Pearl Centre, S.B. Marg, Dadar (W), Mumbai  400 028. Tel.

4232 4232

ME / CE : Engineering Mathematics
Numerical Methods
MODULE  3

INDEX
Page
Contents Topics
No.
Numerical Solutions of Linear Algebraic Equations
1.1 Gauss Elimination Method 1
1.2 Gauss Siedel Method 2
1.3 Triangularization Method 3
Numerical Solutions of Non-linear Algebraic Equations
1.4 Bisection Method 6
1.5 Method of False Position (Regula Falsi Method) 7
1.6 Secant Method 7
1.7 Newton Raphson Method 8
Numerical Integration
Notes 1.8 Trapezoidal Rule 11
1.9 Simpson's Rule 11
Numerical Solutions of Differential Equations
1.10 Euler's Method 14
1.11 Runge  Kutta Method 15
1.12 Taylor's Series Method 16
1.13 Picard’s Method 16
Interpolation Polynomials and Numerical Differentiation
1.14 Lagrange's Interpolation Polynomial 23
1.15 Newton's Interpolation Polynomial 23
List of Formulae 29
Assignment  1 32
Assignments Assignment  2 35
Assignment  3 38
Test Paper Test Paper  1 41
Answer Key 44
Assignments
Model Solutions 45
Answer Key 60
Test Paper
Model Solutions 61
 Numerical Methods
In scientific and engineering work, a frequently occurring problem is to find the roots of
equation of the form f(x) = 0. If f(x) is a quadratic, cubic or a biquadratic expression, then
algebraic formulae are available for expressing the roots in terms of the coefficients. On
the other hand when f(x) is a polynomial of higher degree on an expression involving
transcendental functions.
i.e. 1 + cos x  5x,
x tan x = cos hx,
ex  sin x etc.

Algebraic roots are not available and recourse must be taken to find the root by
approximate methods.

Numerical Solutions of Linear Algebraic Equations

1.1 Gauss Elimination Method
In this method, solution of a system of linear equation is obtained by transforming the
augmented matrix to an upper triangular form.
Consider a system of m equations in n unknowns as follows:
a11x1 + a12x2 + ……a1n xn = b1
a21x1 + a22x2 + ……a2n xn = b2
.
.
.
.
.
.

am1x1 + am2x2 + ……amn xn = bm

This system of equation can be written in matrix form as:
AX = B
where,

ª a11 a12 .. a1n º ª x1 º ª b1 º

«a a 22 .. a 2n »» «x » «b »
A = « 21 X = « 2» and B= « 2»
« : : : » «: » « : »
« » « » « »
¬am1 am2 .. amn ¼ ¬ xn ¼ ¬b m ¼
In Gauss Elimination method, this augmented matrix is transformed to upper triangular
matrix.

1
Vidyalankar : GATE – Engineering Mathematics

ª a11 a12 a1n b1 º

«a a 22 a 2n b2 »»
Augmented matrix [A : B] = « 21
« : : : : »
« »
¬a 22 am2 amn bm ¼

ª a11 a12 a1n b1 º ªu11 u12 u1n v1 º

«a a22 a2n »
b2 » «0 u u2n v 2 »»
« 21 
 o«
o 22
« : : : : » « : : : : »
« » « »
¬a22 am2 amn bm ¼ ¬0 0 umn vm ¼

This method is also called forward elimination method. Solution of the equation is
obtained from the transformed matrix.

1.2 Gauss Siedel Method

It is an iterative technique to solve a system of n linear equations in n unknowns:
a11 x1 + a12 x2 + …. a1n xn = b1
a21 x1 + a22 x2 + …. a2n xn = b2
. . . .
. . . .
. . . .

an1 x1 + an2 x2 + …. ann xn = bn

These equations can be written in matrix form as:

ª a11 a12 a1n º ª x1 º ª b1 º

«a a 2n »» «« x 2 »» ««b2 »»
« 21 a 22   
« : : : » «: » «: »
« » « » « »
a
¬ n1 a n2 ann ¼ ¬ xn ¼ ¬bn ¼

which is of the form AX = B

In this method, matrix A is decomposed into the sum of a lower triangular matrix and a
strictly upper triangular matrix
A=L+U
where

ª a11 0 0 º ª 0 a12 a1n º

«a a 22 0 »» «0 0 a 2n »»
L = « 21 and U= «
« : : : » «: : : »
« » « »
¬ an1 an2 ann ¼ ¬ 0 0 0 ¼

2
 Notes on Numerical Methods

Then the matrix equation can be rewritten as

(L + U) X = B
? LX + UX = B
LX = B  UX
(k) (k + 1)
From a known (initial guess/given) X the next values in X are found as
(k + 1) (k)
LX = B  UX

= L (B  UX )
(k + 1) 1 (k)
X

The iterations are performed till the solution is obtained with desired accuracy.

1.3 Triangularization Method

This method is also known as decomposition method or factorization method.
Consider the system of n equations in n unknowns :
a11 x1 + a12 x2 + …. a1n xn = b1
a21 x1 + a22 x2 + …. a2n xn = b2
. . . .
. . . .
. . . .

an1 x1 + an2 x2 + …. ann xn = bn

The equation can be written in matrix form as

ª a11 a12 a1n º ª x1 º ª b1 º

«a a 2n »» «« x 2 »» ««b2 »»
« 21 a 22   
« : : : » «: » «: »
« » « » « »
a
¬ n1 a n2 ann ¼ ¬ xn ¼ ¬bn ¼

which is of the form AX = B

In this method, the co-efficient matrix A is factorized or decomposed into the product of a
lower triangular matrix and an upper triangular matrix i.e. matrix A is written as
A = LU
where

ª 11 0 0 º ªu11 u12 u1n º

« 0 »» «0 u u2n »»
L = « 21 22
and U= « 22
« : : : » « : : : »
« » « »
¬ n1 n2 nn ¼ ¬0 0 unn ¼

3
Vidyalankar : GATE – Engineering Mathematics

ª 11 0 0 º ªu11 u12 u1n º ª a11 a12 a1n º

« 0 »» «« 0 u22 u2n »» «a a22 a 2n »»
? « 21 22
  « 21
« : : : » « : : : » « : : : »
« » « » « »
¬ n1 n2 nn ¼ ¬ 0 0 unn ¼ ¬ an1 an2 ann ¼

The matrix multiplication is performed and by comparing the elements of the resulting
matrix with the corresponding elements in A, a system of equations in unknowns ij uij is

obtained. To get a unique solution, it is convenient to choose all the diagonal elements of
either L or u as 1.
x If the diagonal elements of L are chosen as 1, it is called the Doolittle's Method.
x If the diagonal elements of u are chosen as 1, it is called the Crout's Method.
Once the matrices L and U are obtained, the matrix equation AX = B is written as
LUX = B which is solved as UX = Z and LZ = B
First Z can be determined by forward substitution and then X by backward substitution.

Solved Example 1 : ª 1 0 0º ª0 2 3 º
Gauss-Seidel method is used to solve the L = ««2 3 0 »» U = ««0 0 1»»
following equations (as per the given ¬«3 2 1»¼ ¬«0 0 0 »¼
order): [ME  2016] ª 3 0 0º
1«
x1 + 2x2 + 3x3 = 5 L = « 2 1 0 »»
1
3
2x1 + 3x2 + x3 = 1 «¬ 5 2 3 »¼

3x1 + 2x2 + x3 = 3 After k iterations

= L1 (B  UX )
(k+1) (k)
Assuming initial guess as x1 = x2 = x3 = 0, the X
value of x3 after the first iteration is _____ X
(k+1)
= L1B  L1UX
(k)

Solution : After 1st iteration

The given equations in matrix form X
(1+1)
= L1B  L1 UX
(1)

ª 1 2 3 º ª x1 º ª5 º ª0 º
«2 3 1»« x » = « 1»
= ««0 »»
(1)
« » « 2» « » Here X
«¬3 2 1»¼ «¬ x 3 »¼ «¬3 »¼ «¬0 »¼
AX = B ? X
(2+)
= L1B  L1 UX
(1)

Gauss-Siedel method,
ª5º
A=L+U = «« 3 »»
(2)
gives X
where «¬ 6 »¼

? Value of x3 after first iteration is 6.

4
 Notes on Numerical Methods

Solved Example 2 : So the solution of the equation is

Solve the system of linear equations using >a b c@ >1 1/ 2 1/ 2@
Gauss Elimination method.
a+b+c=1 Solved Example 3 :
4a + 3b  c = 6 Solve the system of linear equations using

3a + 5b + 3c = 4 Crout's Method.

Solution : x 1 + x2 + x 3 = 1

The equation can be written in matrix form 4x1 + 3x2  x3 = 6

as 3x1 + 5x2 + 3x3 = 4

ª 1 1 1 º ªa º ª 1º Solution :
«4 3 1»«b » ««6 »» The equations can be written in matrix
« »« »
«¬3 5 3 »¼ «¬ c »¼ «¬4 »¼ form as
which is of the form AX = B ª 1 1 1 º ª x1 º ª 1º
«4 3 1»« x » ««6 »»
Augmented matrix : « » « 2»
«¬3 5 3 »¼ «¬ x 3 »¼ «¬4 »¼
ª 1 1 1 1º
[A : B] = ««4 3 1 6 »» which is of the form AX = B
«¬3 5 3 4 »¼ ª1 1 1 º
The co-efficient matrix A = ««4 3 1»»
R2 moR2  4R1 , R3 moR3 3R1
¬«3 5 3 »¼
ª 1 1 1 1º
«0 1 5 2 » Factorizing A as product of L and U and
« »
«¬0 2 0 1»¼ taking diagonal elements of U as 1.
(Crout's Method)
R3 moR3 2R2
ª 11 0 0 º ª 1 u12 u13 º ª 1 1 1 º
ª1 1 1 1º ?«« 21 22 0 »»««0 1 u23 »» ««4 3 1»»
«0 1 5 2 »
« » «¬ 31 32
»«
33 ¼ ¬ 0 0 1 »¼ «¬3 5 3 »¼
«¬0 0 10 5 »¼
ª 11 u
11 12 u
11 13 º
So we have a+b+c=1 ? « u  »
« 21 21 12 22 21u13  22 u23 »
b  5c = 2 «¬ 31 31u12  32
»
31u13  32 u23  33 ¼

10c = 5
ª1 1 1 º
The third equation gives c = 1/2 ««4 3 1»»
Substituting c = 1/2 in second equation «¬3 5 3 »¼
gives b = 1/2
Finally, from the first equation, we get a = 1

5
Vidyalankar : GATE – Engineering Mathematics

By comparing the corresponding elements, ª1 0 0 º ª Z1 º ª 1º

«4 1 0 »« Z » ««6 »»
we get « » « 2»
 1,  4,
4   3 «¬3 2 10 »¼ «¬ Z3 »¼ «¬4 »¼
11 21 31

u  1 ? u12 = 1 gives Z1 = 1, Z2 = 2 and Z3 = 1/2

11 12

(forward substitution)
u  1
11 13 ? u13 = 1
Further, UX = Z gives
u 
21 12 22  3 ? 22   1
ª 1 1 1º ª x1 º ª 1 º
u 
31 12 32  5 ? 22  2 «0 1 5 »« x » «« 2 »»
« » « 2»
u  u   1 ? u23 = 5 «¬0 0 1»¼ «¬ x 3 »¼ «¬ 1 2 »¼
21 13 22 23

u  u   3 ?   10 which gives x3 = 1/2, x2 = 1/2 and x1 = 1

31 13 32 23 33 33

(backward substitution)
ª1 0 0 º
? L = ««4 1 0 »» ? The solution is
«¬3 2 10 »¼ >x1 x2 x3 @ >1 1 2 1 2@

ª 1 1 1º
and U = ««0 1 5 »»
«¬0 0 1»¼

Consider LZ = B

Numerical Solutions of Non-linear Algebraic Equations

1.4 Bisection Method
If f(x) is continuous in [a, b] and if f(a) and f(b) are of opposite signs i.e. f(a) f(b) < 0, then
there exists atleast one root in [a, b]. In Bisection method, this root is computed by
repeatedly bisecting the interval [a, b] as follows:
Let
ab
x0  
2

If f(x0) = 0, then x0 is the root. Otherwise depending on the sign of f(a) f(x 0), the interval
[a, b] is replaced by [a, x0] or [x0, b].
The new interval is again bisected to obtain x 1. Again if f(x1) = 0, then x1 is the root,
otherwise the interval will be further replaced by a new interval.
This is repeated till the root is obtained.

6
 Notes on Numerical Methods

1.5 Method of False Position (Regula Falsi Method)

Let the graph of equation y = f(x) pass through the two points A(a, f(a)) and B(b, f(b))
such that f(a) and f(b) are of opposite signs.
Then the graph of f(x) cuts the X axis between a and b i.e. there exists a root between a
and b.
y
B[b, f(b)]

a x1 x2
o D b x

A [a, f(a)]

In this method, the curve AB is replaced by chord AB and the intersection of chord AB
with X axis gives the approximate root.

Let chord AB intersect the X axis at x 1.

Then the approximate root of f(x) is

(b  a)f(a) af(b)bf(a)
x1 a  
f(b) f(a) f(b)f(a)

If f(b) and f(x1) are of opposite signs, then the root lies between x 1 and b and so a is
replaced by x1.
The process is repeated till the root is obtained with the desired accuracy.

1.6 Secant Method

This method is similar to the method of false position except that, in this method, f(a) and
f(b) need not be of opposite signs.
The approximation of root is done using the same formula as for method of false position:
af(b)bf(a)
x1 
f(b)f(a)

7
Vidyalankar : GATE – Engineering Mathematics

In general,

xn1f(xn )xn f(xn1 )

xn 1 
f(xn )f(xn1 )

In this method, when going for the next iteration, xn1 is replaced by xn and xn is replaced
by xn+1.

Note : This method fails if f(xn) = f(xn1).

1.7 Newton Raphson Method

In this method, a point x0 is approximated as the root of the equation f(x) = 0
The next approximation of the root is taken as the point of intersection of X axis and the
tangent to the graph of f(x) at (x 0, f(x0)).
If x1 is the point of intersection, then
f(x 0 )
x1 = x 0 
f c(x 0 )

If x1 meets the desired accuracy, then x 1 is the root, otherwise the next approximation of
the root is obtained as
f(x1)
x2 = x1
f c(x1)

In general,
f(xn )
xn 1 xn 
f c(xn )

This is repeated till the root is obtained with the desired accuracy.

8
 Notes on Numerical Methods

Solved Example 4 : The third approximation of the root is

Find an approximate value of the root of (b  a)f(a)
x3 = a 
3
the equation x + x  1 = 0 near x = 1, f(b)  f(a)

using the method of false position. 1  0.64 ( 0.0979)

= 0.64  0.672
Solution : 1  ( 0.0979)
3
f(x) = x + x  1 f(0.672) = 0.024 < 0
f (0) = 1, f (1) = 1 + 1  1 = 1 Similarly the fourth approximation of the
Since f(0) and f (1) are having opposite root is
signs, x4 = 0.679
The root lies between 0 and 1 This is repeated till the root is obtained
?a=0 b=1 with desired accuracy.
The first approximation of the root is given
by Solved Example 5 :
x
(b  a)f(a) Find the real root of the equation xe  3 = 0
x1 = a 
f(b)  f(a) by Regula Falsi Method, correct to three
1  0 1 decimal places.
=0 0.5
1  ( 1) Solution :
x
3
f (0.5) = (0.5) + (0.5) 1 f(x) = xe  3

= 0.125 + 0.5 1 f(1) = e  3 = 2.718  3 = 0.282 < 0

1˜1
= 0.375 < 0 f(1˜1) = 1˜1 e 3
Since f(0.5) and f (1) have opposite sign, = 1˜1 u 3.004  3 = 0.304 > 0
the root lies between 0.5 and 1 ' f(1) and f(1˜1) have opposite sign, the
? a = 0.5 b = 1 real root of f(x) = 0 lies between 1 and 1˜1
The second approximation of the root is ? a = 1, b = 1˜1
(b  a)f(a) The first approximation of the root is
x2 = a 
f(b)  f(a)
b  a f(a)
x1 = a 
1  0.5 ( 0.375) f(b)  f(a)
= 0.5  0.64
1  (0.375)
1˜ 1  1 (0.282)
f (0.64) =  0.0979 < 0 = 1 1.048
0.304  (0.282)
'f (0.64) and f (1) have opposite sign, the
root lies between 0.64 and 1 f (1.048) =  0.0112

? a = 0.64, b = 1

9
Vidyalankar : GATE – Engineering Mathematics

' f(1.048) and f (1˜1) have opposite sign, f(x1) x1 log10 x1  1.2
x2 = x1  x1
the root lies between 1.048 and 1˜1 f c(x1) log10 x1  0.4343

? a = 1.048, b = 1˜1 0.4343x1  1.2 0.4343 u 3  1.2

log10 x1  0.4343 log103  0.4343
The second approximation of the root is
1.3029  1.2 2.5029
b  a f(a) 2.7
x3 = a  0.4771  0.4343 0.9114
f(b)  f(a)
The third approximation is
1˜ 1  1.048 ( 0.0112)
= 1.048  f x1 0.4343x 2  1.2
0.304  ( 0.0112) x 3 = x2  =
f c x1 log10 x 2  0.4343
= 1.0498
0.4343 u 2.7  1.2
Hence the real root of f (x) = 0 is 1.050 =
log10 2.7  0.4343

1.1726  1.2
Solved Example 6 : = = 2.74
0.4314  0.4343
Find the real root of x log10 x 1.2 = 0, The fourth approximation is
correct to three decimal places.
f(x 3 ) 0.4343x 3  1.2
x4 = x3  =
Solution : f c(x 3 ) log10 x 3 0.4343
f(x) = x log 10 x  1.2
= 0.4343 u 2.74  1.2
f(1) = log101  1.2 = 0  1.2 = 1.2 log10 2.74  0.4343

f(2) = 2log102  1.2 1.1900  1.2

2.743
= 2 (0.3010)  1.2 =  0.5980 0.4378  0.4343

f(3) = 3 log103  1.2

= 3 (0.4771)  1.2 = 0.2313 Solved Example 7 :
3
Since f(z) and f (3) are of apposite sign, the Find the approximate root of x + x  1 in

real root of f(x) = 0 lies between 2 and 3. [0, 1] using Bisection Method.
Also f(3)< f(2) numerically, we take the first Solution :
3
approximation x1 = 3 f(x) = x + x  1
f(x) = x log10 x  1.2 f(0) = 1 < 0 f(1) = 1 > 0

1 ? Root lies between 0 and 1

f c (x) = log10 x + x ˜ log10e
x ab 0 1
Taking x0 = = = 0.5
ª d 1 º 2 2
…« l ga x
log loga e »
¬ dx x ¼ f(x0) = f(0.5) = 0.375 < 0
= log10 + 0.4343 ? Root lies between 0.5 and 1
The second approximation is given by

10
 Notes on Numerical Methods

0.5  1 ? Root lies between 0.625 and 0.75

Taking x1 = = 0.75
2 0.625  0.75
Taking x3 = = 0.6875
f(x1) = f(0.75) = 0.1719 > 0 2
? Root lies between 0.5 and 0.75 This is repeatedly till the root is obtained
0.5  0.75 with desired accuracy. The further
Taking x2 = = 0.625
2 approximation of x are x4 = 0.65625,
f(x2) = f(0.625) = 0.13 < 0 x5 = 0.671875, x6 = 0.6796875 etc.

            

   

              

                

11
Vidyalankar : GATE – Engineering Mathematics

Assignment  1
Duration : 45 Min. Max. Marks : 30

Q 1 to Q 10 carry one mark each 1

3. Simpson's rule is used to integrate
3 1 3
1. The definite integral ³
1 x
dx is
3 2 9
the function f(x) = x  between
evaluated using Trapezoidal rule with 5 5
a step size of 1. The correct answer is x = 0 and x = 1 using the least number
_________ [ME  2014] of equal sub-intervals. The value of the
integral is ___________ [ME  2015]
2. Match the application to appropriate
numerical method. [EC  2014] 4. Using a unit step size, the value of
2
Application integral ³1 x nxdx by trapezoidal
P1: Numerical integration
rule is ____________ [ME  2015]
P2: Solution to a transcendental
equation
5. Only one of the real roots of
P3: Solution to a system of linear 6
f(x) = x  x  1 lies in the interval
equations
1 ŭ x ŭ 2 and bisection method is
P4: Solution to a differential equation
used to find its value. For achieving an
Numerical Method accuracy of 0.001, the required
M1: Newton-Raphson Method minimum number of iterations is
M2: Runge-Kutta Method __________. [EE  2017]
M3: Simpson's 1/3-rule
6. In Newton-Raphson iterative method,
M4: Gauss Elimination Method
the initial guess value (x ini) is
(A) P1M3, P2M2, P3M4, P4M1
considered as zero while finding the
(B) P1M3, P2M1, P3M4, P4 M2
roots of the equation:
(C) P1M4, P2M1, P3M3, P4M2 2 3
f(x) = 2 + 6x  4x + 0.5x .
(D) P1M2, P2M1, P3M3, P4M4
The correction, 'x, to be added to xini
in the first iteration is ___. [CE  2015]

32
 Assignment on Numerical Methods

3
7. Solve the equation x = 10 cos(x) using 9. The root of the function f(x) = x + x - 1
the Newton-Raphson method. The obtained after first iteration on
initial guess is x = S/4. The value of the application of Newton-Raphson
predicted root after the first iteration, scheme using an initial guess of x0 =1
up to second decimal, is ________. is [ME  2016]
[ME  2016] (A) 0.682 (B) 0.686
8. Numerical integration using trapezoidal (C) 0.750 (D) 1.000
rule gives the best result for a single
variable function, which is [ME  2016]
(A) linear (B) parabolic
(C) logarithmic (D) hyperbolic

10. Function f is known at the following points:

x 0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0
f(x) 0 0.09 0.36 0.81 1.44 2.25 3.24 4.41 5.76 7.29 9.00
3
The value of ³0 f(x)dx computed using the trapezoidal rule is [CS  2013]

(A) 8.983 (B) 9.003 (C) 9.017 (D) 9.045

Q 11 to Q 20 carry two marks each 1

³0 e
x
13. The value of dx using the
11. Using the trapezoidal rule, and dividing
the interval of integration into three trapezoidal rule with four equal
equal subintervals, the definite integral subintervals is [CE  2021]
(A) 1.718 (B) 1.727
1

³ | x | dx is _______.
1
[ME  2014] (C) 2.192 (D) 2.718

14. Consider an ordinary differential

4 dx
equation = 4t + 4. If x = x0 at t = 0,
12. The value of ³ In(x) dx calculated
2.5
dt
the increment in x calculated using
using the Trapezoidal rule with five
Runge-Kutta fourth order multi-step
subintervals is _______. [ME  2014]
method with a step size of 't = 0.2 is
[ME  2014]
(A) 0.22 (B) 0.44
(C) 0.66 (D) 0.88

33
Vidyalankar : GATE – Engineering Mathematics

15. Newton-Raphson method is used to 17. The Newton-Raphson method is used

find the roots of the equation, to solve the equation
3 2 3 2
x + 2x + 3x  1 = 0. If the initial guess f(x) = x  5x + 6x  8 = 0. Taking the
nd
is x0 = 1, then the value of x after 2 initial guess as x = 5, the solution
iteration is _______ [ME  2015] obtained at the end of the first iteration
is __________. [EC  2015]
2
16. The quadratic equation x  4x + 4 = 0
is to be solved numerically, starting 18. The error in numerically computing the
with the initial guess x0 = 3. The S
integral ³0 sinx  cos x dx using the
Newton-Raphson method is applied
trapezoidal rule with three intervals of
once to get a new estimate and then
equal length between 0 and S is
the Secant method is applied once
using the initial guess and this new ___________. [ME  2016]

estimate. The estimated value of the

root after the application of the Secant
method is __________. [CE  2015]

‰‰‰‰‰‰

34
 Test Paper on Numerical Methods

Test Paper  1
Duration : 30 Min. Max. Marks : 25

Q1 to Q5 carry one mark each ª 1 0 0º

«
(A) «0.8 1 0 »»
1. Let f(x) = 0 have a root r in the interval
[a, b] and the method of bisection «¬0.8 .3333 1»¼

generates a sequence {f(cn} which ª 1 0 0º

«
(B) « 0.8 1 0 »»
converges to the root r. Let n f(cn ) .
«¬0.6667 0.3333 1»¼
Then the sequence {n }
ª 1 0 0º
(A) Decreases monotonically
(C) «0.6667 1 0 »»
«
(B) Increases monotonically
«¬0.3333 0.8 1»¼
(C) Constant
ª 1 0 0º
(D) None of these
(D) «0.3333 1 0 »»
«
«¬0.6667 0.8 1»¼
2. A real root of the equation

x3  x  4 0 in the interval [1, 2]

4. Let A be the matrix
obtained by the method of bisection at
ª 2 2 1º
the end of fourth iteration is «4 5 2»
« »
(A) 1.435 (B) 1.875 «¬ 2 1 2 »¼
(C) 1.96875 (D) None of these
Let A = LU, where L is the lower
triangular matrix and U is the upper
ª3 2 1º
3. Let A be the matrix : ««2 3 2 »» triangular matrix.
«¬ 1 2 3 »¼ Using Dolittle LU decomposition (i.e.
with L having unit diagonal values) the
Let A = LU, where L is the lower
matrix U will be
triangular matrix and U is the upper
triangular matrix. ª2 2 1 º ª3 2 1 º
(A) ««0 1 4 »» (B) ««0 1 4 »»
Using Dolittle LU decomposition (i.e.
«¬0 0 11»¼ «¬0 0 11»¼
with L having unit diagonal values), the
matrix L will be ª2 2 1º ª2 2 1 º
«
(C) «0 1 4 »» (D) ««0 1 4 »»
«¬0 0 11»¼ «¬0 0 11»¼

41
Vidyalankar : GATE – Engineering Mathematics

2 Let A = LU, where L is the lower

5. Consider the integral I ³
1
x 3  1 dx .
triangular matrix and U is the upper
th
Using Simpson’s 3/8 rule, and taking triangular matrix.

3 intervals I will be approximately Using Dolittle LU decomposition (i.e.

equal to with L having unit diagonal values), the

(A) 4.75 (B) 5.75 matrix U will be

(C) 6.75 (D) None of these ª2 3 1 º

«
(A) «0 2.5 4.5 »»
Q6 to Q15 carry two marks each «¬0 0 8.5 »¼
6. Let f(x) { x3  4x2  1 0 . f(0) < 0,
ª4 3 1 º
f(1) > 0. A root of the equation, (B) «0 2.5 4.5 »»
«

obtained by the method of bisection, «¬0 0 8.5 »¼

with an error less than 101 is ª4 3 1 º
(A) 0.46875 (B) 0.4375 (C) «0 2.5 4.5 »»
«

(C) 0.37500 (D) None of these «¬0 0 8.5 »¼

7. Consider the equation ª4 3 1 º

(D) ««0 2.5 4.5 »»
f(x) = x3  4x2  1 = 0. Let x0 = 0.5. Let
«¬0 0 8.5 »¼
x1, x2, x3, . . . be the sequence
generated by NewtonRaphson method. 10. The depth d of a river at a distance x
Then x4 = (approximately) from one of the banks is given in the
(A) 0.6728 (B) 0.4728 table below :
(C) 0.7728 (D) 0.5728 X D

8. Let f(x) = x3  3x  2 = 0. Take x0 = 1. 0 0

10 4.1
In four iterations, NewtonRaphson
20 6.9
method converges approximately to x4 =
30 9
(A) 1.01325 (B) 0.91325
40 12.1
(C) 1.51325 (D) None of these
50 14.9
9. Let A be the matrix:
60 13.9
ª 4 3 1º 70 8.1
« 2 4 5 »
« » 80 3
«¬ 1 2 6 »¼

42
 Test Paper on Numerical Methods

Using Trapezoidal rule, the area of the 11. The truncation error in the Simpson’s
rd
crosssection of the river is 1/3 rule is
approximately (b  a)h4 iv (b  a)h4 iv
(A) | fav | (B) fav
(A) 666 (B) 667 90 180
(C) 662 (D) 664 (b  a)3 h3 iv
(C) fav (D) None of these
180
iv
where fav is the average value of the
th
4 derivative of f on [a, b].

12. Given the table :

x 1 1.5 2 2.5 3 3.5 4

x
y(x) e2 1.648721 2.117 2.718282 3.490343 4.481689 5.754603 7.389056
x
The area under the curve y(x) e 2 from x = 1 to x = 4, using Simpson’s 1/3rd rule

will be approximately
(A) 10.48092 (B) 9.48092 (C) 11.88092 (D) 11.48092

S/ 2
13. Let I = ³0 sinx dx . Using Simpson’s 15. Consider the differential equation
dy
th 2t  y, with y(0) = 3. Take the
3/8 rule and h = S/6, the approximate dt
value of I is equal to step size h = 0.1. Using Euler's method
(A) 1.86104 (B) 1.76104 y(0.2) will be approximately equal to
(C) 1.66104 (D) None of these (A) 3.100 (B) 3.610
14. Consider the initial value problem: (C) 2.999 (D) None of these
dy
1  xy; y(0) 1. Take h = 0.2.
dx
Using RungeKutta fourth order method,
y(0.2) will be approximately equal to
(A) 1.17755 (B) 0.82245
(C) 1.10755 (D) None of these

‰‰‰‰‰‰

43