AMERICAN UNIVERSITY OF BEIRUT

Mathematics Department-FAS

MATH 251
FINAL EXAM
Closed Book, 2hours 15mn

STUDENT NAME

ID NUMBER

Instructor’s Name - Lecture time

Problem Out of Grade

1 20

2 15

3 25

4 25

5 15

TOTAL 100

1
Exercise 1 (20 points)
Consider the following Initial value Problem:
 dy
= et + y ; t ∈ [0, 0.50]
(IV P ) dt
y(0) = 1

1. (6 points)(IVP) is first solved on [0 , 0.250] using 2 steps of the discrete

scheme of the 1st order Runge Kutta method (RK1), also called
Euler’ s method.

• Write the formulae of this scheme:


................................................................
(RK1)
yi+1 = ....................................................

• Use 2 steps ONLY of this scheme to approximate y(0.125) and

y(0.250) with 5 significant digits and rounding to the closest.
i ti yi k1
0 0 1 .
1 . . .
2 . . XXX

2. (4 points) (IVP) is then solved on [0.250 ,0.375] using 2 steps of the 4th
order discrete Runge Kutta scheme given by:


 k1 = f (ti , yi )
 k2 = f (ti + h2 , yi + h2 k1 )


(RK4) k3 = f (ti + h2 , yi + h2 k2 )
k = f (ti + h, yi + hk3 ).

 4



yi+1 = yi + h6 (k1 + 2k2 + 2k3 + k4 )

Use 2 steps ONLY of this scheme to approximate the value of y(0.375)

and y(0.5) with 5 significant digits and rounding to the closest.

2
 i ti yi k1 k2 k3 k4
2 0.250 . . . . .
3 . . . . . .
4 . . XX XXX XXX XXX

Exercise 2 (15 points)

Consider the definite integral
Z b=xn
I= f (x) dx
a=x0

where f ∈ C 4 [a, b].

In the case where n = 3m with m ∈ N , the integral I can be decomposed
into simple integrals Ik of the form:
Z x3k+3
Ik = f (x) dx , for k = 0, 1, ..., m − 1
x3k

The ” 38 - Simpson’s ” Simple integration formula that approximates Ik is such

that:
3
Ik ≈ Sk = [f (x3k ) + 3f (x3k+1 ) + 3f (x3k+2 ) + f (x3k+3 )] h
8
1. (5 points)Establish the Composite ” 38 - Simpson’s ” rule S(h) that
approximates
P I.
N.B. I = m−1
k=0 Ik

3
2. (5 points)Given that the local error in this approximation is
3 (4)
Ek = Ik − Sk = − f (ck )h5 , x3k < ck < x3k+3
80
prove that the global error if I ≈ S(h) is:

(b − a) (4)
E=− f (c)h4 , a < c < b
80
Justify all steps.

R4
3. (5 points) Let I = 2 x2 e−x dx
Determine the minimum number of partition points n = 3m needed to
approximate I by means of the Composite ” 83 - Simpson’s ” rule
S(h), in order for the absolute error in this approximation to verify:
1
| I − S(h)| ≤ 10−3
2

4
Exercise 3 (25 points)

Let f be a differentiable function up to any order, i.e. f ∈ C ∞ .

1. (2 points) The Central difference formula that approximates the

Second order derivative f ”(x), is given by:

ψh (x) = 2[x − h, x, x + h]

where [x − h, x, x + h] is a second order divided difference.

Show that ψh (x) = f (x−h)−2fh(x)+f
2
(x+h)

2. (5 points) Derive the expression of the infinite error series

(h) = c1 hα1 + c2 hα2 + c3 hα3 + ....

given that: f ”(x) = ψh (x) + (h), by determining the values of the

exponents {α1 , α2 , α3 , ...}.

5
3. (5 points) For the purpose of approximating the second derivative,
f ”(x) based on the Central Difference approximation formula, Richard-
son Extrapolation procedure can be carried without necessarily divid-
ing h by 2, but more generally by a factor q > 1.
In this context, let q = 3.
Derive in that case Richardson extrapolation operator of first and sec-
(1) (2)
ond order, respectively ψh and ψh , and determine the order in h of
their respective errors 1 (h) and 2 (h).
(1)
• ψh and 1 (h)

(2)
• ψh and 2 (h)

6
4. (5 points) For the purpose of approximating the second derivative,
f ”(0.9), consider the following table

D18 = {(xi , yi )|i = 0, 1, ..., 18}

associated with a function f (x), i.e., yi = f (xi ), i = 0, ..., 18.

i xi yi = f (xi )
0 0 1
1 0.1 0.9900
2 0.2 0.9608
3 0.3 0.9139
4 0.4 0.8521
5 0.5 0.7788
6 0.6 0.6977
7 0.7 0.6126
8 0.8 0.5273
9 0.9 0.4449
10 1 0.3679
11 1.1 0.2982
12 1.2 0.2369
13 1.3 0.1845
14 1.4 0.1409
15 1.5 0.1054
16 1.6 0.0773
17 1.7 0.0556
18 1.8 0.0392

Fill in the empty slots in the following table adequately, starting with
h0 = 0.9. Express all the results obtained with 4 significant digits and
rounding to the closest.

7
 h φ(h) φ(1) (h) φ(2) (h)
h0 = 0.9 XXXXX XXXXX

. XXXXX

Best approximation to f ”(0.9) :

8
Exercise 4 (25 points)
Based on the set of data

Dn = {(xi , yi ) for i = 0, 1, · · · , n|x0 < x1 < x2 < · · · < xn and yi = f (xi )},

with:
hi+1 = xi+1 − xi , ∀ i = 0, 1, · · · , n − 1
A cubic spline S(x), is determined by Dn in addition to the 2 sets of unknown:

{w0 , w1 , ..., wn }

{z0 , z1 , ..., zn }
00 0
where wi = S (xi ) and zi = S (xi ), i = 0, 1, ..., n.

1. (5 points) Derive the equation of the cubic spline S(x) by finding Si (x)
for x ∈ [xi , xi+1 ].

9
 ’

2. (5 points) Derive the n − 1 equations in the n + 1 unknowns:

{wi | i = 0, 1, 2, ..., n − 1, n}

.................................................. = (hi + hi+1 )[xi−1 , xi , xi+1 ].

10
3. (5 points) Based on the results obtained in (1), that relate the {zi } and
the {wi }, ∀ i = 0, ..., n, prove that for i = n − 1:
hn hn
zn−1 = zn − (wn + wn−1 ) = [xn−1 , xn ] − (wn + 2wn−1 )
2 6

4. (5 points) To determine the equations of a Cubic Spline, with the 2

additional imposed mixed conditions respectively on the slope and the
curvature:
zn = [xn−1 , xn ] and w0 = 2[x0 , x1 , x2 ].
Using the equations obtained in 2 and 3, show first that :

(i) (h1 +h
3
2)
w1 + h62 w2 = ( 23 h1 + h2 )[x0 , x1 , x2 ]
(ii) wn−1 + 2wn = 0
then show that these results lead to the n × n ”almost tridiagonal”
system
Aw = r
given by:

11
(iii)
  
h1 + h2 h2 w1
0 ··· 0
 3 6  . 
 h2 h2 + h3 h3 
.

··· 0
  
6 3 6
  

hn−1
 . 
 .. .. .. ..  
 . . . .  . =
 6  
 hn−1 hn−1 + hn hn
 . 
 0 ... 0 6

.

6 3
  
  
  . 
0 .... 0 1 2 wn
 
( 32 h1 + h2 )[x0 , x1 , x2 ]

 (h2 + h3 )[x1 , x2 , x3 ] 

 .. 

 . 

 (hn−2 + hn−1 )[xn−3 , xn−2 , xn−1 ] 
 
 (hn−1 + hn )[xn−2 , xn−1 , xn ] 
0
(i)

(ii)

12
 (iii)

5. (5 points) Let D3 = {(0, 1), (1, 2), (3/2, 3), (2, 5)} with
z3 = [x2 , x3 ] ; w0 = 2[x0 , x1 , x2 ]
Find only the corresponding vector w = [w1 , w2 , w3 ] as determined in
(4), based on the table of divided differences below..

i xi y i [xi , xi+1 ] [xi , xi+1 , xi+2 ]

0 0 1 × ×
1 1 2 1 ×
2 3/2 3 2 3/2
3 2 5 4 2

• w=

13
Exercise 5 (15 points)

Let p(x) = c2 x2 + c1 x + c0 be a quadratic plynomial with one of its roots ”r”,

located in an interval (a, b), with:

min |p0 (x)| ≥ d > 0

a≤x≤b

Using Newton’s method with r0 sufficiently close to r:

1. (5 points)Show that if rn ∈ (a, b), then |rn+1 − r| ≤ C|rn − r|2 , where

C = |cd2 |

2. (5 points)Let en = C|r − rn |. Show that if rn ∈ (a, b), then en+1 ≤ e2n .

Give also the condition on |r0 − r| that makes e0 < 1.

14
 .

1
3. (5 points)Assume that |r0 − r| = 2C
. Show by recurrence that
n
en ≤ (e0 )2 < 1, ∀ n

then consequently, estimate the smallest value N of n for which:

|rN − r|
≤ 2−p
|r0 − r|

where p is some positive integer.

15