NUMERICAL METHODS

.FINDING THE ROOTS. Newton – Raphson Method

 It uses the tangent line of the curve to find the root.
Bisec�on Method
 Approxima�on method of finding the roots by repeatedly
bisec�ng the interval.

CALTECH:

𝑓𝑓(𝑥𝑥)
𝑥𝑥 = 𝑥𝑥 −
CALTECH: 𝑑𝑑
𝑓𝑓(𝑥𝑥)|𝑥𝑥=𝑥𝑥0
𝑑𝑑𝑑𝑑
MODE 7: TABLE
CALC 𝑥𝑥 = 𝑥𝑥0
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆: 𝑎𝑎 → 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

𝐸𝐸𝐸𝐸𝐸𝐸: 𝑏𝑏 → 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = → 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 𝑛𝑛𝑡𝑡ℎ 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆: (𝐸𝐸𝐸𝐸𝐸𝐸 − 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆) ÷ 2𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

.INTEGRATION BY APPROXIMATION.
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 (𝐸𝐸𝑇𝑇 ) = 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 − 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉
𝐸𝐸𝑇𝑇
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 (|𝐸𝐸𝑇𝑇 |) = � �
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉

Trapezoidal Rule
 Solving Area under a curve by dividing it into litle trapezoids.
False Posi�on Method
 It is a modifica�on of the bisec�on method.
 Linear interpola�on is used instead of the bisec�on of the
interval.

𝑥𝑥𝑛𝑛
𝑏𝑏
ℎ
𝐴𝐴 = � 𝑓𝑓(𝑥𝑥) ≈ � � �𝑓𝑓(𝑥𝑥) + 𝑓𝑓(𝑥𝑥 + ℎ)��
𝑎𝑎 2
𝑥𝑥=𝑥𝑥1
𝑏𝑏 − 𝑎𝑎
where: ℎ =
CALTECH: 𝑛𝑛
CALTECH:
𝐵𝐵 − 𝐴𝐴 𝑓𝑓(𝑎𝑎)
𝐶𝐶 = 𝐴𝐴 − ∶ ℎ
𝑓𝑓(𝑏𝑏) 𝑓𝑓(𝑐𝑐) 𝐴𝐴 = 𝐴𝐴 + �𝑓𝑓(𝑥𝑥) + 𝑓𝑓(𝑥𝑥 + ℎ)�: 𝑋𝑋 = 𝑋𝑋 + ℎ
−1 2
𝑓𝑓(𝑎𝑎)
Example:
 NUMERICAL METHODS
Simpson’s 1/3 Rule .SEQUENCE & SERIES.
 Unlike Trapezoidal Rule which makes us first order polynomial,
SEQUENCE
this method uses a second order polynomial approxima�on
 is a set or collec�on of numbers arranged in an orderly manner
method.
such that the preceding and the following numbers are
completely specified.
 The numbers in the given sequence are called elements. An
element is some�mes called term.

𝑏𝑏
𝑥𝑥𝑛𝑛 Infinite Sequence
ℎ
𝐴𝐴 = � 𝑓𝑓(𝑥𝑥) ≅ � � �𝑓𝑓(𝑥𝑥) + 4𝑓𝑓(𝑥𝑥 + ℎ) + 𝑓𝑓(𝑥𝑥 + 2ℎ)��  is a func�on whose domain is the set of posi�ve integers.
𝑎𝑎 3
𝑥𝑥=𝑥𝑥1

CALTECH: Finite Sequence

 If the domain of the func�on consists of the first n posi�ve
ℎ integers only
𝐴𝐴 = 𝐴𝐴 + �𝑓𝑓(𝑥𝑥) + 4𝑓𝑓(𝑥𝑥 + ℎ) + 𝑓𝑓(𝑥𝑥 + 2ℎ)�: 𝑋𝑋 = 𝑋𝑋 + 2ℎ
3
1, 3, 5, 7, 9 → Finite sequence
Example:
1, 3, 5, 7, 9, … → Infinite sequence

SERIES
 is the sum of the terms in a sequence.

Alterna�ng Series
 has posi�ve and nega�ve terms arranged alternately.
.INTERPOLATION:. Convergent Series
NEWTON’S DIVIDED DIFFERENCE.  If an infinite series has a finite sum.

 is a type of es�ma�on, a method of construc�ng (finding) Divergent Series

new data points based on the range of a discrete set of known  if it has no sum at all.
data points

CALTECH:

STAT 3 2 → 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿
STAT 3 3 → 𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄𝑄
3RD ORDER OR HIGHER: SATISFY THE CHOICES:
 NUMERICAL METHODS
.SEQUENCE and PROGRESSION. Infinite Geometric Progression
 is a kind of sequence where each term can be calculated using  is geometric progression only that the number of terms (n) is
a formula. extremely large or infinity.

Arithme�c Progression (AP)

Same Caltech:
 its succeeding terms have a common difference.
 The sum of all the terms in arithme�c progression is called as
Arithme�c Series.
Harmonic Progression (HP)
 A sequence of numbers whose reciprocals form an arithme�c
progression.

CALTECH:
STAT 3 2 → 𝐴𝐴 + 𝐵𝐵𝐵𝐵
𝑋𝑋 (# 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡) 𝑌𝑌 (𝑛𝑛𝑡𝑡ℎ 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡)

𝑛𝑛𝑡𝑡ℎ 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 = 𝑛𝑛𝑦𝑦�

𝑆𝑆 = ∑(𝑥𝑥𝑦𝑦�, 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠, 𝑒𝑒𝑒𝑒𝑒𝑒)

Fibonacci Sequence
 Named a�er the Italian merchant and mathema�cian,
Leonardo di Pisa or Fibonacci (Figlio dei Bonacci, “Son of the
Geometric Progression (GP) Bonnaccis”).
 its succeeding terms have a common ra�o.
 Each number is equal to the sum of the two preceding
 The sum of all the terms in arithme�c progression is called as
numbers.
Geometric Series.
 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144…
1 + √5
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 Φ =
2

Lucas Sequence
 Named a�er Edouard Lucas (1841 – 1891). Like the Fibonacci
numbers, every term of the Lucas sequence is the sum of the
two preceding numbers.
 1, 3, 4, 7, 11, 18, 29, 47, 76, 123…

CALTECH:
CALTECH:
1 + √5 1 − √5
STAT 3 6 → 𝐴𝐴𝐵𝐵 𝑋𝑋 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆: 𝐴𝐴 = 𝐵𝐵 =
2 2
𝑋𝑋 (# 𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡) 𝑌𝑌 (𝑛𝑛𝑡𝑡ℎ 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡)

𝐴𝐴𝑛𝑛 + 𝐵𝐵𝑛𝑛
𝑛𝑛𝑡𝑡ℎ 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 = 𝑛𝑛𝑦𝑦� 𝐹𝐹𝑛𝑛 = 𝐿𝐿𝑛𝑛 = 𝐴𝐴𝑛𝑛 + 𝐵𝐵𝑛𝑛
𝐴𝐴 − 𝐵𝐵
𝑆𝑆 = ∑(𝑥𝑥𝑦𝑦�, 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠, 𝑒𝑒𝑒𝑒𝑒𝑒)
 NUMERICAL METHODS
.SERIES.
Power Series
 Mother of All Series
 infinite sum of terms
∞

𝑓𝑓(𝑥𝑥) = � 𝑎𝑎𝑛𝑛 (𝑥𝑥 − 𝑥𝑥0 )𝑛𝑛

𝑛𝑛=0

𝑎𝑎𝑛𝑛
𝑅𝑅𝑅𝑅𝑅𝑅 = lim � �
𝑛𝑛→∞ 𝑎𝑎𝑛𝑛 + 1

CALTECH
∞

𝑓𝑓(𝑥𝑥) ÷ � 𝑎𝑎𝑛𝑛 (𝑥𝑥 − 𝑥𝑥0 )𝑛𝑛 = 1 @ 𝑥𝑥 = 𝑥𝑥0 + ∆𝑥𝑥

𝑛𝑛=0

Taylor Series Convergence Test using Ra�o Test

NOTE: do not include terms with x
 infinite sum of differen�al func�on at 𝑥𝑥 = 𝑥𝑥0
𝐺𝐺(𝑛𝑛 + 1) 𝐿𝐿 > 1 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
∞
𝑓𝑓 𝑛𝑛 (𝑥𝑥0 ) lim �
𝐺𝐺(𝑛𝑛)
� = �𝐿𝐿 < 1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
𝑓𝑓(𝑥𝑥) = � (𝑥𝑥 − 𝑥𝑥0 )𝑛𝑛 𝑛𝑛→∞
𝐿𝐿 = 1 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹
𝑛𝑛!
𝑛𝑛=0
𝑖𝑖𝑖𝑖: 𝐿𝐿 = 1 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹
𝐺𝐺 (𝑛𝑛 + 1)
−1 𝐿𝐿 > 1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
lim 𝑛𝑛 �� � − 1� = � 𝐿𝐿 < 1 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝑛𝑛→∞ 𝐺𝐺 (𝑛𝑛)
𝐿𝐿 = 1 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹
CALTECH
∞
𝑓𝑓 𝑛𝑛 (𝑥𝑥0 )
𝑓𝑓(𝑥𝑥) ÷ � (𝑥𝑥 − 𝑥𝑥0 )𝑛𝑛 = 1 @ 𝑥𝑥 = 𝑥𝑥0 + ∆𝑥𝑥 Radius of Convergence (ROC)
𝑛𝑛!
𝑛𝑛=0

1
𝑅𝑅𝑅𝑅𝑅𝑅 =
𝐿𝐿

Maclaurin Series Interval of Convergence

 is a Taylor series centered at the point 𝑥𝑥 = 0  Values of x that make the func�on converges.

∞
𝑓𝑓 𝑛𝑛 (0) NOTE: include terms with x (from the choices)
𝑓𝑓(𝑥𝑥) = � (𝑥𝑥)𝑛𝑛 𝐿𝐿 > 1 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝑛𝑛! 𝐺𝐺(𝑛𝑛 + 1)
𝑛𝑛=0 lim � � = �𝐿𝐿 < 1 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
𝑛𝑛→∞ 𝐺𝐺(𝑛𝑛)
𝐿𝐿 = 1 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹

CALTECH Divergence Test

∞
𝑓𝑓 𝑛𝑛 (0) NOTE: do not include terms with x
𝑓𝑓(𝑥𝑥) ÷ � (𝑥𝑥)𝑛𝑛 = 1 @𝑥𝑥 = ∆𝑥𝑥
𝑛𝑛! 𝐿𝐿 ≠ 0 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝑛𝑛=0 lim 𝐺𝐺(𝑛𝑛) = 𝐿𝐿 �
𝑛𝑛→∞ 𝐿𝐿 = 0 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹
𝑖𝑖𝑖𝑖: 𝐿𝐿 = 0 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹
Use Ra�o Test
 NUMERICAL METHODS
Fourier Series
 Expressing a fxn or signal as an infinite
sum of sine and cosine.

𝑇𝑇 = 2𝐿𝐿 = 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃

∞
𝑎𝑎0 𝑛𝑛𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝑛𝑛
𝑓𝑓(𝑡𝑡) = + � �𝑎𝑎𝑛𝑛 cos � � + 𝑏𝑏𝑛𝑛 sin � ��
2 𝐿𝐿 𝐿𝐿
𝑛𝑛=1

→ 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏
𝑎𝑎0 → 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴/𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀/𝐷𝐷𝐷𝐷 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑎𝑎 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
→ 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑠𝑠ℎ𝑖𝑖𝑖𝑖𝑖𝑖 𝑜𝑜𝑜𝑜 𝑎𝑎 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

𝑎𝑎𝑛𝑛 , 𝑏𝑏𝑛𝑛 → 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐/𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤ℎ𝑡𝑡𝑡𝑡 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

1 𝐿𝐿
𝑎𝑎0 = � 𝑓𝑓(𝑡𝑡) ∙ 𝑑𝑑𝑑𝑑
𝐿𝐿 −𝐿𝐿

1 𝐿𝐿 𝑛𝑛𝑛𝑛𝑛𝑛
𝑎𝑎𝑛𝑛 = � 𝑓𝑓(𝑡𝑡) cos � � ∙ 𝑑𝑑𝑑𝑑
𝐿𝐿 −𝐿𝐿 𝐿𝐿

1 𝐿𝐿 𝑛𝑛𝑛𝑛𝑛𝑛
𝑏𝑏𝑛𝑛 = � 𝑓𝑓(𝑡𝑡) sin � � ∙ 𝑑𝑑𝑑𝑑
𝐿𝐿 −𝐿𝐿 𝐿𝐿

Note: 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 × 𝑜𝑜𝑜𝑜𝑜𝑜 = 𝑜𝑜𝑜𝑜𝑜𝑜

𝐿𝐿
� 𝑜𝑜𝑜𝑜𝑜𝑜 = 0
−𝐿𝐿