4) 1
t
0
−1

−1 0 1 t t +2

Summary:
The total convolution integral is therefore y(t) function, see Fig. 17.43, as
⎧
⎪
⎪ 0 t < −3
⎪
⎪
⎨ 1 (t + 1)(t + 3) (−3 ≤ t ≤ −1)
y(t) = 2 1
⎪
⎪ − 2 (t − 1)(t + 1) (−1 ≤ t ≤ 1)
⎪
⎪
⎩0 t >1
Solutions 379

Fig. 17.43 1
Example 17.2-3 (summary) y(t)

t
0

−1
−3 −2 −1 0 1 2

5. We derive x(t) = Λ(τ ) and x(t − τ ) = Λ(t − τ ) functions by using graphical method, see
Fig. 17.44. After the first step, Fig. 17.44a, b, the change of variable t into provisional integration
variable τ does not affect the initial form of x(t). Then, x(τ ) is simply mirrored around the
vertical axis then shifted horizontally by t, Fig. 17.44c, d.

Fig. 17.44 Example 17.2-5

a) x(t) c) x(−t )
1 1

t t
0 0

−1 0 1 −1 0 1
b) x(t ) d) x(t − t )
1 1

t t
0 0

−1 0 1 t −1 t t +1

It is important to specifically write linear equations that correspond to each linear section of the
given x(t) and x(t − τ ) as:
⎫ ⎧
x(t) = 1 + t, −1 ≥ t ≥ 0⎪
⎪ ⎪
⎬ ⎨x(t − τ ) = 1 + τ − t, t − 1 ≤ τ ≤ t
x(t) = 1 − t, 0 ≤ t ≤ 1 ∴ x(t − τ ) = 1 − τ + t, t ≤ τ ≤ t + 1
⎪
⎪ ⎪
⎩
⎭ x(t − τ ) = 0, sinon
x(t) = 0, |t| ≥ 1

We note that, due to Λ(t) being a pair function, x(τ ) and x(−τ ) functions have identical linear
equations even though the argument τ is transformed into −τ , Fig. 17.44. In order to avoid the
calculation errors, we must pay attention which two equations are found in any given interval, as
illustrated while resolving the following cases:
There are six distinct cases to solve.
Case 1 : t + 1 < −1
The leading non-zero point at (t + 1) of x(t − τ ) is still in the left side of τ = −1, i.e.

t + 1 < −1 ∴ t < −2
380 17 Convolution Integral

Fig. 17.45
Example 17.2-5 (case 1) 1 x(t − t ) x(t )

t
0
t −1 t t +1
−1 0 1

That being the case, it is irrelevant where exactly t is relative to the origin, see Fig. 17.45. What
is important is that within these boundaries of t, because always either of the two function equals
zero, therefore their product is

x(τ ) x(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
z(t) = x(τ ) x(t − τ ) dτ = 0 dτ
−∞ −∞

= 0 (t < −2)

Case 2 : t + 1 ≥ −1, t + 1 ≤ 0
Once the leading edge of x(t − τ ) found at (t + 1) crosses τ = −1 while at the same time
(t < −1), i.e.

t + 1 ≥ −1 ∴ t ≥ −2 and
∴ −2 ≤ t ≤ −1
t < −1

a non-zero overlapping surface is created by (1 − τ + t) and (1 + τ ) linear segments, which

defines interval (a) : [−1, t + 1] in Fig. 17.46.

Fig. 17.46
1 x(t − t ) x(t )
Example 17.2-5 (case 2) a

t
0
t −1 t t +1
−1 0 1

Only in this interval the convolution product is defined with these two linear segments:
; ∞ ; t+1
z(t) = x(τ ) x(t − τ ) dτ = (1 + τ ) (1 − τ + t) dτ
−∞ −1
; t+1
= (1 − τ
+ t + τ
+ tτ − τ 2 ) dτ
−1
Solutions 381

%t+1 %t+1 %t+1

% t % 1 %
= (1 + t) τ %% + τ 2 %% − τ 3 %%
−1 2 −1 3 −1

(t + 2)3
= ··· =
6

after factorizing third order polynomial (hint: use Pascal triangle for binomial power develop-
ment), and while keeping in mind that t is considered a constant while integrating relative to τ
variable.
Case 3 : t + 1 ≥ 0, t + 1 ≤ 1
The leading edge of x(t − τ ) is found in τ ∈ [0, 1], that is to say

t +1>0 ∴ t > −1 and
∴ −1 < t ≤ 0
t +1≤1 ∴ t < 0

where, non-zero surface is found in τ ∈ [−1, t + 1] interval, Fig. 17.47. However, actually there
are three distinct sub-intervals (a), (b), (c),

(a) [−1, t],

(b) [t, 0],
(c) [0, t + 1]

where in each sub-interval the convolution product is created by distinct pairs of linear segments,
see Fig. 17.47.

Fig. 17.47
Example 17.2-5 (case 3) 1 x(t − t ) x(t )

t
0
t −1 a t b c t +1
−1 0 1

Therefore, we split the convolution integral into the sum of three integrals so that the overlapping
areas are correctly calculated as
; ∞ ; t
z(t) = x(τ ) x(t − τ ) dτ = (1 + τ ) (1 + τ − t) dτ
−∞ −1
; 0
+ (1 + τ ) (1 − τ + t) dτ
t
; t+1
+ (1 − τ ) (1 − τ + t) dτ
0
1
= ··· = (−t 3 + 3t + 2)
6
382 17 Convolution Integral

1
−t (t 2 + 6t + 6)
6
1
+ (−t 3 + 3t + 2)
6
1 1
= (−t 3 + 3t + 2) − t (t 2 + 6t + 6), (−1 < t ≤ 0)
3 6

Case 4 : t − 1 ≥ −1, t − 1 ≤ 0
The trailing edge of x(t − τ ) is found in τ ∈ [−1, 0], that is to say

t − 1 ≥ −1 ∴ t > 0 and
∴ 0<t ≤1
t −1<0 ∴ t <1

where, non-zero surface is found in τ ∈ [t − 1, 1] interval, Fig. 17.48. Again, there are three
distinct sub-intervals (a), (b), (c),

(a) [t − 1, 0],
(b) [0, t],
(c) [t, 1]

say Fig. 17.48, where in each sub-interval the convolution product is created by distinct pairs of
linear segments.

Fig. 17.48
Example 17.2-5 (case 4) 1 x(t ) x(t − t )

0
t −1 a b t c t +1
−1 0 1

Therefore, we split the convolution integral into the sum of three integrals so that the overlapping
areas are correctly calculated as
; ∞ ; 0
z(t) = x(τ ) x(t − τ ) dτ = (1 + τ ) (1 + τ − t) dτ
−∞ t−1
; t
+ (1 − τ ) (1 + τ − t) dτ
0
; 1
+ (1 − τ ) (1 − τ + t) dτ
t

= ···
1 3 1
= (t − 3t + 2) + t (t 2 − 6t + 6), (0 < t ≤ 1)
3 6
Solutions 383

Case 5 : t − 1 ≥ 0, t − 1 ≤ 1
The trailing edge of x(t − τ ) is found in τ ∈ [0, 1], that is to say

t −1≥0 ∴ t > 1 and
∴ 1≤t ≤2
t −1<1 ∴ t <2

where non-zero surface is found in one distinct interval, see Fig. 17.49:

(a) [t − 1, 1]

Fig. 17.49
1 x(t ) x(t − t )
Example 17.2-5 (case 5) a

0
t −1 t t +1
−1 0 1

Therefore, we calculate the convolution integral as

; 1
z(t) = (1 − τ ) (1 + τ − t) dτ
t−1

= ···
1
= − (t − 2)3
6

Case 6 : t − 1 > 1
The trailing edge of x(t − τ ) is outside of x(τ ) non-zero region, i.e.

t −1>1 ∴ t >2

Fig. 17.50
Example 17.2-5 (case 6) 1 x(t ) x(t − t )

t
0
t −1 t t +1
−1 0 1

It is irrelevant where exactly t is relative to the origin, see Fig. 17.50. What is important is that
under this condition, because always either one of the two function equals zero, the product of
384 17 Convolution Integral

x(τ ) x(t − τ ) = 0, ∀τ
∴
; ∞ ; ∞
z(t) = x(τ ) x(t − τ ) dτ = 0 dτ
−∞ −∞

= 0 (t ≥ 2)

Summary:
The total convolution integral is therefore the sum of these six cases, see also Fig. 17.51.
⎧
⎪
⎪ 0 t < −2
⎪
⎪
⎪
⎪ 1/6(t + 2)3 −2 ≤ t ≤ −1
⎪
⎪
⎪
⎨1/3(−t 3 + 3t + 2) − 1 t (t 2 + 6t + 6) −1 ≤ t ≤ 0
z(t) = 6
⎪
⎪ 1/3(t 3 − 3t + 2) + 1 t (t 2 − 6t + 6) 0≤t ≤1
⎪
⎪ 6
⎪
⎪ −1/6(t − 2)3 1≤t ≤2
⎪
⎪
⎪
⎩0 t >2

Fig. 17.51 Example 17.2-5

z(t)

0.5

t
0
(1) (2) (3) (4) (5) (6)
−2 0 2

Exercise 17.3, page 360

1. We derive x(τ ) and y(t − τ ) functions by using graphical method, see Fig. 17.52. After the first
transformation, Fig. 17.52a, b, the change of variable t into provisional integration variable τ does
not affect the initial form of x(t). In the following transformation, y(τ ) is simply mirrored around
the vertical axis then shifted horizontally by t, Fig. 17.52c, d.
Solutions 385

Fig. 17.52 Example 17.2-1

y(t) y (− t )
1 1

t t
0 0
a) c)
0 0
y(t ) y (t − t )
1 1

t t
0 0
b) d)
0 t

We note that, given

y(t) = t 2 u(t)
∴
y(t − τ ) = (t − τ )2 u(t − τ )

There are two distinct cases to solve.

Case 1 : t ≤ 0:
The last non-zero point t of y(t − τ ) is still in the negative side, i.e.

t ≤0

Fig. 17.53
Example 17.2-1 (case 1) u(t − t ) u(t )
1

t
0

t 0

It is irrelevant where exactly t is relative to the origin, see Fig. 17.53. What is important is that
under this condition, because always either one of the two function equals zero, the product of

u(τ ) u(t − τ ) = 0, ∀τ
∴
386 17 Convolution Integral

; ∞
z(t) = u(τ ) u(t − τ ) dτ
−∞
; ∞
= 0 dτ = 0
−∞

Case 2 : t > 0:
The last non-zero point t of y(t − τ ) is shifted into the positive side, Fig. 17.54 i.e.

t >1

Fig. 17.54
Example 17.2-1 (case 2) y(t − t ) x(t )

t
0
0 t

Therefore,
; ∞ ; ∞
z(t) = u(τ ) u(t − τ ) dτ = τ 3 u(τ ) (t − τ )2 u(t − τ ) dτ
−∞ −∞
; t ; t
= τ 3 (t − τ )2 dτ = (τ 5 − 2tτ 4 + t 2 τ 3 ) dτ
0 0
% %t %t
1 6 %%t 2t % 2
% + t τ4
% 6 6
% = t − 2t + t
6
= τ % − τ5 % %
6 0 5 0 4 0 6 5 4
t6
=
60

Summary:
The total integral is therefore the sum of these two cases, see Fig. 17.55,
⎧
⎨0 t ≤0 t6
z(t) = t 6 = u(t)
⎩ t >0 60
60

Fig. 17.55
Example 17.2-1 (summary) z(t)

t
0
0
Solutions 387

2. Given,

 
1 − t; 0≤t ≤1 e−t ; t ≥0
x(t) = and, v(t) =
0; otherwise 0; otherwise

the two functions x(t) and v(t) are also defined as, see Fig. 17.56.

x(t) = (1 − t) Π (t) and v(t) = e−t u(t)

Fig. 17.56 Example 17.3-2

v (t ) v (− t )
1 1

t t
0 0
a) c)
0 0
v (t ) v (t − t )
1 1

t t
0 0
b) d)
0 t

Numerically and geometrically, solution of a convolution integral equals to the area the
overlapping surface between x(τ ) and v(t − τ ) functions. There are three distinct cases to be
solved.
Case 1 : t ≤ 0
Within this interval, either one of the two function equals zero, see Fig. 17.57, thus the product is

x(τ ) v(t − τ ) = 0, ∀τ
∴
; ∞ ; t−1
w(t) = x(τ ) v(t − τ ) dτ = (0) dτ
−∞ 0

= 0 ( t ≤ 0)
388 17 Convolution Integral

Fig. 17.57
Example 17.3-2 (case 1) v(t − t ) x(t )
1

t
0

t 0 1

Case 2 : 0 ≤ t ≤ 1 :
The leading edge of v(t − τ ) is found in interval τ ∈ [0, 1], that is to say

t > 0 ≥ 0 and t ≤ 1

where non-zero overlapping surface is found in one distinct interval τ ∈ [0, t], see Fig. 17.58.

Fig. 17.58
Example 17.3-2 (case 2) v(t − t ) x(t )
1

t
0

0 t 1

By definition we calculate,
; ∞ ; ;
t
−(t−τ )
t  
w(t) = x(τ ) v(t − τ ) dτ = (1 − τ ) e dτ = e−t eτ − τ e−t eτ dτ
−∞ 0 0
; t ; t
= e−t eτ dτ − τ eτ dτ = e−t [I1 − I2 ]
0 0

Where,
; %t
t %
I1 = e dτ = e %% = et − 1
τ τ
0 0

and,
; t  
I2 = τ eτ dτ = partial integration = t et − et + 1
0

Therefore,
# $
w(t) = e−t et − 1 − t et + et − 1
= 2 − t − 2 e−t (0 ≤ t ≤ 1)
Solutions 389

Case 3 : t ≥ 1
The leading edge of v(t − τ ) is found at t > 1], that is to say that non-zero overlapping surface is
found in one distinct interval τ ∈ [0, 1], see Fig. 17.59.

Fig. 17.59
Example 17.3-2 (case 3) u(t ) v(t − t )
1

t
0

0 1 t

Therefore, by definition we find

; ∞ ; 1
w(t) = x(τ ) v(t − τ ) dτ = (1 − τ ) e−(t−τ ) dτ
−∞ 0
; 1  
= e−t (eτ − τ eτ ) dτ = etc. as in Case 2.
0
−t
= e (e − 2) (t ≥ 1)

Summary:
The total convolution integral is therefore, see Fig. 17.60,
⎧
⎪
⎨0; t ≤0
w(t) = 2 − t − 2 e−t ; 0 ≤ t ≤ 1
⎪
⎩ −t
e (e − 2); t ≥1

Fig. 17.60 Example 17.3-2

w(t)

0 t
0
390 17 Convolution Integral

Exercise 17.4, page 360

1. (a) Given a continuous non-periodic infinitely long signal, we calculate its average power as

; T /2 ; ;
1 1 T /2 1 T /2
P (t) = lim |x(t)|2 dt = lim |A|2 dt = A2 lim
def
dt
T →∞ T −T /2 T →∞ T −T /2 T →∞ T −T /2
% 
1 %%T /2 1 T −T T
= A lim2
t% = A lim
2
− = A2 lim
T →∞ T −T /2 T →∞ T 2 2 T →∞ T
= A2

However,
; ∞ ; ∞ ; ∞
E(t) = |x(t)|2 dt = |A|2 dt = A2
def
dt
−∞ −∞ −∞
%∞
%
= A t %% = A2 (∞ + ∞)
2
−∞

=∞

That is to say, in order to sustain the average non-zero power over infinitely long period of time,
evidently, it is necessary to have infinite source of energy.
(b) Given a continuous periodic infinitely long signal, we calculate its average power over one
period only as
; ;
1 T /2 1 T /2 2
P (t) = | sin(t)|2 dt =
def
sin (t) dt
T −T /2 T −T /2
 
1 − cos 2x
= sin2 (t) =
2
; T /2 
1 1 − cos 2x
= dt
T −T /2 2
T /2
1 1 1
= t − sin(2t)
2 T 4 −T /2
   
1 1 T T 1 T T
= + − sin 2
− sin −2

T 2 2 2 4 2
2


=
1 1
T −
1
(T
sin :
 0
) + sin :

(T
0
@  )
@
T 2 4
1
=
2
Solutions 391

Again, in order to sustain the non-zero average power over infinitely long period of time, there
must be an infinite source of energy

1
E(t) = lim P (t) t = lim t =∞
def

t→∞ t→∞ 2

(c) Given a converging infinitely long signal, we calculate its total energy as
; ; ; %
∞ ∞ % −t %
%e u(t)%2 dt =
∞
−2t 1 −2t %%∞
E(t) = |x(t)| dt = (1) dt = − e
def 2 2
e %
−∞ −∞ 0 2 0
1 1
= − [0 − 1] =
2 2

Therefore,

1
P (t) = lim E(t) = 0
def

T →∞ T

That is to say, having a limited source of energy that must be distributed over infinitely long
period of time, the average power reduces to zero.

(d) Given a complex signal, we use Euler’s formula to write its period forms as

2π
x(t) = A ej ω 0 t = A [cos(ω 0 t) − j sin(ω 0 t)] ∴ ω 0 T0 = 2π ∴ T0 =
ω0

so that average power is calculated as

; T0 /2 ;
1 1 T0 /2 2 j ω 0 :1

P (t) = |A e | dt = |e  t |2 dt
def j ω0 t 2
A 
T0 −T0 /2 T0 −T0 /2
% 
A2 %% 0
T /2
A2 T0 T0
= t% = + = A2
T0 −T0 /2 T 0 2 2

This is infinitely long signal, thus in order to sustain non-zero average power over infinitely long
time it must be that E(t) = ∞.
392 17 Convolution Integral

(e) Given an infinitely long signal, its energy is

; ; ; %∞
∞ ∞ ∞ %
E(t) = |x(t)|2 dt = |u(t)|2 dt = (1)2 dt = t %% = ∞
def

−∞ −∞ 0 0

Therefore, its average power is

; T /2 ;
1 1 T /2
P (t) = lim |x(t)| dt = lim |u(t)|2 dt
def 2
T →∞ T −T /2 T →∞ T −T /2
; %
1 T /2 2 1 %%T /2 1 T
= lim (1) dt = lim t % = lim
T →∞ T 0 T →∞ T 0 T →∞ T 2

1
=
2

which makes sense because half of the time u(t) signal amplitude equals zero (for t ≤ 0) and the
other half of time (i..e. t ≥ 0) it equals to one.
(f) Amplitude of delayed port signal x(t) = 2Π(t − 1) equals two for t ∈ [0, 1], otherwise it
equals zero. Thus we write,
; ∞ ; ∞ ; 1
E(t) = |x(t)|2 dt = |2Π(t − 1)|2 dt = (2)2 dt = 4 < ∞
def

−∞ −∞ 0

which is to say that its average power over the infinitely long time interval must be P (t) = 0.
 Discrete Convolution Sum
18

Important to Know

Convolution sum: It is defined as

∞

y[n] = x[k] ∗ y[n] = x[k] y[n − k] (18.1)
k=−∞

where, k is the provisional summation variable. <t

Energy and power definitions: By definition, E(t) = 0 P (τ )dτ for continuous and discrete
functions. Recall that |x(t)|2 = x(t) x ∗ (t).

1. Energy within an interval

; t2 
m
Ex = |x(t)|2 dt, En = |x[k]|2
t1 k=n

2. Total energy
; ∞ ∞

Ex = |x(t)| dt,
2
En = |x[k]|2
−∞ k=−∞

3. Average power within an interval

;
1 
t2 n
1
Px  = |x(t)|2 dt, Pn  = |x[k]|2
t2 − t1 t1 2n + 1 k=n

4. Total power
;
1 
τ/2 n
1
Px = lim |x(t)| dt,
2
Pn = lim |x[k]|2
τ →∞ τ −τ/2 n→∞ 2n + 1
k=n

© Springer Nature Switzerland AG 2021 393

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_18
394 18 Discrete Convolution Sum

18.1 Exercises

18.1 ** Convolution, Basic Properties

1. Given an arbitrary function h[k] and δ[k] Dirac function, calculate the convolution sum y[k] =
h[k] ∗ δ[k − m].
2. Derive proof to show that convolution sum is commutative.
3. Show the distributive property of convolution sum.
4. Show the associative property of convolution sum.
5. Derive proof to show the time-shifting property of convolution sum, i.e.

h1 [k − m] ∗ h2 [k − n] = f [k − m − n]

18.2 *** Convolution, Discrete Functions

1. Given two short sequences f [k] and g[k], each with only two samples as in Fig. 18.1, calculate

h[k] = f [k] ∗ g[k]

Fig. 18.1 Example 18.2-1 2 f [k] g[k]

1
k
0
−3 −2 −1 0 1 2 3 4 5

2. Given two long sequences x[k] and h[k],

x[k] = α k u[k]; (0 < α < 1)

h[k] = u[k]

calculate y[k] = x[k] ∗ h[k]

Solutions 395

3. Given two sequences h[n] = [1, 2, 0, −3], and x[n] is given as

(a) x[n] = δ[n] (b) x[n] = δ[n + 1] + δ[n − 2]

(c) x[n] = [1, 1, 1] (d) x[n] = [2, 1, − 1, −2, −3]
Note: number whose index n = 0 is underlined.
4. Given two sequences x[k] and h[k], as
 
1; (0 ≤ k ≤ 4) α k ; (0 ≤ k ≤ 6)
x[k] = and h[k] =
0; otherwise 0; otherwise

calculate y[k] = x[k] ∗ h[k]

18.3 *** Energy and Power of Discrete Functions

1. Given sequences x[k], calculate their respective energies and powers.

 k
1
(a) x[k] = δ[k] (b) x[k] = u[k]
4

αk ; k ≥ 0
(c) x[k] = e j 10k
u[k] (d) x[k] =
0; k<0
π 
(e) x[k] = u[k] (f) x[k] = cos k
6

Solutions

Exercise 18.1, page 394

1. The product of an arbitrary function h[k] and Dirac pulse at time m effectively “samples” only one
point of the function while products at all other points equal zero. That is to say, using temporary
summing variable n, we write
∞

y[k] = h[k] ∗ δ[k − m] = h[n] δ[(k − m) − n]
n=−∞
%
%
= · · · + 0 + 0 + · · · + h[n] δ[(k − m) − n] %% + 0 + 0 + ···
n=k−m

= h[k − m] δ[0]
= h[k − m]

2. The sum operation is commutative, that is to say

y[k] = h1 [k] ∗ h2 [k] = h2 [k] ∗ h1 [k]

396 18 Discrete Convolution Sum

and we can formally prove it as follows:

∞

y[k] = h1 [k] ∗ h2 [k] = h1 [m] h2 [k − m]
def

m=−∞

change of variables:
⎧ ⎫
⎪
⎪ n =k−m ∴ m=k−n ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ ⎪
⎬
∴
⎪
⎪ m = −∞ ∴ n = k − (−∞) = +∞ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎩ ⎪
⎭
m = ∞ ∴ n = k − (+∞) = −∞
−∞

= h1 [k − n] h2 [n]
n=∞
 
the order of addition is not relevant, thus
∞

= h1 [k − n] h2 [n]
n=−∞
∞

= h2 [n] h1 [k − n]
n=−∞

= h2 [k] ∗ h1 [k]
def

3. The sum operation is distributive, and we can formally prove it as follows:

 
y[k] = x[k] ∗ h1 [k] + h2 [k]
∞
  
= x[m] ∗ h1 [k − m] + h2 [k − m]
m=−∞
∞ 
 
= x[m] ∗ h1 [k − m] + x[m] ∗ h2 [k − m]
m=−∞
∞
 ∞

= x[m] ∗ h1 [k − m] + x[m] ∗ h2 [k − m]
m=−∞ m=−∞

= x[k] ∗ h1 [k] + x[k] ∗ h2 [k]

def

4. The sum operation is associative, that is to say

   
x[k] ∗ h1 [k] ∗ h2 [k] = x[k] ∗ h1 [k] ∗ h2 [k]

and we can formally prove it as follows:

Solutions 397

 
  ∞

y[k] = x[k] ∗ h1 [k] ∗ h2 [k] = x[m]h1 [k − m] ∗ h2 [k]
m=−∞
∞
 ∞

 
= x[m]h1 [n − m] h2 [k − n]
n=−∞ m=−∞
∞
 ∞

= x[m] h1 [n − m] h2 [k − n]
n=−∞ m=−∞
∞
 ∞

= x[m] h1 [n − m] h2 [k − n]
m=−∞ n=−∞

Change of variable
⎧ ⎫
⎪
⎪ r =k−n ∴ n=k−r ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ ⎪
⎬
∴
⎪
⎪ n = −∞ ∴ r = k − (−∞) = +∞ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎩ ⎪
⎭
n = ∞ ∴ r = k − (+∞) = −∞

therefore,

∞
 −∞

y[k] = x[m] h1 [k − r − m] h2 [r]
m=−∞ r=∞
∞
 ∞

= x[m] h2 [r] h1 [(k − m) − r]
m=−∞ r=−∞
∞
  
= x[m] h2 [m] ∗ h1 [k − m]
m=−∞
∞
  
= x[m] h [k − m] ∗ h2 [m]
1  
m=−∞ z[k−m]
∞

= x[m] z[k − m]
m=−∞

= x[k] ∗ z[k]
def

 
= x[k] ∗ h1 [k] ∗ h2 [k]

5. The time-shifting property of convolution sum, i.e.

h [k − m] ∗ h2 [k − n] = f [k − (m + n)]
 1        
delayed by m delayed by n delayed by m+n
398 18 Discrete Convolution Sum

may be formally proven as follows:

∞

f [k] = h1 [k] ∗ h2 [k] = h1 [r] h2 [k − r]
r=−∞

However, delayed versions of h functions are

∞

h1 [k − m] ∗ h2 [k − n] = h1 [r − m] h2 [(k − n) − r]
r=−∞
⎧ ⎫
⎪
⎪t =r −m ∴ r =t +m ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎨ ⎪
⎬
∴
=
⎪
⎪r = −∞ ∴ t = (−∞) − m = −∞ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎩ ⎪
⎭
r = ∞ ∴ t = (+∞) − m = +∞
∞

= h1 [t] h2 [(k − n) − (t − m)]
t=−∞
∞

= h1 [t] h2 [(k − m − n) − t]
t=−∞
∞

= h1 [r] h2 [(k − m − n) − r]
r=−∞

= f [k − m − n]
def

Exercise 18.2, page 394

1. Given two short sequences f [k] and g[k], the interval boundaries of convolution h[k] series are
determined as follows. Intervals of the two sequences are

f [−2] = 2, f [−1] = 2 ∴ start point: k = −2 and, end point: k = −1

g[1] = 2, f [3] = 1 ∴ start point: k = 2 and, end point: k = 3
∴
h[k] ∴ start point: k = −2 + 2 = 0
∴ end point: k = −1 + 3 = 2

that is to say, h[k] = 0, 0 > k > 2; in other words it is necessary to evaluate h[k] for k =
[0, 1, 2]. We write,

∞
 −1

h[k] = f [k] ∗ g[k] = f [m] g[k − m] = f [m] g[k − m]
def

m=−∞ m=−2

where the last sum is limited only to the interval m = [−2, −1] where f [m] = 0.
Solutions 399

−1

k = 0 : h[0] = f [m] g[0 − m] = f [−2] g[0 − (−2)] + f [−1] g[0 − (−1)]
m=−2

= f [−2] g[2] + f [−1] g[1] = (2)(2) + (2)(0) = 4

−1

k = 1 : h[1] = f [m] g[1 − m] = f [−2] g[1 − (−2)] + f [−1] g[1 − (−1)]
m=−2

= f [−2] g[3] + f [−1] g[2] = (2)(1) + (2)(2) = 6

−1

k = 2 : h[2] = f [m] g[2 − m] = f [−2] g[2 − (−2)] + f [−1] g[2 − (−1)]
m=−2

= f [−2] g[4] + f [−1] g[3] = (2)(0) + (2)(1) = 2

In conclusion, h[k] = [4, 6, 2], see Fig. 18.2. Note that value of h[0] is underlined.
Fig. 18.2 Example 18.2-1 6 h[k]
4

2
0
−2 −1 0 1 2 3 4

2. Given two long sequences x[k] and h[k], we use graphical technique to determine x[m] and
h[k − m], see Fig. 18.3.

x[k] = α k u[k]; (0 < α < 1)

h[k] = u[k]

and similarly to Example 17.2-2 we decompose the problem into calculation of convolution
integral within multiple intervals.
Fig. 18.3 Example 18.2-2 x[m]
1

0 m

−3 −2 −1 0 1 2 3 4 5 6

1 h[k − m]

···
0 m
··· k−1 k k+1
400 18 Discrete Convolution Sum

Case 1 : k < 0
There is no overlapping interval, consequently all products inside the convolution integrals equal
zero, Fig. 18.4.

x[m] h[k − m] = 0; ∀m
∴
y[k] = x[k] ∗ h[k] = 0; (k < 0)

Fig. 18.4 Example 18.2-2 h[k − m]

1
(case 1)
x[m]
···
0 m
··· k −2 k −1 k ··· 0 1 2 ···

Case 2 : k > 0
There is overlapping interval, consequently there are products inside the convolution integrals
non-equal zero, Fig. 18.5.

x[m] h[k − m] = 0; m = [0, 1, . . . , k]

Thus,


k
y[k] = x[k] ∗ h[k] = x[m] h[k − m]
def


k
= α m (1)
0
  
geometric sum
N −1  
 N; α = 1
= α = 1−αN
n

n=0 1−α
; α=1

1 − α k+1
= (k ≥ 0)
1−α

Fig. 18.5 Example 18.2-2 h[k − m]

1
(case 2) x[m]
···
0 m
··· k −1 k ···
· · · −1 0 1 · · ·
Solutions 401

Fig. 18.6 Example 18.2-2 1 y[k]

1−a
···
1
··· k
0
−2 0 2 4 6

In summary, after adding the two cases together we write the solution in the compact form

1 − α k+1
y[k] = u[k]
1−α

We note that (see Fig. 18.6), because α < 1 then

1 − α k+1 1
lim α k+1 → 0 ∴ lim →
k→∞ k→∞ 1 − α 1−α

3. Given two sequences h[n] = [1, 2, 0, −3], and x[n] is given as

(a) y[n] = δ[n] ∗ h[n] = h[n] ∗ δ[n] = h[n]

(b) Given,

x[n] = δ[n + 1] + δ[n − 2]

First we calculate

y[n] = x[n] ∗ h[n] = (δ[n + 1] + δ[n − 2]) ∗ h[n]

= δ[n + 1] ∗ h[n] + δ[n − 2] ∗ h[n]
= h[n + 1] + h[n − 2]

Therefore, two shifted sequences h[n + 1] and h[n − 2] are

h[n] = [1, 2, 0, −3]

h[n + 1] = [1, 2, 0, −3] shifted one places to right

h[n − 2] = [ , 1, 2, 0, −3] shifted two places to left

We tabulate the convolution sums as

n −2 −1 0 1 2 3 4
h[n + 1] 1 2 0 −3
h[n − 2] + 1 2 0 −3
y[n] 1 2 0 −2 2 0 −3
402 18 Discrete Convolution Sum

That is to say, (see also Fig. 18.7)

y[n] = [1, 2, 0, −2, 2, 0, −3]

Fig. 18.7 Example 18.2-3
y[n]
3

n
0

−3

−4−3−2−1 0 1 2 3 4 5 6

(c) Given

x[n] = [1, 1, 1] = δ[n] + δ[n − 1] + δ[n − 2]

we calculate,

y[n] = x[n] ∗ h[n] = (δ[n] + δ[n − 1] + δ[n − 2]) ∗ h[n]

= h[n] + h[n − 1] + h[n − 2]

We multiply as

n −1 0 1 2
h[n] 1 2 0 −3
x[n] × 1 1 1
y[n] 1 2 0 −3

Therefore, knowing y[n] we write

h[n − 1] = [1, 2, 0, −3]

h[n − 2] = [ , 1, 2, 0, −3]
Solutions 403

and the total sum h[n] + h[n − 1] + h[n − 2] is

n −1 0 1 2 3 4
h[n] 1 2 0 −3
h[n − 1] 1 2 0 −3
h[n − 2] + 1 2 0 −3
y[n] 1 3 3 −1 −3 −3

That is to say, (see also Fig. 18.8)

y[n] = [1, 3, 3, −1, −3, −3]

Fig. 18.8 Example 18.2-3

y[n]
3

n
0

−3

−4−3−2−1 0 1 2 3 4 5 6

(d) Given

x[n] = [2, 1, − 1, −2, −3]

h[n] = [1, 2, 0, −3]

using the tabular method we calculate,

y[n] = x[n] ∗ h[n]

for n = [−2, −1, 0, 1, 2]. We calculate the interval of the convolution sum

x[−2] = 2, . . . , f [2] = −3 ∴ start point: n = −2 and, end point: n = 2

h[−1] = 1, . . . , f [2] = −3 ∴ start point: n = −1 and, end point: n = 2
∴
y[n] ∴ start point: n = (−2) + (−1) = −3
∴ end point: k = 2 + 2 = 4

that is to say, it is necessary to evaluate h[n] for n = [−3, −2, −1, 0, 1, 2, 3, 4].
404 18 Discrete Convolution Sum

We “sweep” index of x[n] and each time perform the multiplication x[n] × h[n] for each term in
h[n] and write the result in the corresponding column. For example, in n = 2 column we write

x[0] h[2] = (−1) × (−3) = 3

n=0

n −2 −1 0 1 2
h[n] 1 2 0 −3
x[0] × −1
2 1 −2 −3
−1 −2 0 3

And the rest of products are as follows:

n = −2

n −3 −2 −1 0 1 2
h[n + 2] 1 2 0 −3
x[−2] × 2
1 −1 −2 −3
2 4 0 −6

n = −1

n −2 −1 0 1 2
h[n + 1] 1 2 0 −3
x[−1] × 1
2 −1 −2 −3
1 2 0 −3

n=1

n −2 −1 0 1 2 3
h[n − 1] 1 2 0 −3
x[1] × −2
2 1 −1 −3
−2 −4 0 6
Solutions 405

n=2

n −2 −1 0 1 2 3 4
h[n − 2] 1 2 0 −3
x[2] × −3
2 1 −1 −2
−3 −6 0 9

All intermediate results are tabulated, we execute the convolution sum


2
y[n] = x[k] h[n − k]
k=−2

in tabular form as,

n −3 −2 −1 0 1 2 3 4
x[−2] h[n + 2] 2 4 0 −6
x[−1] h[n + 1] 1 2 0 −3
x[0] h[n] −1 −2 0 3
x[1] h[n − 1] −2 −4 0 6
x[2] h[n − 2] + −3 −6 0 9
y[n] 2 5 1 −10 −10 −3 6 9

Visually, tabular method illustrates the same “sweep” of h[k−n] as we already saw in the graphics
method. In conclusion,


2
y[n] = x[n] ∗ h[n] = x[k] h[n − k] = [2, 5, 1, −10, −10, −3, 6, 9]
def

k=−2

as well as in Fig. 18.9.

Fig. 18.9 Example 18.2-3 10 y[n]

5
n
0

−5

−10
−4−3−2−1 0 1 2 3 4 5 6
406 18 Discrete Convolution Sum

4. Given,
 
1; (0 ≤ k ≤ 4) α k ; (0 ≤ k ≤ 6)
x[k] = and h[k] =
0; otherwise 0; otherwise

∞

we calculate convolution sum as y[k] = x[k] ∗ h[k] = x[m] h[k − m]. Using graphical
def

m=−∞
method, we find x[m] and h[k − m] as in Fig. 18.10

Fig. 18.10 Example 18.2-4 x[m]

1
··· ···
m
0

−2 0 2 4 6
h[k − m]

1
··· m
0
k−6 k

Case 1 : k < 0
There is no overlapping interval, Fig. 18.11, consequently all products inside the convolution
integrals equal zero.

x[m] h[k − m] = 0; ∀m
∴
y[k] = x[k] ∗ h[k] = 0; (k < 0)

Fig. 18.11 Example 18.2-4 h[k − m] x[m]

1
··· m
0
k−6 k 0 4 ···
Solutions 407

Case 2 : 0 ≤ k ≤ 4
There is overlapping interval [0, k], Fig. 18.12, thus the convolution integrals are calculated as
follows:


k
y[k] = x[k] ∗ h[k] = x[m] h[k − m]
m=0


k 
k
= (1) α k−m = α k−m
m=0 m=0
 
r =k−m m=0 ∴ r=k 
0 
k
1 − α k+1
= = αr = αr =
m=k ∴ r=0 r=k r=0
1−α

Fig. 18.12 Example 18.2-4 h[k − m] x[m]

1
··· m
0
k−6 0 k 4 ···

Case 3 : 5 ≤ k ≤ 6
There is overlapping interval [0, 4], Fig. 18.13, thus the convolution integrals are calculated as
follows:


4
y[k] = x[k] ∗ h[k] = x[m] h[k − m]
m=0


4 
4
= (1) α k−m
= α k α −m
m=0 m=0
 5

4
  − α −1 α k − α k−5 α α k+1 − α k−4
−1 m k1
=α k
α =α −1
= −1
=
m=0
1−α 1−α α α−1

Fig. 18.13 Example 18.2-4 h[k − m] x[m]

1
m
0
k−6 0 4 k ···
408 18 Discrete Convolution Sum

Case 4 : 7 ≤ k ≤ 10
There is overlapping interval [k − 6, 4], Fig. 18.14, thus the convolution integrals are calculated
as follows:


4
y[k] = x[k] ∗ h[k] = x[m] h[k − m]
m=k−6


4
= (1) α k−m
m=k−6
⎧ ⎫
⎪
⎪ r = m − k + 6; k−m=6−r ⎪
⎪
⎨ ⎬
= m=k−6 ∴ r =0
⎪
⎪ ⎪
⎪
⎩ ⎭
m = 4 ∴ r = 10 − k

10−k 
10−k
 r 1 − α k−11 α 7 − α k−4
= α 6−r = α 6 α −1 = α6 =
r=0 r=0
1 − α −1 α−1

Fig. 18.14 Example 18.2-4 h[k − m] x[m]

1
m
0
0 k−6 4 k ···

Case 5 : k > 10
There is no overlapping interval, Fig. 18.15, consequently all products inside the convolution
integrals equal zero.

x[m] h[k − m] = 0; ∀m
∴
y[k] = x[k] ∗ h[k] = 0; (k > 10)

Fig. 18.15 Example 18.2-4 x[m] h[k − m]

1
m
0
0 4 k−6 k ···
Solutions 409

In summary,
⎧
⎪
⎪ 0; k<0
⎪
⎪
⎪
⎪ 1−α k+1
⎪
⎪ ; 0≤k≤4
⎪
⎪ 1−α
⎨ k+1
α − α k−4
y[k] = ; 5≤k≤6
⎪
⎪ α−1
⎪
⎪
⎪
⎪ α −α
7 k−4
⎪
⎪ ; 7 ≤ k ≤ 10
⎪
⎩ α−1
⎪
0; k > 10

Exercise 18.3, page 395

1. The energy and power sums of discrete functions are done as follows:
(a) Given Dirac function, by definition we write

∞
 ∞

E[k] = |x[k]|2 = |δ[k]|2 = · · · + 0 + 0 + · · · + 12 + 0 + 0 + · · ·
def

k=−∞ k=−∞

=1<∞

which is to say that since the energy is distributed over infinitely long time interval k = ∞;
therefore, at any given moment, the available average power must be

E[k] 1
P [k] = lim = lim = 0
k→∞ k k→∞ k

as same as for continuous functions where t → ∞.

(b) Given infinitely long series, we write the expression for its energy as
%
∞ %
%2 ∞  ∞ 
∞
  k %  1 2k  1 k
% 1 %
E[k] = |x[k]| = u[k]% = =
def 2
%
% 4 % 4 16
k=−∞ k=−∞ k=0 k=0
⎧ ∞ ⎫
⎪  1 ⎪
⎪geometric series,
⎪ k
= |r| ⎪
⎪
⎪
⎪
r
−
where, < 1 ⎪
⎪
⎪
⎪ 1 r ⎪
⎪
⎪
⎪
k=0 ⎪
⎪
⎪
⎪ ⎪
⎪
⎨ ∞
a ⎬
as well as, ar =2k
⎪
⎪ 1 − r2 ⎪
⎪
⎪
⎪ k=0 ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ ∞ ⎪
⎪
⎪
⎪ a r ⎪
⎪
⎪
⎩ or, ar 2k+1
= ⎪
⎭
k=0
1−r 2
410 18 Discrete Convolution Sum

 
1
∴ a = 1 and r =
16
1 16
= = <∞
1 15
1−
16

which is to say that since the energy is distributed over infinitely long time interval k = ∞;
therefore,

E[k] 16/15
P [k] = lim = lim =0
k→∞ k k→∞ k

as same as for continuous functions where t → ∞.

(c) Complex exponential series are added for k ∈ [−n, n]; consequently, there are in total 2n+1
points to add. That is k = 1, 2, 3, . . . , n (therefore n points in total) on the positive side, plus k =
−1, −2, −3, . . . , −n (i.e. n points in total) on the negative side, plus k = 0 (i.e. one additional
point) in the middle, thus 2n + 1 points in total. Accordingly, by definition for power we find
! " ! "
1 n
1 n
% j 10k %2
P [k] = lim
def
|x[k]| = lim
2 %e u[k]%
n→∞ 2n + 1 n→∞ 2n + 1
k=−n k=−n
! " ! "
1  %% j 10k %%2 
n n
1
= lim e = lim |1k |2
n→∞ 2n + 1 n→∞ 2n + 1 k=0
k=0
  
n+1

n+1 ∞ 1 1
= lim = = lim = < ∞
L.H.

n→∞ 2n + 1 ∞ n→∞ 2 2

which is to say that in order to support non-zero average power over infinite long interval k = ∞,
there must be E[k] = P [k] k = ∞.
(d) Infinite sum of power series where |α| < 1 converges, thus by definition we write

∞
 ∞
 ∞
 ∞

% k % % k %2  k
E[k] =
def
|x[k]|2 = %α u[k]%2 = %α % = α2
k=−∞ k=−∞ k=0 k=0
  
geometric series
1
= <∞
1 − α2
Solutions 411

Finite amount of energy must be distributed over infinitely long interval, thus the average power
equals

1
= lim 1 − α = 0
E[k] 2
P [k] = lim
k→∞ k k→∞ k

(e) Infinitely long step function has average power calculated as

! " ! "
1 n
1 n
P [k] = lim |x[k]| = lim |u[k]|
def 2 2
n→∞ 2n + 1 n→∞ 2n + 1
k=−n k=−n
! "
1  n
n+1 ∞ 1
= lim |1k |2 = lim = = lim
L.H.

n→∞ 2n + 1 k=0 n→∞ 2n + 1 ∞ n→∞ 2

  
n+1

1
= <∞
2

which is to say that in order to support non-zero average power over infinite long interval k = ∞,
there must be E[k] = P [k] k = ∞.

(f) Average power of an infinitely long periodic series is calculated over one period as
! " ! n %
"
n   π %2
1 1 % %
P [k] = lim |x[k]|2 = lim
def
%cos k %
n→∞ 2n + 1 n→∞ 2n + 1 6
k=−n k=−n
  
1 + cos(2x)
= cos (x) =
2
2
! "
1 n
1 1 π
= lim + cos 2
k
n→∞ 2n + 1 2 2 6
3
k=−n
⎡ ⎤
⎢   π  ⎥
⎢ 1 1
n
1 1
n
⎥
= lim ⎢ 1k + cos k ⎥
n→∞ ⎢ 2n + 1 2 2n + 1 2 3 ⎥
⎣ k=−n k=−n ⎦
     
2n+1 =0
 
in total, there are (2n+1) terms of 1/2 each;
=
the sum of sin or cos over one period equals zero ;
1 1 1
= lim  ( +
2n 1) = < ∞
n→∞ 
2n+ 1 2 2

which is to say that in order to support non-zero average power over infinite long interval k = ∞,
there must be E[k] = P [k] k = ∞.
 Fourier Transformation Integral
19

Important to Know

Classification of Fourier transformations, sums, and series:

Continuous time (CT) periodic, Fig. 19.1(left), and non-periodic, Fig. 19.1(right), x(t) functions.
Periodic functions are transformed by continuous time Fourier series (CT FS), while continuous time
non-periodic functions are transformed by continuous time Fourier transformation (CT FT).

Fig. 19.1 Continuous time x(t) x(t)

Fourier series vs. Fourier
transformation t t
T

CT FS CT FT

|C[n]| |X(w )|
n

Similarly, discrete time (DT) periodic functions are mapped by discrete time Fourier series (DT FS),
while discrete time non-periodic functions are transformed by discrete time Fourier transformation
(DT FS).
In addition, in order to carry out FT of continuous signals on numerical processors, CT signals
must be sampled before applying discrete Fourier transformation (DFT) algorithm. Version of DFT
algorithm that is optimized for fast execution is referred to as fast Fourier transformation (FFT).
Fourier and inverse Fourier integrals: are defined as
; ∞
F(x(t))=X(ω) = x(t) e−j ω t dt
def
(19.1)
−∞
; ∞
1
F−1 (X(ω))=x(t) =
def
X(ω) ej ω t dω (19.2)
2π −∞

© Springer Nature Switzerland AG 2021 413

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_19
414 19 Fourier Transformation Integral

19.1 Exercises

19.1 ** Fourier Integral

1. Derive Fourier transformations X(ω) of the following functions x(t):

(a) x(t) = δ(t) (b) x(t) = a0 ; (const.)
(c) x(t) = δ(t − τ ) (d) x(t) = ej ω 0 t

2. Derive Fourier transformations X(ω) of the following functions x(t):

(a) x(t) = cos(ω 0 t) (b) x(t) = sin(ω 0 t)

3. Derive Fourier transformations X(ω) and show its graphical representation of the following
functions x(t), where (a > 0):
(a) x(t) = e−at u(t) (b) x(t) = eat u(−t) (c) x(t) = e−a|t|

t
4. Given x(t) = a0 Π derive Fourier transformation X(ω) and show its graphical representa-
τ
tion. 
t
5. Given x(t) = Λ derive Fourier transformation X(ω) and show its graphical representation.
τ
6. Given x(t) = sign (t) derive Fourier transformation X(ω) and show its graphical representation.
7. Given x(t) = u(t) derive Fourier transformation X(ω) and show its graphical representation.
8. Given x(t) = sinc (t) derive Fourier transformation X(ω) and show its graphical representation.

19.2 ** Fourier Integral, Exercises

1. Given function x(t) in Fig. 19.2 derive X(ω).

Fig. 19.2 Example 19.2-1 x(t)

2

1
t
0
3 2 1 0 1 2 3

2. Given function x(t) in Fig. 19.3 derive X(ω).

Solutions 415

Fig. 19.3 Example 19.2-1 x(w )

2
w
0
−2

2 0 2

3. Given RF pulse, Fig. 19.4, where fc is the RF waveform frequency, and T is the duration of RF
pulse, derive its Fourier transform.

Fig. 19.4 Example 19.2-1 h(t)

A 1/ fc
t
0
−A T

 T /2 0 T /2

Solutions

Exercise 19.1, page 414

1. Continuous functions are transformed by Fourier integral as follows.

(a) Dirac function equals non-zero only in one point, and its Fourier transformation is found by
definition.
; ∞
 
X(ω)= x(t) e−j ω t dt = δ(t) = 0 ⇒ t = 0 ∴ ej ω 0 = 1
def

−∞
; ∞ ; ∞
:1

= e−jω t dt =
δ(t)  δ(t) dt =1
def

−∞ −∞

We keep in mind this bidirectional relationship between Dirac function x(t) and its Fourier
transformation of the constant function X(ω), Fig. 19.5. That is to say, the two functions are said
to be “dual” because Fourier transform of Dirac function is a constant function, as well as the
opposite, i.e. Fourier transform of a constant function is Dirac function. Physical interpretation
is that in order to synthesize Dirac function it is necessary to add all possible sin(ω t) whose
frequencies are in the range ω ∈ [−∞, ∞].

Fig. 19.5 Example 19.1-1(a) x(t) X(w )

1 1

t w
0 0

0 0
416 19 Fourier Transformation Integral

(b) Fourier transformation of a constant function may be derived, for example, by using “reverse
engineering” technique, i.e. by using the inverse Fourier transformation integral. Given that
x(t) = a0 , we write
; ∞
−1 1
F (X(ω)) = x(t)=
def
X(ω) ej ω t dω
2π −∞

∴
; ∞
1
a0 =
def
X(ω) ej ω t dω
2π −∞

The idea is to gradually deduce the form of X(ω) until the inverse Fourier equality satisfied, i.e.
the right side of the equation is reduced to a0 and becomes equal to the left side. To start with,
X(ω) must include a 2π term, so that it cancels the already existing 2π term, as
; ∞
1
a0 =
def
HH×? ej ω t dω
2π
HH
2π −∞

Then, there must be the a0 constant on the right side somewhere; thus we further deduce that
X(ω) = 2π a0 ×?, i.e.
; ∞
a0 = a0 ×? ej ω t dω
def

−∞

∴
; ∞
a0 =a0
def
? ej ω t dω
−∞
  
=1

In order for the last equality to be valid, the right side integral must be reduced to one.
One possible solution would be that the exponential term equals to one. As already seen in
Example 19.1-1(a), we can use the sampling property of Dirac function because
; ∞
:1

e−jω t dω=1
δ(ω) 
def

−∞

which leads to the following conclusion,

; ∞
a0 =a0
def
δ(ω) ej ω t dω
 −∞  
=1

In summary, we write
; ∞
1
a0 =
def
X(ω) ej ω t dω
2π −∞

∴
Solutions 417

; ∞
1
a0 = 2π a0 δ(ω) ej ω t dω
2π −∞

∴
X(ω) = 2π a0 δ(ω)

Which is to say that Fourier transformation of a constant function is indeed Dirac function (the
2π a0 term is the multiplying constant), as stated in Example 19.1-1(a) and illustrated in Fig. 19.5.

(c) Delayed Dirac function δ(t − τ ) results in

; ∞  
X(ω)= x(t) e−j ω t dt = δ(t − τ ) = 0 ⇒ t = τ
def

−∞
; ∞ ; ∞
−j ω τ −j ω τ
= δ(t − τ ) e dt = e δ(t − τ ) dt
−∞
 −∞  
def
=1
−j ω τ
= (1) e

that is, Fourier transformation of delayed Dirac function δ(t − τ ) is also the constant function,
where the complex exponent term only adds to phase delay τ that does not change the form of
the constant functions.
We note that this property of Fourier transform is general for any delayed function, i.e. Fourier of
a delayed function equals to Fourier transformation of its non-delayed version multiplied by the
phase delayed term,

F (f (t − τ )) = F (f (t)) e−j ω τ

(d) Transformation of complex exponential function x(t) = ej ω 0 t (its amplitude is constant,

i.e. equals to the radius of the unity circle) may be deduced using similar reasoning as in
Example 19.1-1(b). Note that ω 0 is a constant as opposed to ω that is a variable.
; ∞
1
F−1 (X(ω)) = x(t)=
def
X(ω) ej ω t dω
2π −∞

∴
; ∞
1
e j ω0 t
= X(ω) ej ω t dω
2π −∞

∴
; ∞
1
ej ω 0 t = 2π ×? ej ω t dω
2π −∞
418 19 Fourier Transformation Integral

We note that by sampling the exponential function on the right side at ω = ω 0 , we create
term equal to the exponent on the left side. That can be done by using delayed Dirac function
δ(ω − ω 0 ), so that we write
; ∞
1
ej ω 0 t = 2π δ(ω − ω 0 ) ej ω t dω
2π −∞

∴
; ∞
1
e j ω0 t
= 
2π δ(ω − ω 0 ) ej ω 0 t dω

2π −∞

∴
; ∞
ej ω 0 t = ej ω 0 t δ(ω − ω 0 ) dω
−∞
  
def
=1

∴
ej ω 0 t = ej ω 0 t 

In conclusion,

X(ω) = 2π δ(ω − ω 0 )

2. Continuous sinusoidal functions are transformed by Fourier integral as follows.

(a) Application of Euler formula transforms cosine function to the sum of complex exponential
functions.

ej ω 0 t + e−j ω 0 t
x(t) = cos(ω 0 t) =
2

Therefore,
 
ej ω 0 t + e−j ω 0 t ej ω 0 t e−j ω 0 t
X(ω) = F(x(t)) = F =F +F
2 2 2

In Example 19.1-1(d) we derived Fourier transformation of complex exponential function; thus

we conclude
 j ω0 t
e
F = 2π δ(ω − ω 0 )
2
∴

e−j ω 0 t
F = 2π δ(ω + ω 0 )
2
Solutions 419

In conclusion,

1# $
X(ω) = 2
π δ(ω − ω 0 ) + 2
π δ(ω + ω 0 ) = π [δ(ω + ω 0 ) + δ(ω − ω 0 )]
2

That is to say, Fourier transform of cos(ω 0 t) function is the sum of two Dirac functions, one
delayed and the other advanced by ω 0 , Fig. 19.6.

Fig. 19.6 Example 19.1-1(a) x(t) X(w )

1 p
t
0
w
0
−1
0 −w0 0 w0

(b) Application of Euler formula transforms sine function to the difference of complex
exponential functions.

ej ω 0 t − e−j ω 0 t
x(t) = sin(ω 0 t) =
2j

Therefore,
 
ej ω 0 t − e−j ω 0 t ej ω 0 t e−j ω 0 t
X(ω) = F(x(t)) = F =F −F
2j 2j 2j

In Example 19.1-1(d) we derived Fourier transformation of complex exponential function; thus

we conclude
 j ω0 t
e
F = 2π δ(ω − ω 0 )
2
∴

e−j ω 0 t
F = 2π δ(ω + ω 0 )
2

In conclusion,

1 # $
X(ω) = 2
π δ(ω − ω 0 ) − 2
π δ(ω + ω 0 )
2j
π j
= [δ(ω − ω 0 ) − δ(ω + ω 0 )]
j j
= j π [δ(ω + ω 0 ) − δ(ω − ω 0 )]
420 19 Fourier Transformation Integral

That is to say, Fourier transform of sin(ω 0 t) function is also the sum of two Dirac functions, one
delayed and the other advanced by ω 0 ; however, there is also phase rotation by π/2 (which is set
by the j factor) as well as the negative orientation of the δ(ω − ω 0 ) term.

3. Exponential functions are transformed by Fourier integral as follows.

(a) Product x(t) of exponential function e−at and u(t) is highlighted in Fig. 19.7.
By definition, Fourier integral is
; ∞
X(ω)= x(t) e−j ω t dt
def

−∞
; ∞
= e−at u(t) e−j ω t dt
−∞
; ∞
= e−at e−j ω t dt
0
; ∞
= e−(a+j ω )t dt
0
 
= change of variable
%∞
1 %
=− e−(a+j ω )t %%
a + jω 0
1
1
=− e
(0 − 7
0
)
a + jω
∴
1
X(ω) =
a + jω

Fig. 19.7 Example 19.1-3(a)

x(t)
e−at
u(t)
1
t
0
0
Solutions 421

Fig. 19.8 Example 19.1-3(a) |X(w )|

1/a
√
1/a 2

0 w
−a 0 a
p /2
q (w )
p/4
0 w
− p /4
− p /2
−a 0 a

Modulus |X(ω)| and argument θ(ω) = arg X(ω) of complex function X(ω) are calculated as
usual for complex numbers,
% % % % % %
% 1 a − j ω %% %% a − j ω %% %% a ω %%
%
|X(ω)| = % = = −j 2
a + j ω a − j ω % % a 2 + ω2 % % a 2 + ω2 a + ω2 %
* * 
a2 ω2 a 2
 +ω2 1
= + = =
(a + ω )
2 2 2 (a + ω )
2 2 2
(a 2 + ω2 )2 a + ω2 )
2

or, even faster calculation

% %
% 1 %% 1 1
|X(ω)| = %% = =√
a + j ω % |a + j ω | a 2 + ω2

This even function may be sufficiently evaluated only in a few important points as

1
ω = 0 ∴ |X(ω)| =
a
1
ω = ±a ∴ |X(ω)| = √
a 2
lim |X(ω)| = 0
ω→±∞

and then, its graph is sketched in Fig. 19.8 (top).

However, we pay attention to the signs of (X(ω)) and (X(ω)), so that the argument is
calculated in the correct quadrant of the unity circle. The (X(ω)) > 0 and (X(ω)) < 0;
therefore, the argument θ(ω) is in the fourth quadrant thus negative.

ω
− 
θ(ω)= arctan
def (X(ω))
= arctan a 2
 +ω2 = arctan −ω = − arctan ω
(X(ω)) a a a

a 2
 +ω2
422 19 Fourier Transformation Integral

This odd function may be sufficiently evaluated only in a few important points as

ω = 0 ∴ θ(0) = 0
π
ω = +a ∴ θ(a) = − arctan(1) = −
4
π
ω = −a ∴ θ(−a) = − arctan(−1) =
4
π
lim θ(ω) = −
ω→∞ 2
π
lim θ(ω) =
ω→−∞ 2

and then, its graph is sketched in Fig. 19.8 (bottom).

(b) x(t) = eat u(−t) Product x(t) of exponential function eat and u(−t) is highlighted in
Fig. 19.9. By definition, Fourier integral is
; ∞
X(ω)= x(t) e−j ω t dt
def

−∞
; ∞
= eat u(−t) e−j ω t dt
−∞
; 0
= eat e−j ω t dt
−∞
; 0
= e(a−j ω )t dt
−∞
 
= change of variable
%0
1 %
= e(a−j ω )t %%
a − jω −∞

1
1 7

= ( e − 0)
0
a − jω 
∴
1
X(ω) =
a − jω

Fig. 19.9 Example 19.1-3(b) x(t)

eat
u(−t)
1

0
0
Solutions 423

Fig. 19.10 |X(w )|

Example 19.1-3(b) 1/a
√
1/a 2

0 w
−a 0 a
p /2
q (w )
p /4
0 w
−p /4
−p /2
a 0 a

Modulus |X(ω)| and argument θ(ω) = arg X(ω) of complex function X(ω) are calculated as
usual for complex numbers, Fig. 19.10
% % % % % %
% 1 a + j ω %% %% a + j ω %% %% a ω %%
|X(ω)| = %% = = + j
a − j ω a + j ω % % a 2 + ω2 % % a 2 + ω2 a 2 + ω2 %
* * 
a2 ω2 a 2
 +ω2 1
= + = =
(a + ω )
2 2 2 (a + ω )
2 2 2
(a 2 + ω2 )2 a 2 + ω2 )

or, even faster calculation

% %
% 1 % 1 1
|X(ω)| = %% %=
% |a − j ω | = √a 2 + ω2
a − jω

We note that this |X(ω)| is identical to the one in Example 19.1-3(a).

However, we pay attention to the signs of (X(ω)) and (X(ω)), so that the argument is
calculated in the correct quadrant of the unity circle. The (X(ω)) > 0 and (X(ω)) > 0;
therefore, the argument θ(ω) is in the first quadrant thus positive.

ω

θ(ω)= arctan
def (X(ω)) a +
= arctan 
2  ω2 = arctan ω
(X(ω)) a a

a 2
 +ω2

Similarly to Example 19.1-3(a), this odd function may be sufficiently evaluated only in a few
important points as

ω = 0 ∴ θ(0) = 0
π
ω = +a ∴ θ(a) = arctan(1) =
4
π
ω = −a ∴ θ(−a) = arctan(−1) = −
4
424 19 Fourier Transformation Integral

π
lim θ(ω) =
ω→∞ 2
π
lim θ(ω) = −
ω→−∞ 2

and then, its graph is sketched in Fig. 19.9 (bottom).

(c) Given function x(t) = e−a|t| may be written as

⎧
−at
⎪
⎨e ; t >0
x(t) = 1; t =0
⎪
⎩ at
e ; t <0

Sketches of x(t) and X(ω) are shown in Fig. 19.11. Being a real function X(ω), consequently its
argument θ(ω) = arg X(ω) = 0◦ , i.e. it is a constant.
Fig. 19.11
x(t)
Example 19.1-3(c) e−at eat
1
t
0
0
|X(w )|

1/a

0 w
−a 0 a

Thus, Fourier transformation may be written as

; ∞ ; ∞
X(ω)= x(t) e−j ω t dt = e−a|t| e−j ω t dt
def

−∞ −∞
; 0 ; ∞
= eat e−j ω t dt + e−at e−j ω t dt
−∞ 0
 
= see Example 19.1-3(a) and (b)
1 1 j
a−ω +a+ j
ω
= + =
a − jω a + jω a +ω
2 2

∴
2
X(ω) = ∴ X(ω) ∈ R ∴ θ(ω) = 0◦
a 2 + ω2
Solutions 425

4. The port function x(t) = a0 Π (t/τ ) where t ∈ [−τ/2, τ/2] is transformed by Fourier integral as
follows.
; ∞ ; ∞  ; τ/2
t
X(ω)= x(t) e−j ω t dt = e−j ω t dt = a0 e−j ω t dt
def
Π
−∞ −∞ τ −τ/2
 
= change of variable
%
a0 −j ω t %%τ/2 a0 # −j ω τ/2 $
=− e % =− e − ej ω τ/2
jω −τ/2 j ω
a0 # j ω τ/2 $ 2a0 ej ω τ/2 − e−j ω τ/2
= e − e−j ω τ/2 =
jω ω 2j
 ωτ 
2a0 τ/2  ωτ  τ sin 2
= sin = 2
a0 ωτ
ω τ/2 2 2

2
 ωτ 
= a0 τ sinc
2

Fig. 19.12 Π
Example 19.1-3(a)

Õ(t) 1 sinc(w )
1

t w
0 0

−0.5 0 0.5 −3 −2 −1 0 1 2 3

In conclusion, the Π (t) and sinc (ω) are dual functions with respect to Fourier transformation,
Fig. 19.12.
5. Fourier transform of the triangle function x(t) = Λ (t/τ ) where t ∈ [−τ/2, τ/2] may be
performed by exploring properties of Fourier transform of function derivatives.
Given function f (t), there are the following identities:
 
F f  (x) = j ω F(ω)
 
F f  (x) = (j ω )2 F(ω)
 
F f (3) (x) = (j ω )3 F(ω)
···

Derivatives of triangle function may be deduced by inspection of graphs.

426 19 Fourier Transformation Integral

First derivative of a piecewise linear triangle function may be deduced for each interval separately,
Fig. 19.13. Horizontal parts of x(t) are constant; thus derivative equals zero. Two slopes of the
triangle are constant and found by definition of tangent in a right angle triangle, as a/τ and a/−τ ,
respectively.

dx a
= ; x ∈ [−τ, 0]
dt τ
dx a
= − ; x ∈ [0, τ ]
dt τ

Fig. 19.13 x(t)

a0
Example 19.1-3(a)
t
0

a/t x (t)
t
0
−a/t
a/t x (t) t
0

−2a/t
−t 0 t

Second derivatives are found at points where x  (t) change, i.e. t = (−τ, 0, τ ), where we find
Dirac functions. Sharp step from zero to a/τ at (t = −τ ) generates positive Delta function, sharp
step from zero to −a/τ at t = τ generates negative Delta function, and sharp step from a/τ
to −a/τ at t = 0 generates negative −2a/τ Delta function. In Example 19.1-1(a) and (b) the
duality between delayed Delta and constant functions is derived; thus in combination with the
Fourier transformations of function derivatives we find

d 2x a 2a a
= δ(t + τ ) − δ(t) + δ(t − τ )
dt 2 τ τ τ
a
= [δ(t + τ ) − 2 δ(t) + δ(t − τ )]
τ
F  
−
→ Fourier transformations of function derivatives
a # jω τ $
(j ω )2 X(ω) = e − 2 + e−j ω τ
τ
∴
a # jω τ $ 2a # $
X(ω) = e + e−j ω τ −2 = 2 1 − cos ωτ
−ω τ
2    ω τ   
2 cos(ωτ ) 2 sin2 (ωτ/2)

τ 2a  ωτ  sin (ωτ/2) sin (ωτ/2)

= 2 sin2 = aτ
τ ωωτ 2 ωτ/2 ωτ/2
 ωτ   ωτ 
= aτ sinc sinc
2 2
Solutions 427

 ωτ 
= aτ sinc2
2

In conclusion, Λ(t) and sinc2 [ω] are dual functions with respect to Fourier transformation,
Fig. 19.14.

 
t L t at sinc2 wt
t 2

t
0 0 w
−t 0 t −1 0 1

Fig. 19.14 Example 19.1-5

6. Sign function sign (t) does not converge when t → ∞; thus direct application of Fourier integral
also produces non-converging result. One possible method to derive Fourier transformation is to
express sign (t) in terms of u(t) and then to use limit of exponential function.
Specifically, we write

sign (t) = u(t) − u(−t)

In addition, we can interpret the number one as the limit of exponential function when a = 0 (ie.
e0 = 1), highlighted function in Fig. 19.15, as

u(t) = lim e−at u(t)

a=0

−u(−t) = − lim eat u(−t)

a=0

By doing so, we write

Fig. 19.15 Example 19.1-6

u(t) u(t)
1
lima→f
e−at u(t)
t
0
0
u(t) t
0
eat u(t)
lima→f
−u(−t)
0
428 19 Fourier Transformation Integral

# $
x(t) = sign (t) = u(t) − u(−t) = lim e−at u(t) − eat u(−t)
a→0
 
= see Example 19.1-3(a) and (b)
∴
⎡ ⎤
1 1 ⎦= 1 − 1
X(ω) = lim ⎣ 0
− 0
a→0 jω jω
a + jω
 a − jω

2
=
jω

Second, and much faster, method to derive this Fourier transform function is based on the same
idea as in Example 19.1-5, that is to say to exploit the derivative property of Fourier transform as

dx(t) F 2
= 2 δ(t) −
→ j ω X(ω) = 2 ∴ X(ω) =
dt jω

7. Step function u(t) does not converge when t → ∞; thus direct application of Fourier integral
also produces non-converging result. We can use the same technique as in Example 19.1-6 to
write u(t) in terms of sign (t) that is already derived in Example 16.2-2(c). Then

1 + sign (t)
x(t) = u(t) =
2
∴
1 sign (t)  
X(ω) = F + = F( constant ) + F(sign (t) )
2 2

1 1 1 1 2

=F + F(sign (t)) = 2
π δ(ω) +
2 2 2
2
j ω
1
= π δ(t) +
jω

8. Formal derivation of Fourier transformation of cardinal sine function sinc (t) may be used a study
case to illustrate various techniques in algebra and calculus working together.

First, let us do some preparatory work and quick reviews:

(a) Reminder of basic trigonometric identities

1 
sin x cos y = sin(x + y) + sin(x − y)
2
1 
sin x sin y = cos(x − y) − cos(x + y)
2
cos(−x) = cos(x)
sin(−x) = − sin(x)
Solutions 429

(b) Reminder of Euler formula

e±j x = cos x ± j sin x

(c) Reminder of special functions

⎧
⎪ 
|x| x ⎨−1; x<1
x; x≥0
sign (x) = = = 0; x=0 and, |x| =
x |x| ⎪
⎩ −x; x<0
1; x>0

(d) Reminder of partial integration

; ;
u dv = u v − v du

(e) Reminder of basic integrals

;
1
dx = ln |x| + C
x
;
1
dx = arctan x + C
1 + x2
;
ex dx = ex + C

(f) Solve the following integral:

; ∞ ; ∞ .x /
a a 1
dx =
 x 2 dx = = t ∴ dx = |a| dt
∞ a2 + x 2 a 2∞ a
1+
a
; %∞
|a| ∞ 1 |a| %
= dt = arctan(t) %%
a 1+t 2
∞ a −∞

|a|  π  π 
= − −
a 2 2
= π sign (a)
430 19 Fourier Transformation Integral

(g) Solve the following integral

; ∞
e−ω 0 t y dy =
0
⎧ ⎫
⎪
⎨change variable: −ω 0 t y = x ∴ dy = −
1
dx ⎪
⎬
ω0 t
⎪
⎩ ⎪
⎭
y = 0 ⇒ x = 0, and y = ∞ ⇒ x = ∞
; %∞ 1
1 ∞
1 % 1 *0


=− ex dx = − ex %% = − 
−∞
− 7
0
ω0 t ω0 t ω
e 
e
0 0 0 t

1
=
ω0 t

(h) Solve the following integral where x ∈ [0, ∞]; first we solve
;
e−bx sin(ax) dx =
⎧ ⎫
⎪
⎨partial integration: u = e−bx ∴ du = −b e−bx dx ⎪
⎬
⎪ 1
⎩ dv = sin(ax) dx ∴ v = − cos(ax) ⎪ ⎭
a
 ; 
1 1  
= e−bx − cos(ax) − − cos(ax) −b e−bx dx
a a
;
1 b
= − e−bx cos(ax) − e−bx cos(ax) dx
a a
  
partial integration
⎧ ⎫
−bx
⎪
⎨partial integration: u = e ∴ du = −b e−bx dx ⎪
⎬
⎪ 1
⎩ dv = cos(ax) dx ∴ v = sin(ax) ⎪
⎭
a
! ; "
1 b 1 −bx b
= − e−bx cos(ax) − e sin(ax) + e −bx
sin(ax) dx
a a a a
Solutions 431

(h) (cont.) We note that after two partial integrations one of the terms on the right side is identical
to the original integral on the left side. Thus further partial integration does not advance the
solution. However, like any other common term in an algebra equation, this integral may be
factored out as
!  2" ;
b 1 b 1 −bx
1+ e−bx sin(ax) dx = − e−bx cos(ax) − e sin(ax)
a a a a

∴
;
a 2 + b2 1 a cos(ax) + b sin(ax)
e−bx sin(ax) dx = − e−bx
a2 a a

which gives
; 
−bx a2 1 1 −bx
e sin(ax) dx = 2 − e [a cos(ax) + b sin(ax)] + C
a + b2 a
a

1
=− e−bx [a cos(ax) + b sin(ax)] + C
a2 + b2

Now, we calculate definite integral (the integration constants cancel) as

; %∞
∞
1 %
e −bx
sin(ax) = − 2 e −bx
[a cos(ax) + b sin(ax)] %%
0 a + b2 0

1 ∞:0

=− 2 e−b a cos(a ∞) +b sin(a ∞)
a + b2      
≤1 ≤1
  
=0

·
e−b
−
:
0 1
a cos(a · 0) +b sin(a · 0)
     
=1 =0
a
= 2
a + b2

Therefore,
; ∞
a
e−bx sin(ax) =
0 a 2 + b2

By repeating the same idea, we find

; ∞
b
e−bx cos(ax) =
0 a2 + b2
432 19 Fourier Transformation Integral

After this preparatory work, we derive Fourier transform of sinc (t) function as follows:
; ∞ ; ∞
sin(ω 0 t) −j ω t
X(ω)= x(t) e−j ω t dt =
def
e dt
−∞ −∞ ω0 t
; ∞  
1
= sin(ω 0 t) cos(ω t) − j sin(ω t) dt
−∞ ω0 t
; ∞ ; ∞
1 1
= sin(ω 0 t) cos(ω t) dt − j sin(ω 0 t) sin(ω t) dt
−∞ ω0 t −∞ ω0 t
= I1 − j I2

In Example 19.1-8(d) we solved integral

; ∞
1
e−ω 0 t y dy =
0 ω0 t

whose solution appears as the leading term in both I1 and I2 . Thus, that term may be replaced with
the integral itself. In addition, we keep in mind that ω 0 is a constant, while ω is the integration
variable and that integration is commutative. By doing so, after applying the trigonometric
identities, the two integrals I1 and I2 are solved separately as follows. First, solution to I1
; ∞
1
I1 = sin(ω 0 t) cos(ω t) dt
−∞ ω0 t
; ∞; ∞  
1
= e−ω 0 t y dy sin[(ω 0 + ω) t] + sin[(ω 0 − ω) t] dt
2 −∞ 0
; ∞ ; ∞  
1
= e−ω 0 t y dy sin[(ω + ω 0 ) t] − sin[(ω − ω 0 ) t] dt
2 −∞ 0
; ∞ ; ∞
1
= dy e−ω 0 y t sin[(ω + ω 0 ) t] dt
2 −∞ 0
; ∞ ; ∞ 
−ω 0 y t
− dy e sin[(ω − ω 0 ) t] dt
−∞ 0
 
see: Example 19.1-8(h)
; ;
1 ∞ ω + ω0 1 ∞
ω − ω0
= dy − dy
2 −∞ (ω + ω 0 )2 + (ω 0 y)2 2 −∞ (ω − ω 0 )2 + (ω 0 y)2
 
see: Example 19.1-8(f)
π
= sign (ω + ω 0 ) − sign (ω − ω 0 )
2

The last result defines Π (ω/2ω 0 ) as illustrated in Fig. 19.16 (a) to (d), which is well known dual
function of sinc (t) function. That being the case, the question is: what is then I2 ?
Solutions 433

Fig. 19.16 Example 19.1-8 (a) x( w ) sign ( w + w 0 ) (c) x( w )

1 2
w sign ( w + w 0 ) − sign ( w − w 0 )
0
−1 w
0
−w 0 0 w0 −w 0 0 w0
(b) x( w ) (d) x(w )
1 p
−sign ( w − w 0 ) w p [sign ( w + w 0 ) − sign ( w − w 0 )]
0 2

−1 w
0
w 0 0 w0 w0 0 w0

Solution to I2 follows the same idea as for I1 ; however, there are slight differences between the
two paths to solutions.
; ∞
1
I2 = sin(ω 0 t) sin(ω t) dt
−∞ ω0 t
; ∞; ∞  
1
= e−ω 0 t y dy cos[(ω − ω 0 ) t] − cos[(ω + ω 0 ) t] dt
2 −∞ 0
; ∞ ; ∞
1
= dy e−ω 0 y t cos[(ω − ω 0 ) t] dt
2 −∞ 0
; ∞ ; ∞ 
− dy e−ω 0 y t cos[(ω + ω 0 ) t] dt
−∞ 0
; ∞ ; ∞
1 ω0 y 1 ω0 y
= dy − dy
2 −∞ (ω − ω 0 )2 + (ω 0 y)2 2 −∞ (ω + ω 0 )2 + (ω 0 y)2
⎧ ⎫
⎪
⎪ 1 ⎪
⎪
⎨x = (ω ± ω 0 ) + (ω 0 y) ∴ dx = 2ω 0 y dy ∴ ω 0 y dy = dx ⎪
2 2
⎪
⎬
2
; ∞ %∞
⎪
⎪ 1 dx 1 % ⎪
⎪
⎪
⎩ ∴ = ln x %% ⎪
⎭
2 −∞ x 2 −∞

1   %%∞   %%∞
2 % 2 %
= ln (ω − ω 0 ) + (ω 0 y) %
2
− ln (ω + ω 0 ) + (ω 0 y) %
2
4 −∞ −∞

1   %%∞   %%
= ln (ω − ω 0 )2 + (ω 0 y)2 %% − ln (ω − ω 0 )2 + (ω 0 y)2 %%
4 −∞
  %%∞   %%
− ln (ω + ω 0 )2 + (ω 0 y)2 %% + ln (ω + ω 0 )2 + (ω 0 y)2 %%
−∞
. a /
ln a + ln b = ln(ab); ln a − ln b = ln ; ln a b = b ln a
b
! %∞ % "
1 (ω − ω 0 )2 + (ω 0 y)2 %% (ω − ω 0 )2 + (ω 0 y)2 %%
= ln − ln
4 (ω + ω 0 )2 + (ω 0 y)2 % (ω + ω 0 )2 + (ω 0 y)2 %−∞
 
1 (ω − ω 0 )2 + (ω 0 y)2 (ω − ω 0 )2 + (ω 0 y)2
= lim ln − lim ln
4 y→∞ (ω + ω 0 )2 + (ω 0 y)2 y→−∞ (ω + ω 0 )2 + (ω 0 y)2
434 19 Fourier Transformation Integral

 
1 (ω − ω 0 )2 + (ω 0 y)2 (ω − ω 0 )2 + (ω 0 y)2
= ln lim − ln lim
4 y→∞ (ω + ω 0 )2 + (ω 0 y)2 y→−∞ (ω + ω 0 )2 + (ω 0 y)2
∞
=
∞
   
1 2ω
 0y 2ω
 0y
= −
L.H.
ln lim  ln lim 
4 y→∞ 2ω0y y→−∞  2ω
0y

1
= [ln(1) − ln(1)] = 0
4

In conclusion,
; ∞ 0
sin(ω 0 t) −j ω t
X(ω) = e I
dt = I1 + j 7
2 = I1
−∞ ω0 t
π
= sign (ω + ω 0 ) − sign (ω − ω 0 )
2

Exercise 19.2, page 414

1. Given x(t) may be decomposed into the sum of two Π (t) functions, x1 (t) and x2 (t) as illustrated
in Fig. 19.17, where
 
t t
x1 (t) = Π and x2 (t) = Π
4 2

Therefore,
 
t t
x(t) = Π +Π
4 2

Fig. 19.17 Example 19.2-1 x(t)

2
1
t
0
x1 (t)
2
1
t
0
x2 (t)
2
1
t
0
−2 −1 0 1 2
Solutions 435

Fourier transformation is therefore the sum of two transformations as

   
t t
X(ω) = F Π +F Π = 4sinc (2ω) + 2sinc (ω)
4 2

2. Using, for example, properties of Fourier transformation with regard to derivatives given x(t) we
write

x(t) = −2sign (t − 2) ∴ x  (t) = −4δ(t − 2) ∴ j ω X(ω) = −4e−2j ω

∴
4 −2j ω
X(ω) = − e
jω

3. Given RF pulse h(t) may be decomposed into the product of x(t) = cos(t) and Π(t) functions,
Fig. 19.18, i.e.

t
h(t) = A cos(2πfc t) Π
T

where fc is the RF waveform frequency, and T is the duration of RF pulse.

Fig. 19.18 Example 19.2-3 x(t)

A
t
0
−A

P (t)
1
t
0
− T /2 0 T /2

By using Euler’s formula, we write

1# $
h(t) = cos(2πfs t) = exp(j 2πfc t) + exp(−j 2πfc t)
2

Fourier transformation is therefore derived with the help of frequency shifting property, as
 
1 t 1 t
h(t) = Π exp(j 2πfc t) + Π exp(−j 2πfc t)
2 T 2 T
∴
AT
F(h(t)) = sinc (T (f − fc )) + sinc (T (f + fc ))
2
436 19 Fourier Transformation Integral

That is to say, if fc T  1, we write, see Fig. 19.19

⎧
⎪ AT
⎪
⎪ sinc (T (f − fc )) ; f >0
⎨ 2
F(h(t)) = 0 ; f =0
⎪
⎪
⎪
⎩ AT sinc (T (f + fc )) ; f <0
2

Fig. 19.19 Example 19.2-3 AT

H( w )
2

0 w
w c 0 wc
 Discrete Fourier Transformation
20

Important to Know

Fourier sum: is defined as

−1 2π k

N
−j n
F(xn ) = Xk =
def
xn e N
n=0
∞ 
−1 1  2π k
F (X[k]) = x[n] = X[k] exp j n
T −∞ N

Fourier series: is defined as

∞
 2π
x(t) = Ck e j ω 0 t , ω 0 =
k=−∞
T
; t0 +T
1
Ck = x(t) e−j ω 0 t dt
T t0

20.1 Exercises

20.1 ** Discrete Fourier Sum

1. Derive discrete Fourier transformations X[k] of the following discrete series x[n].
(a) x[n] = {1, 2, 3, 4} (b) x[n] = {2, 3, −1, 1}
(c) x[n] = δ[n], N = 16 (d) x[n] = δ[n − n0 ]
(e) x[n] = 1, N = 16 (f) x[n] = α n u[n − 1], (|α| < 1)

2. Derive inverse discrete Fourier transformations x[n] of the following discrete series X[k].
(a) X[k] = {10, −2 + 2j, −2, −2 − 2j } (b) X[k] = {5, 3 − j 2, −3, 3 + j 2}

© Springer Nature Switzerland AG 2021 437

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0_20
438 20 Discrete Fourier Transformation

3. Derive discrete Fourier transformations X[k] of the following functions x(t).

(a) x(t) = sin(1 Hz), sampling frequency fs = 8 Hz.

20.2 ** Fourier Series

1. Given square waveform in Fig. 20.1 (left) derive Fourier series assuming
(a) a = T /4 (b) a = T /8

(a) a = T /4 (b) a = T /8

x(t) y(t)
1 1

t t
0 0

−a 0 a T 0 2a T

Fig. 20.1 Example 20.2-1 (left) and 2 (right)

2. Further to result derived in Example 20.2-1, given square waveform in Fig. 20.1 (right), derive
Fourier series by applying the time-shifting property.
3. Given x(t) waveform in Fig. 20.2, derive its Fourier series.

Fig. 20.2 Example 20.2-3 x(t)

A
t
0

−A
−2 −1 0 1 2

4. Given f (t) = 1 − (t/π ), derive its Fourier series if t ∈ [−π, π ].

20.3 ** Fast Fourier Transformation

1. Assuming the audio spectrum signal, [20 Hz, 20 kHz], estimate the time needed to calculate
discrete Fourier transformation (DFF) by hands, if the desired resolution is 1 Hz and it takes
1 min to perform one calculation (extremely fast, but for the sake of argument). Then, compare
with the time needed to do the same calculations by using Fast Fourier Transformation (FFT).
2. Given a general discrete series xn = x[n], develop transformation equations and illustrate the
associated graph diagram of FFT algorithm if
(a) N = 2 (b) N = 4

3. By using graph diagram of FFT in Fig. 20.15, calculate Fourier transformation of x[n] =
{1, 2, 3, 4} (already solved using DFT in Example 20.1-1(a)).
Solutions 439

4. Following the logic developed in Example 20.3-2,

(a) Deduce FFT graph diagram if the number of transformation points is N = 8.
(b) Using FFT graph diagram calculate F(xn ) if xn = [1, 1, −1, −1, 1, 1, −1, −1].

Solutions

Exercise 20.1, page 437

1. Fourier sum for a simply defined function is derived by definition.

(a) Given a simply defined function x[n] = {1, 2, 3, 4}, we deduce N = 4 and we write


N −1  3 
2π k 2π k
X[k] = x[n] exp −j n = x[n] exp −j n
n=0
N n=0
4

Calculations for each k are as follows.


3  3 1
2π 0
k = 0 : X[0] = x[n] exp −j n = e
x[n] 7
0

n=0
4 n=0

= x[0] + x[1] + x[2] + x[3] = 1 + 2 + 3 + 4

= 10

3  3  π 
2
1 π 1
k = 1 : X[1] = x[n] exp −j n = x[n] exp −j n
n=0
4
2 n=0   2 
(−j )n
 
see the examples for calculating powers of complex numbers
= x[0] (−j )0 + x[1] (−j )1 + x[2] (−j )2 + x[3] (−j )3
= 1(1) + 2(−j ) + 3(−1) + 4(j ) = 1 − 2j − 3 + 4j
= −2 + 2j

3   3
2
π 2

k = 2 : X[2] = x[n] exp −j n = x[n] exp (−j π n)
4
1   
n=0 n=0 n (−1)

= x[0] (−1)0 + x[1] (−1)1 + x[2] (−1)2 + x[3] (−1)3

= 1(1) + 2(−1) + 3(1) + 4(−1) = 1 − 2 + 3 − 4
= −2
440 20 Discrete Fourier Transformation

(a) (cont.)


3  3 
2
1 π 3 3π
k = 3 : X[3] = x[n] exp −j n = x[n] exp −j n
4
2 2
n=0 n=0   
(j )n

= x[0] (j )0 + x[1] (j )1 + x[2] (j )2 + x[3] (j )3

= 1(1) + 2(j ) + 3(−1) + 4(−j ) = 1 + 2j − 3 − 4j
= −2 − 2j = x[1]∗

Therefore, after recalling that θ = −π ≡ π , and θ = −(3π/4) ≡ (5π/4), we write its graphical
representation in Fig. 20.3 and the complete solution as

X[k] = {10, −2 + 2j, −2, −2 − 2j }

. √ √ /
|X[k]| = 10, 2 2, 2, 2 2
 
3π 5π
θk = 0, , π,
4 4

Fig. 20.3 Example 20.2-1 |X[k]|

10

√
2 2
2 k
0
q [k]
p
p
2
0
k
0 1 2 3

(b) Given function x[n] = {2, 3, −1, 1} we conclude N = 4; by definition we write


N −1  3 
2π k 2
π k
X[k] = x[n] exp −j n = x[n] exp −j n
n=0
N n=0
4
2
Solutions 441

Thus,


3  3 3
π0 :1 =

k = 0 : X[0] = x[n] exp −j n = x[n]  (0)
exp x[n]
n=0
2 n=0 n=0

=2+3−1+1=5

3 
π1
k = 1 : X[1] = x[n] exp −j n
n=0
2
   
0π 1π 2π 3π
= 2 exp −j + 3 exp −j − 1 exp −j + 1 exp −j
2 2 2 2
= 2 + 3(−j ) − 1(−1) + 1(j )
= 3 − 2j

3 
π 2

k = 2 : X[2] = x[n] exp −j n
n=0
2

= 2 exp (−j (0π )) + 3 exp (−j (1π )) − 1 exp (−j (2π )) + 1 exp (−j (3π ))
= 2 + 3(−1) − 1(1) + 1(−1)
= −3

3 
π3
k = 3 : X[3] = x[n] exp −j n
n=0
2
   
π3 π3 π3 π3
= 2 exp −j 0 + 3 exp −j 1 − 1 exp −j 2 + 1 exp −j 3
2 2 2 2
= 2 + 3(j ) − 1(−1) + 1(−j )
= 3 + 2j

Therefore,

X[k] = {5, 3 − 2j, −3, 3 + 2j }

. √ √ /
|X[k]| = 5, 13, 3, 13

θ[k] = {0, −33.7◦ , π, 33.7◦ }

442 20 Discrete Fourier Transformation

(c) Given Delta function x[n] = δ[n] in N = 16 points, that is to say x[0] = 1 otherwise
x[n] = 0, n = 0, or

x[n] = {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}

By definition we write


N −1  15 
2π k 2
π k
X[k] = x[n] exp −j n = x[n] exp −j n
N 8
16
n=0 n=0

Calculations for each k are as follows.


15   15
π0 :1

k = 0 : X[0] = x[n] exp −j n = x[n]  (0)
exp
n=0
8 n=0

= x[0] + x[1] + x[2] + · · · + x[15] = 1 + 0 + 0 + · · · + 0

=1

15  15  π 
π1
k = 1 : X[1] = x[n] exp −j
n = x[n] exp −j n
n=0
8 n=0
8
 π 
= x[0] exp −j 0 + 0 + 0 + · · · + 0
8
=1

15  15 
π2 2π
k = 2 : X[2] = x[n] exp −j n = x[n] exp −j n
n=0
8 n=0
8

2π
= x[0] exp −j 0 + 0 + 0 + ··· + 0
8
=1
..
.
etc.

In other words X[k] = 1, ∀k, which is, in general, a constant function that is known to be dual
function of impulse.
Solutions 443

(d) DFT of delayed impulse x[n] = δ[n − n0 ] we derive as

∞
  ∞ 
2π k 2π k
X[k] = x[n] exp −j n = δ[n − n0 ] exp −j n
n=−∞
N n=−∞
N

2π k
= · · · + 0 + 0 + δ[n0 − n0 ] exp −j n0 +0 + 0 + · · ·
N
  
n=n0

2π k
= 1 exp −j n0
N

Therefore, |X[k]| = 1, ∀k, and the argument is (2π k n0 /N), which illustrates the property of
delayed functions in general where the delay transforms into the equivalent argument (i.e. angle
or phase).
(e) A constant function x[n] = 1 defined in N = 16 points is transformed as


N −1  15 
2π k 2
π k
X[k] = x[n] exp −j n = 1 exp −j n
N 8
16
n=0 n=0

Calculations for each k are as follows.


15   15  15
π0 :1 =

k = 0 : X[0] = exp −j n =  (0)
exp 1n
n=0
8 n=0 n=0

= 16

15  15  π 
π1
k = 1 : X[1] = exp −j n = exp −j n
n=0
8 n=0
8
 π   π   π 
= exp −j 0 + exp −j 1 + · · · + exp −j 15
8 8 8

The question is: how to do summation of the last sum? First, modules of all terms equal to one.
Second, each term in the sum is a complex number that is shifted along the circle at the multiples
of π/8, that is to say all sixteen terms are uniformly distributed on the unity circle in the complex
plain, see Fig. 20.4. We note that, for example, x[6] and x[14] are at the opposite sides; therefore,
their sum equals zero. Formally, we can confirm that statement as
444 20 Discrete Fourier Transformation

Fig. 20.4 Example 20.1-1

x11 x12 x13
x10 j x14
x9 x15
x8 x0
−1 1

x7 x1
x6 −j x2
x5 x4 x3

 π   π 
x[6] + x[14] = exp −j 6 + exp −j 14
8 8
 
2
π 2
π
= exp −j 3 + exp −j 7
8
4 8
4
  π
3π
= exp −j + exp j
4 4
  π  π 
3π 3π
= cos − + j sin − + cos + j sin
4 4 4 4
√ √ √ √
2 2 2 2
=− −j + +j
2 2 2 2
=0

Similarly, we show that all other sums of the opposite pairs also equal zero, i.e.

x[0] + x[8] = 0 x[4] + x[12] = 0

x[1] + x[9] = 0 x[5] + x[13] = 0
x[2] + x[10] = 0 x[6] + x[14] = 0
x[3] + x[11] = 0 x[7] + x[15] = 0

Same calculation is valid for all the other values of k; therefore, we conclude

X[0] = 16
 π   π   π 
X[1] = exp −j 0 + exp −j 1 + · · · + exp −j 15 = 0
8 8 8
X[2] = X[3] = · · · = X[15] = 0
Solutions 445

In summary, given N = 16

F(x[n] = 1) = 16 δ[n]

and X[k] being real function, the argument θ[k] = 0.

(f) Given infinitely long series x[n] = α n u[n−1], (|α| < 1) we apply Fourier sum as follows:

∞
  ∞
 
2π k 2π k
X[k] = x[n] exp −j n = α u[n − 1] exp −j
n
n
n=−∞
N n=−∞
N
 
delayed step function u[n − 1] = 0, ∀n < 1 ∴ u[n − 1] = 1, ∀n ≥ 1
 ∞  ∞ 
2π k 2π k n
= α (1) exp −j
n
n = α exp −j
n=1
N n=1
N
.  b /
x a y a = (xy)a ; as well as x ab = x a
 N 
 a N +1 − a m
a =
k
, (|a| < 1)
k=m
a−1
:0


 ∞+1 
2π k 1
2π
  k
α exp 
−j − α exp −j
 N N
= 
2π k
α exp −j −1
N
%  %  
% %
%exp −j 2π k % = 1, α < 1 ∴ α exp −j 2π k < 1
% N % N

2π k
α exp −j
N
= 
2π k
1 − α exp −j
N
446 20 Discrete Fourier Transformation

2. Inverse Fourier sum is derived by definition.

(a) Given a simply defined function X[k] = {10, −2 + 2j, −2, −2 − 2j }, we deduce N = 4
and we write
∞  
1  1
3
−1 2π k 2π k
x[n] = F (X[k]) = X[k] exp j n = X[k] exp j n
T −∞ N 4 n=0 4
 
1 2π 0 :1 2
π 1
=  + +
 4
10 exp j n (−2 j 2) exp j n
4 4
2
 
2
π 2
2
π 3
− 2 exp j n + (−2 − j 2) exp j n
4
4
2
1  π 
n = 0 : x[0] = 10 + (−2 + j 2) exp j 0
4 2

3π
− 2 exp (j π 0) + (−2 − j 2) exp j 0
2
1# $
= 10 − 2 + j 2 − 2 − 2 − j 2
4
=1

Therefore,

 π  j
:
1 
n = 1 : x[1] = 10 + (−2 + j 2)  
exp j 1
4 2
−j
 *


 
: −1 3π
− 2 π 1)
(j
exp + (−2 − j 2) exp j 1
 2
1
= 10 + (−2 + j 2)(j ) − 2(−1) + (−2 − j 2)(−j )
4
1
= 10 + H
−2j
H−2
+2
+@ @−2
2j
4
=2
 π  
1 : −1
n = 2 : x[2] = 10 + (−2 + j 2)  
exp j 2

4 2

 * −1

 :
 1 3π 
− 2 π 2) + (−2 − j 2) exp 
(j
exp j  2

  2

1
= 10 + (−2 + j 2)(−1) − 2(1) + (−2 − j 2)(−1)
4
1
= 10 + 2 − 2j − 2
+ 2
+ 2j
4
=3
Solutions 447

(a) (cont.)

 π  −j
:
1 
n = 3 : x[3] = 10 + (−2 + j 2)  
exp j 3
4 2
j
 *

  : −1
 3π

− 2  + −
 2
exp (j π 3) (−2 j 2) exp j 3

1
= 10 + (−2 + j 2)(−j ) − 2(−1) + (−2 − j 2)(j )
4
1
= 10 + 2j + 2 + 2 − 2j + 2
4
=4

In conclusion, x[n] = {1, 2, 3, 4}, as same as in Example 20.1-1(a).

(b) Given a simply defined function X[k] = {5, 3 − j 2, −3, 3 + j 2}, we deduce N = 4 and we
write
∞  
1  1
3
2π k 2π k
x[n] = F−1 (X[k]) = X[k] exp j n = X[k] exp j n
T −∞ N 4 n=0 4
 
1 2π 0  :1 2
π 1
=  + (3 − j 2) exp j
 4
5 exp j n n
4 4
2
 
2
π 2
2
π 3
− 3 exp j n + (3 + j 2) exp j n
4
4
2

Therefore,

1  π 
n = 0 : x[0] = 5 + (3 − j 2) exp j 0
4 2

3π
− 3 exp (j π 0) + (3 + j 2) exp j 0
2
1# $
= 5+3−@2 − 3
+ 3
+ @
j@ 2
j@
4
=2
448 20 Discrete Fourier Transformation

(b) (cont.)

 π  j
:
1 
n = 1 : x[1] = 5 + (3 − j 2)  
exp j 1
4 2
−j
 *

  : −1
 3π

− 3  + +
 2
exp (j π 1) (3 j 2) exp j 1

1
= 5 + (3 − j 2)(j ) − 3(−1) + (3 + j 2)(−j )
4
1
= @+2+3−@
5+@
3j @+2
3j
4
=3
 π  
1 : −1
n = 2 : x[2] = 5 + (3 − j 2)  
exp j 2

4 2

 * −1

 :
 1 3π 
− 3 π 2) + (3 + j 2) exp 
(j
exp j  2

  2

1
= 5 + (3 − j 2)(−1) − 3(1) + (3 + j 2)(−1)
4
1
= 5 − 3 + 2j − 3 − 3 − 2j
4
= −1
 π  −j
:
1 
n = 3 : x[3] = 5 + (3 − j 2)  
exp j 3
4 2
j
 *

   −1
: 3π

− 3 (j π 3)
exp + (3 + j 2) exp j 3
 2

1
= 5 + (3 − j 2)(−j ) − 3(−1) + (3 + j 2)(j )
4
1
= 5 − 3j − 2 + 3 + 3j − 2
4
=1

In conclusion, x[n] = {2, 3, −1, 1}, as same as in Example 20.1-1(b).

3. Given x(t) = sin(1 Hz), in the first step it is necessary to convert continuous function x(t) into its
discrete version sampled at N = 8 points. Given the sampling frequency fs = 8 Hz we normalize
period T (1 Hz) = 1 s as

1
f = 1Hz ⇒ T = = 1 s ⇔ 2π
f
Solutions 449

fs = 8 Hz ⇒ N = 8
∴
1 2π π
s= =
8 8 4

that is to say, each sample is found at the multiples of π/4, see Fig. 20.5.
We calculate N = 8 samples as

Fig. 20.5 Example 20.1-3

f (t)
1

0 t

−1
0 p 2p

x[0] = (1) sin(0) = 0

√ π 
π  2 x[4] = (1) sin 4 =0
x[1] = (1) sin = 4
4 2 √
π  2
π  x[5] = (1) sin 5 =−
x[2] = (1) sin 2 =1 4 2
4 π 
√ x[6] = (1) sin 6 = −1
π  2 4
x[3] = (1) sin 3 = √
4 2 π  2
x[7] = (1) sin 7 =−
4 2

Therefore,
 √ √ √ √ 
2 2 2 2
x[n] = 0, , 1, , 0, − , −1, −
2 2 2 2

We note
√ √ √ √

7
2 2 2 2
x[n] = 0 + +1+ +0− −1− =0
n=1
2 2 2 2
450 20 Discrete Fourier Transformation

By definition, we write


N −1  7 
2π k 2
π k
X[k] = x[n] exp −j n = x[n] exp −j n
n=0
N n=0
8
4

Calculations for each k are as follows.


7  7 1  7
π0
k = 0 : X[0] = x[n] exp −j n = e
x[n] 7
0
= x[n] = 0
n=0
4 n=0 n=0


7 
π1
k = 1 : X[1] = x[n] exp −j n
n=0
4
√  π   π 
2
= (0) exp(0) + exp −j 1 + (1) exp −j 2
2 4 4
√     √  π 
2 π π 2
+ exp −j 3 + (0) exp −j 4 − exp −j 5
2 4 4 2 4
 π  √  π 
2
− (1) exp −j 6 − exp −j 7
4 2 4
 
cos(−x) = cos(x) sin(−x) = − sin(x)
√  π   π   
2 2π 2π
= cos − j sin + (1) cos − j sin + ···
2 4 4 4 4
= (0.5 − j 0.5) + (−j ) + (−0.5 − j 0.5) + (0.5 − j 0.5) + (−j ) + (−0.5 − j 0.5)
= −4j


7 
π 2

k = 2 : X[2] = x[n] exp −j n
n=0
4
2
√  π  π 
2
= (0) exp(0) + exp −j + (1) exp −j 2
2 2 2
√     √  π 
2 π π 2
+ exp −j 3 + (0) exp −j 4 − exp −j 5
2 2 2 2 2
 π  √  π 
2
− (1) exp −j 6 − exp −j 7
2 2 2
√       
2 π π 2π 2π
= cos − j sin + (1) cos − j sin + ···
2 2 2 2 2
√ √ √ √
2 2 2 2
= −j −1+j +j +1−j
2 2 2 2
=0
Solutions 451

Similarly,

X[2] = X[3] = X[4] = X[5] = X[6] = 0

X[7] = · · · = 4j

In summary,

X[k] = {0, −4j, 0, 0, 0, 0, 0, 4j }

|X[k]| = {0, 4, 0, 0, 0, 0, 0, 4}

This Fourier transformation result is referred to as “bilateral” as illustrated in Fig. 20.6 (top).
There are two impulses in |X[k]|; however, we know that sinusoidal function is dual with only
impulse.
In order to show amplitude |X[k]| relative to the units of frequency, i.e. [Hz], we must calculate
resolution of each “bin” as

fs 8 Hz
= = 1 Hz/bin
N 8

Fig. 20.6 Example 20.1-3

4 |X[k]|

k
0
0 1 2 3 4 5 6 7

4 |X[k]| fs /2

folding

0
f [Hz]
0 1 2 3 4 5 6 7

In this case, because N and sampling frequency fs are numerically equal it happens that k maps
directly into Hz. That is to say, by selecting different either N or fs we can arbitrary set the res-
olution of Fourier transformation. In addition, impulses found above half of sampling frequency,
known as Nyquist frequency, are folded over in “unilateral” view of Fourier transformation (as
commonly shown by spectrum analyzers or simulated graphs), see Fig. 20.6 (bottom).
Consequently, in unilateral view amplitudes at each frequency are calculated as the sum of
samples already in place and the ones “folded over”.
For example amplitude of sample at x(1 Hz) is increased by amplitude of x(7 Hz) that after
folding over falls on top, thus |H [1 Hz]| = 4 + 4 = 8. This amplitude is the consequence of
N = 8; thus in order to show correct amplitude, i.e. |x(t)| = 1, the unilateral graph must be
recalibrated by dividing the amplitudes by N, see Fig. 20.7.
452 20 Discrete Fourier Transformation

Fig. 20.7 Example 20.1-3 |X[k]|

(1) 8 fs /2

0
f [Hz]
0 1 2 3 4

Exercise 20.2, page 438

1. Square pulse waveform that is even (i.e. symmetric around x = 0) may be decomposed on Fourier
series as follows. By definition, coefficients of Fourier series for periodic function x(t) = sin(ωt)
are
∞
 2π
x(t) = Cn exp (j nω 0 t) ; where, ω 0 = and
k=−∞
T
; t0 +T
1
Cn = x(t) exp (−j nω 0 t) dt
T t0

Case n = 0: note that x(t) = 1, t ∈ [−a, a], thus

; ;
1 t0 +T
 :1
 1 a
2a
C0 = x(t) exp 
(−j0ω 0 t) dt = (1) dt =
T   T T
t0 −a

otherwise,
; t0 +T ; a
2 2
Cn = x(t) exp (−j nω 0 t) dt = (1) exp (−j nω 0 t) dt
T t0 T −a

2 1
= exp (−j nω 0 a) − exp (j nω 0 a)
−j nω 0 T
2T 1 1
= exp (−j nω 0 a) − exp (j nω 0 a)
nπ T 2j

2 2nπ
= sin a
nπ T

Relative to the period T we calculate

Solutions 453

(a) a = T /4: therefore,

2a 2
T 1
C0 = = =
T T 4
2 2
 
2 2nπ 2 2
nπ T sin (nπ/2)
Cn = sin a = sin = = sinc (nπ/2)
nπ T nπ T 4
2 nπ/2
∴
|Cn | = |sinc (nπ/2) |

We calculate a few samples,

π 
C1 =
2
sin =
2 
2 4π
π 2 π C4 = sin =0
 π 2
2 2π 
C2 = sin =0 2 5π 2
π 2 C5 = sin =
 π 2 5π
2 3π 2
C3 = sin =−
π 2 3π ···

That is 
to say, function x(t) can be synthesized as the sum of infinite sine series. Recall that
|Cn | = an2 + bn2 ; thus

∞
 ∞

x(t) = Cn exp (j nω 0 t) = a0 + an cos(nω 0 t) + bn sin(nω 0 t)
k=−∞ n=1
 
for even functions bn = 0 ∴ |Cn | = |an |

1  2
∞  nπ 
= + sin cos(n ω 0 t)
2 n=1 nπ 2
  
1 2 2π 1 2π 1 2π
= + cos t − cos 3 × t + cos 5 × t + ···
2 π T 3 T 5 T

As a comparison, this series is overlaid on x(t), see Fig. 20.8, where only the first four terms are
used, and spectrum is illustrated in Fig. 20.9 (left). Note that all even terms equal to zero.

Fig. 20.8 Example 20.2-1 x(t)

1

t
0
−a 0 a T
454 20 Discrete Fourier Transformation

(b) a = T /8: therefore,

2a 2
T 1
C0 = = =
T T 8
4 4
 
2 2nπ 2 2 2
nπ T 1 sin (nπ/4) 1
Cn = sin a = sin = = sinc (nπ/4)
nπ T 2 nπ T 8
4 2 nπ/4 2
∴
1
|Cn | = |sinc (nπ/4) |
2

We calculate a few samples,

π  √
2
C1 = sin =
2 
2 4π
π 4 π C4 = sin =0
 4π 4
2 2π 1 √
C2 = sin = 
2π 4 π 2 5π 2
√ C5 = sin =−
 5π 4 5π
2 3π 2
C3 = sin =
3π 4 3π ···

and spectrum is illustrated in Fig. 20.9 (right).

|C[n]| |C[n]|
1
1
2
4
n 0
0 n
0 5 0 5

Fig. 20.9 Example 20.2-1

2. Relative to square pulse waveform in Example 20.2-1, shapes of x(t) and y(t) waveforms are
same except for the time delay, that is to say y(t) = x(t − a). Fourier transform of time delay is
the exponential phase shift term; thus we write

1 2nπ
Cn = exp(−j nω 0 a) sin a
nπ T
Solutions 455

3. Given function x(t) is odd, by graph inspection, we write

2π 2π
T0 = 2 ∴ ω 0 = = =π
T0 2
T0 T0 1 1
x(t) = 2At x ∈ − , = − ,
4 4 2 2

Coefficients of Fourier series for an odd periodic function x(t) may be calculated by definition as
follows:
∞
 ∞

x(t) = Cn exp (j nω 0 t) = a0 + an cos (nω 0 t) + bn sin (nω 0 t)
k=−∞ n=1

where, in odd function all a0 = an = 0, while

; T0 ; T0 /2
2 4
bn = x(t) sin (nω 0 t) dt = x(t) sin (nω 0 t) dt
T0 0 T0 0
; T0 /4 ; 1/2
4 4
= x(t) sin (nω 0 t) dt = 2
A t sin (nπ t) dt
T0 −T0 /4 2
−1/2
; 1/2
= 4A t sin (nπ t) dt
−1/2

<
Integral in the form x sin(ax) dx is solved by partial integration as
⎧ ⎫
; ⎪
⎨u = x ∴ du = dx ⎪
⎬
x sin(ax) dx = 1
⎪ ∴ v = − cos(ax) ⎪
⎩dv = sin(ax) dx ⎭
a
;
1 1 1 1
= − x cos(ax) + cos(ax) dx = − x cos(ax) + 2 sin(ax)
a a a a
1
= sin(ax) − ax cos(ax)
a2

Therefore,
%1/2
1 %
bn = 4A sin(nπ t) − nπ t cos(nπ t) %
nπ 2 %
−1/2
 nπ   nπ   nπ   nπ  nπ   nπ
4A  
= sin + sin −  cos +  cos
(nπ )2 2 2 2 2 2 2
8A  nπ 
= sin
(nπ )2 2
456 20 Discrete Fourier Transformation

⎧
⎪
⎪ 0; n even
⎪
⎪
⎨ 8A
= (nπ )2 ; n = 1, 5, 9, . . .
⎪
⎪
⎪
⎪ 8A
⎩− ; n = 3, 7, 11, . . .
(nπ )2

Fourier series is therefore

∞
 8A 1 1
x(t) = bn sin (nπ t) = 2
sin (π t) − sin (3π t) + sin (5π t) + · · ·
n=1
π 9 25

As a comparison, this series is overlaid on x(t), see Fig. 20.10, where only the first three terms
are used, and amplitude spectrum is illustrated in Fig. 20.11 (left) and phase spectrum is in
Fig. 20.11 (right). Note that all even terms equal to zero.

Fig. 20.10 Example 20.2-1 x(t)

A
t
0

−A
−2 −1 0 1 2

8A
|Cn | = an2 + bn2 = bn2 = |bn | =
(nπ )2
⎧
  ⎪
⎪ 0; n even
−bn ⎨ π
θn = arctan = arctan(±∞) = − ; n = 1, 5, 9, . . .
a
>
0 ⎪
⎪ 2
n ⎩π ; n = 3, 7, 11, . . .
2

|C[n]| q [n]
1
A
0 n
n
0 −1
0 5 10 0 5 10

Fig. 20.11 Example 20.2-3

4. Given a linear function f (t) = 1 − t/π its Fourier series within the interval x ∈ [−π, π ] is
calculated as follows. By definition we write
Solutions 457

∞
  ∞
 
nπ t nπ t
f (t) = a0 + an cos + bn sin
n=1
T n=1
T

where
; T
1
a0 = f (t) dt
2T −T
; 
1 T nπ t
an = f (t) cos dt
T −T T
; T 
1 nπ t
bn = f (t) sin dt
T −T T

In addition, given interval it follows that T = π ; therefore

;  %π %
1 π
t 1 % 1 2 %%π 1 1  2 
a0 = 1− dt = t %% − t % = 2π − π − π2
2π −π π 2π −π 2π −π 2π 2π
=1
;   ; !; "
1 π n 1 π π
:0

an = 1−
t
cos
π t
dt = cos (n t) dt − 2
1

x cos 
(n t) dt
π −π π 
π π −π π −π 
⎧ ; a ⎫
⎪
⎨for odd functions: f (x) = x cos (n x) = −f (−x) ∴ f (x) dx = 0, ⎪
⎬
−a
⎪
⎩ ⎪
⎭
or, use partial fraction integration to prove it
%π
% :0

 :0
)
1 1
= sin(nt) %% = 
sin(nπ ) − sin(−nπ
nπ −π nπ
=0
;   ; π ; π
1 π
t n
π t 1  :0 1
bn = 1− sin dt = 
 (n t) dt
sin − 2 t sin (n t) dt
π −π π 
π π −π π −π
  
cos(nπ )−cos(nπ )=0
 
use partial fraction integration
π
1 1 1 1
=− − t cos(nt) + sin(nt) = π cos(nπ ) + π cos(nπ )
π2 n    −π π2 n
sin(nπ )=0

1 1
= 2
π cos(nπ )
π 2 n   
(−1)n

2
= (−1)n
nπ
458 20 Discrete Fourier Transformation

Therefore,

∞
2 2 2 2
f (t) = 1 + (−1)n sin(nt) = 1 − sin(t) + sin(2t) − sin(3t) + · · ·
n=1
nπ π 2π 3π

This series is superimposed to f (t), see Fig. 20.12, which illustrates the approximation. For
“perfect” approximation, there should be an infinite number of terms. The largest discrepancy
between the continuous function f (t) and its approximation is at the end of given interval. Over
wider interval Fourier series is a periodic function.

Fig. 20.12 Example 20.2-4 f (t)

2

1
t
0
−p 0 p

Exercise 20.3, page 438

1. Discrete Fourier transformation (DFT) of series x[n] is by definition


N −1 
2π k
Fk (x[n]) = x[n] exp −j where, k = 0, 1, 2, . . .
def
n
n=0
N

In order to simplify the long equations writing, often we use exp() to denote the necessary
exponential term in the above definition. As n, k, N changes, we find that the number of necessary
calculations increases as follows:

N = 1 F0 = x[0] exp() (one calculation)

N = 2 F0 = x[0] exp() + x[1] exp() (two calculations)
F1 = x[0] exp() + x[1] exp() (two calculations)
(total: four calculations)
N = 3 F0 = x[0] exp() + x[1] exp() + x[2] exp() (three calculations)
F1 = x[0] exp() + x[1] exp() + x[2] exp() (three calculations)
F2 = x[0] exp() + x[1] exp() + x[2] exp() (three calculations)
(total: nine calculations)
..
.

N (total: N 2 calculations)
Solutions 459

That is to say, for example, the resolution of 1 Hz within 20 kHz bandwidth means that it is
necessary to do N = 20 000 points in the transformation. Knowing that DFT calculation does
N 2 operations, we calculate

 2
t = N 2 × 1 min = 20 × 103 min = 400 × 106 min ≈ 278 × 103 days
≈ 760 years

By comparison, FFT algorithm does N log 2 N calculations, that is to3 say
ln N
t = (N log2 N) × 1 min = N × 1 min = 286 × 10 min ≈ 198 days
ln 2
The advantage of FFT over DFT increases as N becomes larger.
2. The main idea of FFT is to split the grand sum into two separate sums, in the first iteration to one
with odd and one with even coefficients. In the subsequent iteration, again every second term is
grouped as “odd” and the other half as “even” group, see Fig. 20.13. The process is repeated until
each group consists of two terms only. This process is highly symmetric; as a consequence the
binary form of aligned term indexes in the final iteration is binary inverted relative to the binary
form of aligned term indexes at the beginning.

Fig. 20.13 Example 20.3-2 bit inversion

[0 0 0] x0 x0 x0 [0 0 0]
[1 0 0] x4 x2 x1 [0 0 1]
[0 1 0] x2 x4 x2 [0 1 0]
[1 1 0] x6 x6 x3 [0 1 1]
[0 0 1] x1 x1 x4 [1 0 0]
[1 0 1] x5 x3 x5 [1 0 1]
[0 1 1] x3 x5 x6 [1 1 0]
[1 1 1] x7 x7 x7 [1 1 1]

N =2 N =4 N =8 ···

That is to say, we can write as follows.


N −1 
2π k
Fk (x[n]) = x[n] exp −j
def
n
n=0
N

N/2−1  
N/2−1 
2π k 2m 2π k 2m + 1
= x[2m] exp −j + x[2m + 1] exp −j
m=0
N/2 2 m=0
N/2 2
     
even index odd index
  
2π k
Ck = exp −j
N

N/2−1  
N/2−1 
2π k 2π k
= x[2m] exp −j m + Ck x[2m + 1] exp −j m
m=0
N/2 m=0
N/2
460 20 Discrete Fourier Transformation

where
  
2π k 2π k 2π k
exp −j m = cos − m + j sin − m ; (k = 0, 1, . . . , N)
N/2 N/2 N/2
However, when k > N/2, that is to say k = N/2 + r where r = 1, 2, . . . , N/2, we can write
 
2π (N/2 + r) −2π (N/2) m − 2π r m
cos − m = cos
N/2 N/2

2π r m
= cos −2π m −
N/2
 
m = 0, 1, . . . , N/2 ∴ −2π m = 0
 
2π r m 2π (N/2 + r) m
= cos − = cos −
N/2 N/2
 
known as “symmetry identity”

This identity is 
valid for sine terms aswell, which is to say that
2π k 2π (N/2 + k)
exp −j m = exp −j m ; (k = 0, 1, 2, . . . , N)
N/2 N/2
As a consequence, it is sufficient to carry out only N/2 calculations. However, each of the
summing groups can be split again and again; thus the total number of calculations is N log2 N.
(a) Given N = 2 and a general discrete series xn = x[n], we write Fourier transform
Fk (x[n]) = Xk equations using the following syntax

 −1   −1
N
2π
N
 
Xk = xn exp −j nk = xn wNn · k = N = 2
n=0
N n=0


1
= xn w2n · k
n=0

This syntax simplifies the repetitive writing of complex exponential terms; we keep in mind that
wNn · k represents the following calculations:
   n·k
2π 2π 2π
wN ≡ exp −j ∴ exp −j n k = exp −j = wNn · k
def

N N N

where the power term is calculated as the product nk, for example, if N = 2 it follows that
 0·0
2
π # $0 · 0
w20 · 0 = exp −j = exp (−j π ) = (−1)0 = 1
2

Solutions 461

or, =
 1·1
2
π # $1 · 1
w21 · 1 = exp −j = exp (−j π ) = (−1)1 = −1
2

For the moment, let us keep n > 0 exponential terms in place and also use the identity
w21 · 1 ≡ −w20 · 0 , even if their values are as simple as +1 or −1. Reason for this choice becomes
evident as N = 4, N = 8, etc. In this case,

k=0: X0 = x0 w20 · 0 + x1 w21 · 0 = x0 + x1 ≡ x0 + x1 w20

k=1: X1 = x0 w20 · 1 + x1 w21 · 1 = x0 − x1 ≡ x0 − x1 w20

This calculation process is illustrated by “butterfly diagram”, Fig. 20.14. The solutions Xk on the
right side are formed by the graph inspection. For example, x0 is first multiplied by “1” and then
added to x1 w20 at the summing node to produce X0 = x0 + x1 w20 . Similarly, the graphs illustrate
the formation of X1 equation. This basic butterfly diagram is reused for any N = 2 Fast Fourier
Transformation. As illustrated by the colour scheme, note paths that carry “ + 1 multiplication
factor, paths that carry positive exponential factor, and paths that carry negative exponential factor.

Fig. 20.14 x0 1 X0 = x0 + x1 e−j

2π·0
2

Example 20.3-2(a) 1

w20 = 1
x1 −w20 = −1 X1 = x0 − x1 e−j
2π·0
2

(b) In the case of N = 4, we reorganize the four Xk DFT transformation into symmetric form
that can be directly translated into butterfly diagram as follows.
As a reminder, we use shorthand syntax for N = 4 as
  π  π n k
2π
w4 ≡ exp −j = exp −j ∴ exp −j ≡ w4n · k
4 2 2

∴
 π 0
w40 = exp −j =1
2
 π 1
w41 = exp −j = −j
2
 π 2
w42 = exp −j = exp (−j π ) = −1
2
 π 3   π
3π
w43 = exp −j = exp −j = exp j =j
2 2 2
462 20 Discrete Fourier Transformation

so that we can write

k = 0 : X0 = x0 w40 · 0 + x1 w41 · 0 + x2 w42 · 0 + x3 w43 · 0

= x0 + x1 + x2 + x3 (“DC component”)
k = 1 : X1 = x0 w40 · 1 + x1 w41 · 1 + x2 w42 · 1 + x3 w43 · 1
 
3π π
= x0 + x1 w41 + x2 w42 + x3 w43 = ≡−
2 2
= x0 + x1 w41 − x2 − x3 w41
k = 2 : X2 = x0 w40 · 2 + x1 w41 · 2 + x2 w42 · 2 + x3 w43 · 2
= x0 + x1 w42 + x2 w44 + x3 w46
= x0 − x1 + x2 − x3
k = 3 : X3 = x0 w40 · 3 + x1 w41 · 3 + x2 w42 · 3 + x3 w43 · 3
 
9π 3π
= x0 + x1 w43 + x2 w46 + x3 w49 = ≡−
2 2
= x0 + x1 w41 − x2 − x3 w41

This set of equation may be written in a bit symmetric form as

X0 = (x0 + x2 ) + w40 (x1 + x3 )

X1 = (x0 − x2 ) + w41 (x1 − x3 )
X2 = (x0 + x2 ) + w42 (x1 + x3 )
X3 = (x0 − x2 ) + w43 (x1 − x3 )

which can be further structured to explicitly include the form of two-bit butterfly (i.e. N = 2) as

X0 = (x0 + w20 x2 ) + w40 (x1 + w20 x3 )

X1 = (x0 − w20 x2 ) + w41 (x1 − w20 x3 )
X2 = (x0 + w20 x2 ) − w40 (x1 + w20 x3 )
X3 = (x0 − w20 x2 ) − w41 (x1 − w20 x3 )
Solutions 463

Direct graph implementation of these equations is illustrated in Fig. 20.15, where starting from the
left side of the graph, and following xn terms along the associated paths (the addition operations
are done at the merging nodes), we compose the four Xk equations simply by inspection. Note
the index order of xn and Xk indexes in this graph structure.

bit inversion

[0 0] x0 X0 [0 0]

even
+w20
[1 0] x2 −w20 X1 [0 1]

+w40

[0 1] x1 −w40 X2 [1 0]

odd +w41
+w20
[1 1] x3 −w20 −w41 X3 [1 1]

Fig. 20.15 Example 20.3-2(b)

3. In order to use FFT graph diagram, in the first step we reorganize xn terms on the left side by
using the index bit inversion relative to Xk , see Fig. 20.16. The intermediate sums are calculated
using the two-bit grouping on the left side, and then the second group of addition/multiplication
operations is done by following paths on the right side of the graph.

bit inversion

[0 0] x0 = [1] 4 X0 = [10] [0 0]

+w20
[1 0] x2 = [3] −w20 −2 X1 = [−2 − 2j] [0 1]

+w40

[0 1] x1 = [2] 6 −w40 X2 = [−2] [1 0]

+w41
+w20
[1 1] x3 = [4] −w20 −2 −w41 X3 = [−2 + 2j] [1 1]

Fig. 20.16 Example 20.3-3

464 20 Discrete Fourier Transformation

4. (a) In the case of N = 8, we use shorthand syntax as

  π  π n k
2π
w8 ≡ exp −j = exp −j ∴ exp −j ≡ w8n · k
8 4 4

∴
 π 0
w80 = exp −j =1
4
 π 1 √
2
w81 = exp −j = (1 − j )
4 2
 π 2  π
w82 = exp −j = exp −j = −j
4 2
 π 3  √
3π 2
w83 = exp −j = exp −j =− (1 + j )
4 4 2
 π 4
w84 = exp −j = exp (−j π ) = −1 ≡ −w80
4
 π 5  √
5π 2
w85 = exp −j = exp −j =− (1 − j ) ≡ −w81
4 4 2
 π 6 
6π
w86 = exp −j = exp −j = j ≡ −w82
4 4
 π 7  √
7π 2
w87 = exp −j = exp −j = (1 + j ) ≡ −w83
4 4 2

With the help of Fig. 20.13, we derive FFT graph flow for N = 8 as in Fig. 20.17, where
horizontal red lines imply multiplication by “−1”, dark lines imply multiplication by “1”, and
other multiplication factors are written explicitly.
(b) Given xn = [1, 1, −1 − 1, 1, 1, −1 − 1], by inspection of graph in Fig. 20.17, where
horizontal red lines imply multiplication by “−1”, we write

X0 = x0 + x4 + x2 + x6 + x1 + x5 + x3 + x7 = 1 + 1 − 1 − 1 + 1 + 1 − 1 − 1
=0

X1 = x0 − x4 + (x2 − x6 )(−j ) + x1 − x5 + (x3 − x7 )(−j ) w82

= 1 − 1 − 1 + 1 + (1 − 1 + (−1 + 1)(−j ))w82

=0
Solutions 465


X2 = x0 + x4 + (x2 + x6 )(−1) + x1 + x5 + (x3 + x7 )(−1) (−j )

= 1 + 1 + 1 + 1 + (1 + 1 + 1 + 1)(−j )
= 4 − 4j
X3 = x0 − x4 + (x2 − x6 )(−w82 ) + x1 − x5 + (x3 − x7 )(−w82 )
= 1 − 1 + (−1 + 1)(−w82 ) + 1 − 1 + (−1 + 1)(−w82 )
=0
X4 = x0 + x4 + x2 + x6 + (x1 + x5 + x3 + x7 )(−1)
=1+1−1−1−1−1+1+1
=0

X5 = x0 − x4 + (x2 − x6 )(−j ) + x1 − x5 + (x3 − x7 )(−j ) (−w81 )

= 1 − 1 + (−1 + 1)(−j ) + 1 − 1 + (−1 + 1)(−j ) (−w81 )

=0

X6 = x0 + x4 + (x2 + x6 )(−1) + x1 + x5 + (x3 + x7 )(−1) (j )

= 1 + 1 + 1 + 1 + (1 + 1 + 1 + 1)(j )
= 4 + 4j

X7 = x0 + x4 + (x2 − x6 )(j ) + x1 − x5 + (x3 − x7 )(j ) (−w83 )

= 1 + 1 + (−1 + 1)(j ) + 1 − 1 + (−1 + 1)(j ) (−w83 )

=0

The total number of sum/product calculations is 8 log2 8 = 24 in comparison with DFT where
the total number of calculations equals 82 = 64.
466 20 Discrete Fourier Transformation

x0 X0

x4 X1

x2 X2

w82
x6 −w82 X3

x1 X4

w81

x5 −w81 X5

w82

x3 −w82 X6

w83
w82
x7 −w82 −w83 X7

Fig. 20.17 Example 20.3-4(a)

 References

[App17] W. Appel. Mathématiques pour la physique et les physiciens. Number 978–2–35141–339–5. H&K
Éditions, 5th edn. (2017)
[Cou20a] MIT Open Courseware. Multivariable Calculus (2020), https://ocw.mit.edu/courses/mathematics/18-
02sc-multivariable-calculus-fall-2010/. On-line July 2020
[Cou20b] MIT Open Courseware. Signals and Systems (2020), https://ocw.mit.edu/resources/res-6-007-signals-
and-systems-spring-2011. On-line July 2020
[Cou20c] MIT Open Courseware. Single Variable Calculus (2020), https://ocw.mit.edu/courses/mathematics/18-
01sc-single-variable-calculus-fall-2010/. On-line July 2020
[Dem63] B.P. Demidovic, Collection of Problems and Exercises for Mathematical Analysis (in Russian). Number
517.2 D30. National Printing House for Literature in Mathematics and Physics, 5th edn. (1963)
[Ink20] Inkscape. Drawings Generated by Inkscape (2020), https://inkscape.org/. On-line July 2020
[LT09] Y. Leroyer, P. Tessen, Mathématique pour l’ingénieur: exercises et problèmes. Number 978–2–10–
052186–9. Dunod, 1st edn. (2009)
[Mit67] D.S. Mitrinovic, Matematika I: u obliku metodicke zbirke zadataka sa rešenjima. Number 2043.
Gradjevinska Knjiga, 3rd edn. (1967)
[Mit77] D.S. Mitrinovic, Kompleksna Analiza. Gradjevinska Knjiga, 4th edn. (1977)
[Mit78] D.S. Mitrinovic, Matematika I: u obliku metodicke zbirke zadataka sa rešenjima. Number 06–1785/1,
Gradjevinska Knjiga, 5th edn. (1978)
[Mit79] D.S. Mitrinovic, Kompleksna analiza: zbornik zadataka i problema, vol. 3. Number 06–1431/1, Naucna
Knjiga, 2nd edn. (1979)
[Mod64] P.S. Modenov, Collection of problems in special program (in Russian). National Printing House for
Literature in Mathematics and Physics, (1964)
[Mü20] D. Müller, Mathématiques (2020), http://www.apprendre-en-ligne.net/MADIMU2/INDEX.HTM. On-
line July 2020
[Sob20] R. Sobot, Private Collection of Course Notes in Mathematics. (2020)
[Spi80] M.R. Spiegel, Analyse de Fourier et Application aux Problémes de Valeurs aux Limites. Number France
2–7042–1019–5. McGraw–Hill, 1st edn. (1980)
[Ste11a] Stewart, Analyse Concepts et Contextes, Vol.1: Fonctions d’une Variable. Number 978–2–8041–6306–8.
De Boeck Supérieur, 3rd edn. (2011)
[Ste11b] Stewart, Analyse Concepts et Contextes, Vol.2: Fonctions de plusieurs variables. Number 978–2–8041–
6327–3. De Boeck Supérieur, 3rd edn. (2011)
[Sym20] Symbolab, Online Calculator (2020), https://www.symbolab.com/. On-line July 2020
[Tik20] TikZ, Graphs Generated by TikZ (2020) , https://en.wikipedia.org/wiki/PGF/TikZ. On-line July 2020
[Ven96] T.B. Vene, Zbirka rešenih zadataka iz matematike, vol. 2. Number 86–17–09617–9. Zavod za Udžbenike
i Nastavna Sredstva, 22nd edn. (1996)
[Ven01] T.B. Vene, Zbirka rešenih zadataka iz matematike, vol. 1. Number 86–17–09031–6. Zavod za Udžbenike
i Nastavna Sredstva, 28th edn. (2001)
[Ven03] T.B. Vene, Zbirka rešenih zadataka iz matematike, vol. 3. Number 86–17–10786–3. Zavod za Udžbenike
i Nastavna Sredstva, 27th edn. (2003)

© Springer Nature Switzerland AG 2021 467

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0
468 References

[Ven04] T.B. Vene, Zbirka rešenih zadataka iz matematike, vol. 4. Number 86–17–11156–9. Zavod za Udžbenike
i Nastavna Sredstva, 36th edn. (2004)
[Wik20a] Wikipedia, Fourier Transform (2020), https://en.wikipedia.org/wiki/Fourier_transform. On-line July
2020
[Wik20b] Wikipedia, Linear Algebra (2020), https://en.wikipedia.org/wiki/Linear_algebra. On-line July 2020
 Index

A Column matrix, 129

abs function, 9–12 Common factors, 14–15, 23
Absolute numbers, 5, 9 Complex algebra, 105–127
Abstract power expressions, 8 Complex equations, 107, 121–124
Adjacent side, 74, 76 Complex functions
Algebra, fundamental theorems, 32 basic forms, 296
Analysis of function bode plots, 307
critical points, 230, 232–234 decibel scale, 295
domain of definition, 227 decibel unit, 295
function’s parity, 228 phase function, 297–302, 304–306
function’s sign, 228 piecewise linear approximation, 296
graphical representation, 231 second order function, 309–310
limits, 229 transfer function forms, 295–296
oblique asymptote, 229 z(x) functions, 296, 309
Analytical geometry, 132 Complex numbers, 105–106, 108–118
Arithmetic progression, 331 Complex plane, 105, 106, 121, 122
sum of first nterms in, 331 Continuous functions, 357
energy and power of, 360, 393
Fourier integral, 413–420
B transformations of, 357–370
Basic calculations, 3–5 Continuous sinusoidal functions, 418
Basic exponential function, 53 Continuous time (CT), 413
Basic number operations Convolution
absolute numbers, equations, and inequalities, 5, 9 basic properties, 394
basic calculations, 3–5 discrete functions, 394–395, 409
Binomial expansion, 19 Convolution functions, 359–360, 366–380, 383–387, 393
Binomial factor, 20, 21 Convolution integral, 357
Binomial square, 16, 28 continuous functions
form, 16, 26, 151 energy and power of, 360, 395, 409
Bi-quadratic form, 16 transformations of, 357
Bi-quadratic polynomial, 16, 27 convolution
continuous functions, 360
piecewise linear functions, 359
C Cosine law, 154
Cardinal sine function, 342, 428–434 Cramer’s rule, 43, 121, 130, 135, 164–165, 171, 173, 175
Cathetus adjacent, 74
Cathetus forming, 73
Cathetus opposite, 73 D
Chain rule derivatives, 203, 206 Decimal numbers, 6
Characteristic polynomial, 185, 312 Definite integrals, 250–252, 277–281

© Springer Nature Switzerland AG 2021 469

R. Sobot, Engineering Mathematics by Example,
https://doi.org/10.1007/978-3-030-79545-0
470 Index

Delta function, 341 exponential equations and inequalities, 52

properties of, 341 Fourier integral, 420–423
Derivatives identities, 51
chain rule, 203, 206
exponential functions, 207
L’Hôpital’s Rule, 204, 215–217 F
loga x functions, 205–207 Factorization, 6, 17, 26, 35, 56, 63
odd function, 212 Factor theorem, 13, 14, 20, 233, 236
product functions, 209–210 Fast Fourier transformation (FFT), 438–439, 461–464
product rule, 203 First order homogenous equations, 313, 316–318
radicals, 205 First order polynomial, 22
ratio functions, 210, 215 Fourier integral, 413, 414
rational function’s limit, 215 cardinal sine function, 428–434
ratio rule, 204 continuous functions, 415–418
rules of derivations, 202 continuous sinusoidal functions, 418
sin x within [–1, 1] interval, 212 exercises, 414–415, 434
systematic calculation, 213 exponential functions, 419–424
tabular derivatives port function, 425
exponent, 202 sign function, 427–428
logarithm, 202 step function, 428
power rule, 202, 203 triangle function, Fourier transform of, 425
trigonometric functions, 208–209, 211 Fourier series, 438, 452, 455
Tylor polynomials, 204, 211 Fourier sum, 437
a x functions, 205 Fourier transformation integral
Determinant, linear algebra, 130, 166–168 Fourier integral, 415, 418–428
Determinant of matrix, 130 cardinal sine function, 428–434
Differential equations, 315–317 continuous functions, 415–418
engineering examples, 313–314, 324–328 continuous sinusoidal functions, 418
first order homogenous equations, 313, 316–317 exercises, 414–415
linear equations, 313, 318–321 exponential functions, 419–424
with constant coefficients, 313, 321–324 port function, 425
separable variables, 311, 312 sign function, 427–428
Dirac comb function, 346 step function, 428
Dirac delta function, 345–346 triangle function, Fourier transform of, 425
Discrete convolution sum, 393 sums and series, 413
convolution Fractional powers, 7, 9, 116, 121, 205
basic properties, 394 Fractions, 6, 7
discrete functions, 394–395, 409 Function analysis
energy and power of discrete functions, 395, 409 analyse and plot graphs, 318
power, 410, 411 analysis of function
Discrete Fourier sum, 437–452 critical points, 230, 232–234
Discrete Fourier transformation domain of definition, 231
discrete Fourier sum, 437–452 function’s parity, 231
fast Fourier transformation, 438–439, 461 function’s sign, 232
Fourier series, 438, 455, 456 graphical representation, 231
Fourier sum, 439 limits, 229
Discrete functions, energy and power of, 395, 409 oblique asymptote, 229
Discrete time (DT), 413 analysis of functions, 220–221
Double integrals, multivariable function, 293 basic functions, 221
Double integration process, 273 arctangent, 225
argument mapping, 226–227
exponent, 223
E inverse, 221
Eigenvalue, linear algebra, 136, 179, 181 logarithm, 223
Eigenvector, linear algebra, 136, 179, 181 radicals, 222
Engineering examples, differential equations, 313–314, sine and cosine, 224
324–328 tangent, 224–225
Equivalent radicals, 8, 12 composite function, 220, 225–226
Euler’s equation, 113 decomposing mathematical forms, 226
Exponential function, 52–54, 343 logarithmic function, 227
Index 471

rational function sine and cos functions, 275

critical points, 230, 232–234 squares of sine/cosine functions, 257
domain of definition, 227, 237 tabular integral, 248, 265–266
function’s parity, 228, 237–238 trigonometric functions, 257
function’s sign, 228 trigonometric substitutions, 250, 263–265
function’s zero, 238 Inverse function, 279–280
graphical representation, 231, 234 Inverse matrix, linear algebra, 136, 183
limits, 229, 232, 238
oblique asymptote, 229, 240
summary of variations, 239 L
results of analysis, 234–237 Leading coefficient, 26, 28–30, 33
typical function forms, 240–246 L’Hôpital’s Rule, 204, 213–215
Limits
asymptotes, 188
G basic rules, 187
Gaussian integral form, 260–261 continuous function, 189
Geometrical interpretation, 161 famous limits, 187
Geometric progression, 331, 337 forms, 194
sum of, 331–332 infinite limits, 187
Geometric series, 333 multivariable functions, 283–287
Geometric sum, convergence of, 332 non-continuous function, 189
Graphical method, 379 non-one numerator, 194
Grouping binomials, 23, 24 oblique asymptote, 199
pole–zero cancellation, 192, 193, 195, 198
rational forms, 194
H rational function, 191–193
Homogeneous linear differential equations, 312, Sandwich theorem, 191
316 shape of the function, 196
Hypotenuse side, 73–76, 79 sin function, 193
undefined functions, 188
vertical asymptotes, 195–199
I Linear algebra
Identity matrix, 130 analytical geometry, 132, 147–154
Improper integral, 252 determinants, 130, 166–168
Impulse function, 342 eigenvalue and eigenvector, 136, 179, 181
Inequalities, 5, 9, 10 inverse matrix, 136, 183
Integrals linear transformation, 133, 156, 157, 160, 162
basic functions, 248 matrix powers, 137, 186
calculus method, 271 matrix product, 136, 176
change of variables, 253, 267, 272, 276 matrix-vector product, 135, 175, 176
circular form, 256 systems of linear equations, 134–135, 169–170
cutting out operation, 269 vector definitions, 131–132, 143–147
definite, 248–250, 272–274 vector operations, 132–133, 154–156
double integration process, 273 Linear combination, 163, 164
Gaussian integral, 249 Linear equations, 43–50, 313, 318–321
Gaussian integral form, 260–261 with constant coefficients, 313, 321–324
idea of integration by parts method, 254–256 systems of, 134–135, 169–175
improper integral, 252 Linear inequalities, 44, 46–50
integration Linear operation, 156, 157
by parts, 248 Linear transformation, 133, 156–166
by substitution, 248 Logarithmic functions, 52, 60–61
integration method, 248 identities, 51
integration techniques, 248 logarithmic equations and inequalities, 53, 61–66
inverse function, 279–280 Long division, 16–17, 30–35
list of basic integrals, 253
Newton–Leibniz formula, 252
partial integration technique, 268 M
periodic functions, 274–275, 280–281 Magnitudes, 140
piecewise linear function, 249, 257–258, 274 Matrix–matrix multiplication, 177
rational function, 249, 261–262, 269–270 Matrix powers, 137, 184–186
472 Index

Matrix product, 136, 176 Q

Matrix transformation method, 174 Quadratic equation, 28, 56–58, 63, 121–123, 170, 245
Matrix-vector product, 135, 175, 179
Multiple factorization techniques, 17, 35–37
Multiplication, identity, 13–14, 18–20 R
Multivariable functions Radicals, 3–5, 7–9, 12, 121, 177, 193, 205, 222
arbitrary directions, 288 Ramp function, 347
composite and parametric functions, 291 Rational expression, 4, 37
domains of, 284, 287 Rational function
double integral, 293 critical points, 230, 239
integrals, 284, 285 domain of definition, 225, 237–238
limits, 283, 285, 287–288 function’s parity, 228, 231
partial derivatives, 283–284 function’s sign, 228
two-sided conus function, 287 function’s zero, 238
two-variable functions, 286–287 graphical representation, 231, 234
limits, 229, 238, 240
oblique asymptote, 229, 240
N summary of variations, 239
Negative powers, 5, 6, 8, 12 Rational powers, 7, 8
Nested radicals, 9 Raw matrix, 129
Newton–Leibniz formula, 252 Rectangular function, 348
Number theory, 3–12 Right-angled triangles, 72–74, 76–79

O S
Oblique asymptote, 199 Sandwich theorem, 191
One-dimensional space, 137 Separable variables, 312, 314–315
Opposite side, 73, 74, 76 Series, 331–332
Orthogonality, 311 basic sums, 332–337
geometric series, 333, 337
limits of sums, 333, 337
P Sign function, 342, 347
Partial fractions, 17, 37–42 Fourier integral, 427–428
Pascal triangle, 20, 381 Similar triangles, 68, 73, 78
Perfect square, 16, 28–30 Sine law, 154
Piecewise linear function, 249, 257–258, 274, Special functions, 341–354
359–366 cardinal sine function, 349
Pole–zero cancellation, 192, 193, 195, 198 dirac comb function, 346
Polynomial dirac delta function, 345–346
basic polynomial transformations, 13 exponential function, 343
binomial square, 16, 28 interrelations, 342, 350–355
bi-quadratic form, 16, 27–28 ramp function, 347–348
common factors, 14–15, 23–24 rectangular function, 348
factor theorem, 13, 14, 20–21 sign function, 347
long division, 16–17, 30–35 step function, 347
multiple factorization techniques, 17, 35–37 triangle function, 349
multiplication, identity, 13–14, 18–20 Squares identity, 15, 24, 25
partial fractions, 17, 37–42 Step function, 347, 352, 428
perfect square, 16, 28–30 Sums, 331–337
polynomial identities, 15–16, 24–25 of cubes, 17, 34
sum of cubes, 17, 35 limits of, 333, 337
Viète’s formulas, 16, 26–27 Systems of linear equations, 44, 46
Polynomial identities, 15–16, 24–25
Port function, Fourier integral, 425
Positive real number, 9 T
Pythagoras’s theorem, 67, 68, 72, 74, 75, 79, 88, 108, Tabular derivatives
141–143, 146, 150 exponent, 202
Pythagorean triangle, 105 logarithm, 202
power rule, 202, 203
Index 473

Tabular integrals, 247, 265–268 square to double angle identities, 68

“Telescopic” forms of equations, 45 sum identities, 70, 85–87
Third order determinants, 168–169 sum to product identities, 67
Third order polynomial, 19, 198, 233, 381 trigonometric equations, 71, 97
Transformations, of continuous functions, 357–366 trigonometric inequalities, 71, 101, 102
Transpose of a column vector, 129 trigonometry identities, 70, 81–85
Transpose of a row vector, 129 Trigonometry identities, 70, 81–85
Triangle function, 349 Tylor polynomial, 204, 211, 245
of Fourier integral, 425–428
Trigonometric equations, 71, 97
Trigonometric functions, 208–209, 211, 257 V
Trigonometric inequalities, 71, 101–102 Variable substitution technique, 46
Trigonometry Vector addition, 105, 132, 154–155
adjacent side, 76, 79 Vector equations, 147
hypotenuse side, 73–76, 79 Vector operations, 132–133, 156–157
opposite side, 75, 76, 79 Vertical asymptotes, 195
product to sum identities, 67–68, 71, 91–93 Viète’s formulas, 16, 26–27