AMERICAN UNIVERSITY OF BEIRUT

Mathematics Department-FAS

MATH 251
FINAL EXAM
Fall 2013-2014
Closed Book, 2hours

STUDENT NAME

ID NUMBER

Problem Out of Grade

1 20

2 20

3 20

4 25

5 15

TOTAL 100

1
1. (20 points) Consider the following Initial value Problem:
 dy
= 2et − y ; t ∈ [1.00, 2.00]
(IV P ) dt
y(1) = 0

To solve (IVP)

(a) (5 pts)(IVP) is first solved on [1.00 , 1.25] using 1 step of the

discrete scheme of the first order Runge Kutta method -(Euler’s
method) : (RK1).
• Write the formulae of this scheme:


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

• Use 1 step ONLY of this scheme to approximate y(1.25)

with 5 significant digits and rounding to the closest.

i ti yi k1 yi+1
0 1 . . .

(b) (7 pts) (IVP) is then solved on [1.25 , 1.75] using 2 steps of the
discrete scheme of the 2nd order Runge Kutta method - Trape-
zoid Rule (Heun’s method): (RK2.T).
• Write the formulae of this scheme:


 ...............................................................
(RK2.T ) ................................................................
yi+1 = ....................................................


• (7 pts)Use 2 steps ONLY of this scheme to approximate

y(1.50) and y(1.75) with 5 significant digits and rounding to

2
 the closest..

i ti yi k1 k2 yi+1
1 1.25 . . . .
2 . . . . .

(c) (5 pts) (IVP) is then solved on [1.75 ,2.00] using 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 1 step ONLY of this scheme to approximate y(2.00) with 5

significant digits and rounding to the closest.

i ti yi k1 k2 k3 k4 yi+1
3 1.75 . . . .

(d) (1pt) The discrete solution of (IVP) is

Y4 = {y0 = 0, ......................................................................................................}

3
2. (20 points) Consider the following set of data:
Dn = {(xi , yi )|i = 0, ..., n where yi = f (xi ) , and xi+1 −xi = h, 0 < h ≤ 1}
(a) (7 points) Write first the Central difference formula
ψh ≡ ψh (f (xi )) that approximates the first derivative f 0 (xi ), ∀i =
1, 2, ..., n − 1. Assuming then that f is differentiable up to any or-
der, derive the expression of the infinite error series
(h) = c1 hα1 + c2 hα2 + c3 hα3 + ....
given that: f 0 (xi ) = ψh + (h), by determining the values of the
exponents {α1 , α2 , α3 , ...}.

• ψh ≡ ψh (f (xi )) =

• Derive the Error series (h) :

(h) =

4
(b) (6 points) Based on the Central difference formula, derive Richard-
son extrapolation operators of orders 1 and 2 and the order of their
error series.
• 1st order Richardson extrapolation operator ψh1 (f (xi )) :

ψh1 ≡ ψh1 (f (xi )) =

Corresponding Error = O(...........)

• 2nd order Richardson extrapolation operator ψh2 (f (xi )):

ψh2 ≡ ψh2 (f (xi )) =

Corresponding Error = O(...........)

5
(c) (7 pts) For the purpose of approximating f 0 (2.0), we consider the
following table Dn = {(xi , yi )|i = 0, 1, ..., n} associated with a
function f (x), i.e., yi = f (xi ), i = 1, ..., n

i xi yi = f (xi )
0 1.6 3.803929 × 10+01
1 1.7 4.745912 × 10+01
2 1.8 5.880257 × 10+01
3 1.9 7.240824 × 10+01
4 2.0 8.866867 × 10+01
5 2.1 1.080384 × 10+02
6 2.2 1.310432 × 10+02
7 2.3 1.582903 × 10+02
8 2.4 1.904805 × 10+02

Fill in the empty slots in the following table adequately, starting

with h0 = 0.4.
Express all the results obtained with 5 significant digits and
rounding to the closest.

h ψh (.) ψh1 (.) ψh2 (.)

h0 = 0.4 XXXXX XXXXX

h0 /2 XXXXX

h0 /4

Best approximation to f 0 (2.0) :

6
3. (20 points) Consider the function
Z x
2 2
erf (x) = √ e−t dt
π 0

Let I = erf (1).

(1)-(5 pts) Determine the minimum number of partition points needed

to approximate I by means of the Composite Simpson’s rule, if the
absolute error in this approximation verifies:
1
|S | < 10−2
2

7
(2)-(5 pts) Approximate I by means of the Composite Simpson’s Rule
using the number of equi-spaced partition points found in part (1) above
.
Express your answer with a precision p = 5 with rounding to the closest.

8
 b−a
(3)- (5 pts)Assume h = 2i
,i = 0, 1, 2, .... Given that:

I = S(h) + s1 h4 + s2 h6 + ..... + sj h2j+2 + ....

Derive the first order Romberg extrapolation formula based on Simp-

son’s rule rule.

S 1 (h) =

(4)- (5 pts) Apply this formula to the previous result obtained in (a-3)
in order to improve the approximation of I .
Do not compute the final answer:
Write only the detailed formula that should be used with the adequate
numerical values of h.

9
4. (25 points) Consider the following set of data

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

where the nodes are not necessarily equally spaced.

(a) (7 pts) Based on the set Dn , derive the general equations of the
Non Natural Quadratic spline function S(x), that approxi-
mate the function f (x).

10
(b) (6 pts) Use the results obtained in (a) to write the equations of
the Non Natural Quadratic Spline S(x) that interpolates the
set of data
D2 = {(0, 1), (0.25, 4/5), (0.5, 2/3), (0.75, 4/7), (1, 1/2)}
and fill in the following table:
[Link] the results in fractions.
i xi yi zi [xi , xi+1 ]
0 0.00 1 . .
1 0.25 4/5 . .
2 0.50 2/3 . .
3 0.75 4/7 . .
4 1.00 1/2 . .









S(x) =








11
(c) (7 pts)Use the equations of the Quadratic
R 1 Spline S(x) obtained in
(b), to approximate the integral I = 0 f (x)dx

12
(d) (5 pts)Let
Z xk+1 Z xk+1
Ik = f (x)dx ≈ Qk = Sk (x)dx
xk xk

where Sk (x) is the equation of the Quadratic Spline on the interval

[xk , xk+1 ], ∀ k = 0, 1, ..., n − 1. Let also
n−1
X n−1
X
I= Ik ≈ Q(h) = Qk
k=0 k=0

Given that on each interval [xk , xk+1 ],

M axxk <x<xk+1 |f (x) − Sk (x)| ≤ ck h3

prove that
n−1
X
|I − Q(h)| ≤ |Ik − Qk | ≤ C(b − a)h3
k=0

Hint: Prove first that |Ik − Qk | ≤ ch4

13
5. (15 points) We seek the approximation of r = a3/4 , where 1 ≤ a < 2,
by seeking the positive root of a polynomial function p(x).

(a) (2 pts)For 1 ≤ a < 2, show that 1 ≤ r < 2.

(b) (3 pts)Write Newton’s iterative formula: rn+1 = g(rn ). Show then

graphically that
lim rn = r, ∀r0 > 0.
n→∞

14
(c) (5 pts)For r0 = 2, show that the resulting Newton’s sequence {rn }
is decreasing.
Hint Show that rn+1 − rn = 4r13 (a3 − rn4 ), n ≥ 0
n

(d) (5 pts) Show that the convergence of rn to r is quadratic, by

finding C such that:

|rn+1 − r| ≤ C|r − rn |2

15