Numerical solution of Algebraic and Transcendental equations

Solution of nonlinear equations

A polynomial equation of the form f (x ) = a 0 x n + a1 x n −1 + a 2 x n − 2 + .......... + a n −1 x + a n

(where a i ’s are constants , a0  0 and n is a positive integer) is called an
algebraicequation.

Eg: x 3 − 2 x 2 − 7 x + 24 = 0

Transcendental equations:

An equation which involves transcendental functions like sin x , cos x, tan x ,

ex, log10 x, loge x etc are called transcendental equation.

Example

2ex+1=0

2x+ cos x-1=0

log10 x-2x=12

Roots of an equation

A value a for which f (a ) = 0 is called a root or solution of the equation

f (x ) = 0 .

Geometrically, the point at which a curve y = f (x ) intersects the x-axis , is a

root of the equation f (x ) = 0 .

For example, the roots of the equation x 2 + 4 x + 3 = 0 are -1 and -3. We

observe below that the graph of the function y = x 2 + 4 x + 3 passes through
these points on the x-axis

If an equation has two equal roots, the x-axis is tangential to the graph at
that point.
For example, the equation x 2 = 0 has a double root at x = 0 . The graph of
y = x 2 has x-axis as a tangent at x = 0

Fundamental theorem of Algebra

Every equation has a root, real or imaginary.

Note:

1. Every algebraic equation of degree n has exactly n roots, real or

imaginary.

2. The number and nature of roots of a transcendental equation are not

known.

3. If f (x ) is continuous in the interval (a, b) and if f (a ) and f (b) are of

opposite signs, then the

equation f (x ) = 0 will have at least one real root between a and b.

RegulaFalsi Method (Method of false position or Linear interpolation
method)

Aim: To find the real root of the equation f (x ) = 0 .

1. Find an interval (a, b) in which the root lies. i.e., such that f (a ) and f (b)
have opposite signs.

af (b ) − bf (a )
2. The first approximation of the root is x1 =
f (b ) − f (a )

3. Now either f (a ) and f (x1 ) have opposite signs or f (x1 ) and f (b) have
opposite signs.

4. If f (a ) and f (x1 ) have opposite signs, then the root lies between a and x1 .

af (x1 ) − x1 f (a )
x2 =
f (x1 ) − f (a )

Note: If f (x1 ) and f (b) have opposite signs, then the root lies between
x1 and b .

x1 f (b ) − bf (x1 )
x2 = .
f (b ) − f (x1 )

5. Similarly, we find x3 , x 4 ,...... till the desired accuracy is obtained.

Note: Order of convergence of RegulaFalsi method is 1.618

Geometrical interpretation of RegulaFalsi method

f (a ) and f (b) have opposite signs, so the points A(a, f (a )) and B(b, f (b)) are on
opposite sides of the x - axis. The chord AB meets the x - axis at x = x1 . This
is the first approximation of the root. If f (a ) and f (x1 ) have opposite signs,
then the root lies between a and x1 and we join the points
(a, f (a)) and (x1 , f (x1 )) . This chord meets the x - axis at x = x2 . This is the
second (and closer) approximation of the root. The sequence x1 , x2 , x3 , .........
converges to the root.

Remember:When working with sin, cos etc. the calculator should be

in radian mode.

When differentiation or integration has to be done, log should to be

the base e.

Problems:

1. Find the approximate root of xex − 2 = 0 by method of false position.

Solution:

Let
f (x ) = xex − 2

f (0 ) = −2  0
f (1) = 0.7183  0

Therefore the root lies between 0& 1.

Take A =0 and B=1

af (b ) − bf (a )
By Reulafalsi method xn =
f (b ) − f (a )

Iteration a b f(a) f(b)

xn =
( AD − BC ) f(xn)
(A) (B) (C) (D) (D − C )
1 0 1 -2 0.7183 0.7358 -0.4643
2 0.7358 1 -0.4643 0.7183 0.8395 -0.0564
3 0.8395 1 -0.0564 0.7183 0.8512 -0.0061
4 0.8512 1 -0.0061 0.7183 0.8525 -0.0005
5 0.8525 1 -0.0005 0.7183 0.8526 0
Hence the root is 0.853
2. Solve for a positive root of by x − cos x = 0 Regula falsi method.

Solution:

Let
f (x) = x − cos x

f (0 ) = −1  0
f (1) = 0.4597  0

Therefore the root lies between 0& 1.

Take A =0 and B=1

af (b ) − bf (a )
By Reulafalsi method xn =
f (b ) − f (a )

Iteration a b f(a) f(b)

xn =
( AD − BC ) f(xn)
(A) (B) (C) (D) (D − C )
1 0 1 -1 0.4597 0.6851 -0.0893
2 0.6851 1 -0.0893 0.4597 0.7363 -0.0047
3 0.7363 1 -0.0047 0.4597 0.7390 -0.0001
4 0.7390 1 0 0.4597 0.7391 0
Hence the root is 0.739

3. Find the approximate root of x log10 x − 1.2 = 0 by method of false position.

Solution:

f (x ) = x log10 x − 1.2

f (1) = −1.2000  0
f (2) = −0.5979  0
f (3) = 0.2314  0

Therefore the root lies between 2&3.

Take A =2 and B=3

af (b ) − bf (a )
By Reulafalsi method xn =
f (b ) − f (a )
Iteration a B f(a) f(b)
xn =
( AD − BC ) f(xn)
(A) (B) (C) (D) (D − C )
1 2 3 -0.5979 0.2314 2.7210 -0.0171
2 2.7210 3 -0.0171 0.2314 2.7402 -0.0004
3 2.7402 3 -0.0004 0.2314 2.7406 0
4 2.7406 3 0 0.2314 2.7406 0
Hence the root is 2.741

4. Find the positive root of x = 2 x + 5 by method of false position.

3

Solution:

Let
f (x ) = x 3 − 2 x − 5

f (0 ) = −5  0
f (1) = −6  0
f (2) = −1  0
f (3) = 16  0

Therefore the root lies between 2& 3.

Take A =2 and B=3

af (b ) − bf (a )
By Reulafalsi method xn =
f (b ) − f (a )

Iteration a b f(a) f(b)

xn =
( AD − BC ) f(xn)
(A) (B) (C) (D) (D − C )
1 2 3 -1 16 2.0588 -0.3911
2 2.0588 3 -0.3911 16 2.0813 -0.1468
3 2.0813 3 -0.1468 16 2.0897 -0.0540
4 2.0897 3 -0.0540 16 2.0928 -0.0195
5 2.0928 3 -0.0195 16 2.0939 -0.0073
6 2.0939 3 -0.0073 16 2.0943 -0.0028
Hence the root is 2.094

5. Find the negative root of x + 1.2 x = 4 x + 4.6 by method of false position.

3 2

Solution

f (x ) = x3 + 1.2 x 2 − 4 x − 4.62 [Replace x by –x for negative root]

g (x ) = − x3 + 1.2 x 2 + 4 x − 4.62

g (0) = −4.0000  0
g (1) = −0.4000  0
g (2) = 0.2000  0

Therefore the root lies between 1 & 2.

Take A=1 and B=2

ag (b ) − bg (a )
By Reulafalsi method xn =
g (b ) − g (a )

a b g(a) g(b)
xn =
( AD − BC ) g(xn)
(A) (B) (C) (D) (D − C )
1 1 2 -0.4 0.2 1.6667 0.7704
2 1 1.6661 -0.4 0.7704 1.2279 0.2695
3 1 1.2279 -0.4 0.2695 1.1362 0.0272
4 1 1.1362 -0.4 0.0272 1.1275 0.0022
5 1 1.1275 -0.4 0.0022 1.1268 0.0001
6 1 1.1268 -0.4 0.0001 1.1268 0.0001
The root of g(x) is 1.127

Hence the root of f(x) is -1.127

Practice problems

1. Find the positive root of x − 4 x + 1 = 0 by method of false position.

3

Ans : 0.254

2. Solve by Regula falsi method find x = 100 .

x

Ans : 4.967

3. Find the positive root of tan x + tanh x by method of false position.

Ans : 2.365

4. Solve the equation xtan x = −1 by Regula falsi method starting with a=2.5 and
b=3 correct to 3 decimal places.

Ans : 2.798
Fixed point iteration method (Iteration method)

Aim: To find the real root of the equation f (x ) = 0 .

1. Find an interval (a, b) in which the root lies. i.e., such that f (a ) and f (b) have
opposite signs

2. Rewrite the equation in the form x =  (x ) .

3. Verify that  (x )  1 for all x  (a, b)

4. Choose an initial approximation x 0 of the root in the interval (a, b) .

5. Find successive approximations of the root as x1 =  (x0 ), x2 =  (x1 ), x3 =  (x2 ),........

6. If the sequence x0 , x1 , x2 ,...... converges to  , then  is the required root of the

equation

f (x ) = 0 .

Note:1. Condition for convergence of fixed point iteration method is that  (x )  1 .

2. Order of convergence of fixed point iteration method is 1. (Linear

convergence)

1. Find the real root of cos x = 3x − 1 , by iteration method. Correct to four decimal
places.

Solution:

Let f (x) = cos x − 3x + 1

f (0) = 2  0
f (1) = −1.45970  0

Therefore the root lies between 0 and 1.

Rewrite the given equation as x =  (x )

cos x = 3x −1
3x = cos x + 1

cos x + 1
x= =  (x )
3

− sin x
 ' (x ) =
3

− sin 0
 (0) = = 0 1
3

− sin1
 (1) = = 0.28049  1
3

(x)  1 fox all x in (0, 1)

Assume x0=1

cos x0 + 1 cos1 + 1
x1 = = = 0.51343
3 3

cos x1 + 1
x2 = = 0.62369
3

cos x2 + 1
x3 = = 0.60391
3

cos x3 + 1
x4 = = 0.60771
3

cos x4 + 1
x5 = = 0.60699
3

cos x5 + 1
x6 = = 0.60712
3

cos x6 + 1
x7 = = 0.60710
3

Hence the root is 0.6071

 𝑐𝑜𝑠𝑥 + 1
𝑥=
3

X0=0.5

cos⁡(0.5)+1
𝑥1 = ⁡= 0.62586
3

cos⁡(0.62586)+1
𝑥2 = =0.60349
3

cos⁡(0.60349)+1
𝑥3 = ⁡= 0.60779
3

cos⁡(0.60779)+1
𝑥4 = ⁡= 0.60697
3

cos⁡(0.60697)+1
𝑥5 = =0.60712
3

cos⁡(0.60712)+1
𝑥6 = =0.60710
3

2. Find the real root of 2 x − log10 x = 7 , by iteration method.

Solution:

Let f (x ) = 2 x − log10 x − 7

f (1) = −5.0000  0
f (2 ) = −3.3010  0
f (3) = −1.4771  0
f (4 ) = 0.3979  0

Therefore the root lies between 3 and 4.

Rewrite the given equation as x =  (x )

2 x − log10 x = 7

2 x = log10 x + 7

log 10 x + 7
x= =  (x )
2

d
log10 x = d log e x log10 e
dx dx

=
d
0.4343 log e x 
dx Type⁡equation⁡here.

0.4343
=
x

1  0.4343 
 ' (x ) =  
2 x 

 ' (x ) =
0.2172
x

 (3) =
0.2172
= 0.0724  1
3

 (4) =
0.2172
= 0.0543  1
4

 (x )  1 fox all x in (3, 4)

Assume x0=4

log 10 x0 + 7
x1 = = 3.8010
2

log 10 x1 + 7
x2 = = 3.7900
2

log10 x2 + 7
x3 = = 3.7893
2
 log10 x3 + 7
x4 = = 3.7893
2

Hence the root is 3.789

3. Solve the equation x + x − 1 = 0 , for positive roots by iteration method.

3 2

Solution:

Let f (x ) = x + x − 1
3 2

f (0) = −1  0
f (1) = 1  0

Therefore the root lies between 0 and 1.

Rewrite the given equation as x =  (x )

x3 + x 2 = 1

Type⁡equation⁡here.

x 2 (x + 1) = 1

1
x2 =
1+ x

= (1 + x ) 2 =  (x )
1 1
−
x=
1+ x

 ' (x ) = − (1 + x )− 2
1 3

 ' (x ) = −
1
2(1 + x ) 2
3

 ' (0) = −
1
= 0.5  1
2(1 + 0)
3
2

 ' (1) = −
1
= 0.1768  1
2(1 + 1) 2
3
(x)  1 fox all x in (0, 1)

Assume x0=0

1
x1 = =1
1 + x0

1
x2 = = 0.7071
1 + x1

1
x3 = = 0.7654
1 + x2

1
x4 = = 0.7526
1 + x3

1
x5 = = 0.7554
1 + x4

1
x6 = = 0.7548
1 + x5

Hence the root is 0.755

4. Solve the equation 3x − 1 + sin x = 0 , by iteration method.

Solution:

Let f (x ) = 3 x − 1 + sin x

f (0 ) = −1  0
f (1) = 1.6430  0

Therefore the root lies between 0 and 1.

Rewrite the given equation as x =  (x )

3x − 1 + sin x = 0

1 + sin x =  (x )
1
x=
3
 ' (x ) = (1 + sin x )− 2 cos x
11 1

32

 ' (x ) =
1
cos x
6 1 + sin x

 ' (0) =
1
cos 0 = 0.1667  1
6 1 + sin 0

 ' (1) =
1
cos1 = 0.544  1
6 1 + sin1

(x)  1 fox all x in (0, 1)

Assume x0=0

1
x1 = 1 + sin x0 = 0.3333
3

1
x2 = 1 + sin x1 = 0.3840
3

1
x3 = 1 + sin x2 = 0.3908
3

1
x4 = 1 + sin x3 = 0.3917
3

1
x5 = 1 + sin x4 = 0.3918
3

Hence the root is 0.392

5. Solve the equation e − 3x = 0 , by iteration method.

x

Solution:

Let f (x ) = e − 3x
x

f (0 ) = 1  0
f (1) = −0.2817  0

Therefore the root lies between 0 and 1.

Rewrite the given equation as x =  (x )

e x − 3x = 0

e x = 3x

ex
x= =  (x )
3

ex
 ' (x ) =
3

e0
 ' (0) = = 0.3333  1
3

e1
 (1) =
'
= 0.9061  1
3

(x)  1 fox all x in (0, 1)

Assume x0=1

e x0
x1 = = 0.9061
3

e x1
x2 = = 0.8249
3

e x2
x3 = = 0.7605
3

e x3
x4 = = 0.7131
3

e x4
x5 = = 0.6801
3

e x5
x6 = = 0.6580
3

e x6
x7 = = 0.6437
3
 e x7
x8 = = 0.6345
3

e x8
x9 = = 0.6287
3

e x9
x10 = = 0.6250
3

e x10
x11 = = 0.6228
3

e x11
x12 = = 0.6214
3

e x12
x13 = = 0.6205
3

Hence the root is 0.621

Practice problems

1. Solve the equation x = 2 x + 5 , by iteration method.

3

Ans: 2.0945

2. Solve the equation cos x − xe = 0 , by iteration method.

x

Ans: 0.517

3. Solve by iteration method: 4 x − e x = 0

Ans : 0.358

Newton Raphson method

f (xn )
1. Newton’s formula is xn+1 = xn −
f (xn )

2. This method is used when f (x ) can be easily determined.

 f (x ) f (x )  ( f (x )) in the interval.
2
3. Condition for convergence is:

4. The order of convergence of Newton’s method is 2, i.e. the convergence is

quadratic. (The error at any step is proportional to the square of the error in the
previous step).

Problems

1. By using Newton-Raphson’s method, find the root of 3x = cos x + 1 .

Solution:

Let f (x) = 3x − cos x − 1

f (0 ) = −2  0
f (1) = 1.4597  0

f ' (x ) = 3 + sin x

Therefore the root lies between 0 and 1.

Newton-Raphson’s method

f (xn )
xn +1 = xn −
f (xn )

Assume x0 = 1

f (xn )
xn +1 = xn −
f (xn )

Put n=0

f (x0 )
x1 = x0 −
f ' (x0 )

f (1)
= 1−
f ' (1)

1.4597
=1−
3.8415
=0.6200

Put n=1

f (x1 )
x2 = x1 −
f ' (x1 )

f (0.62)
= 0.62 −
f ' (0.62)

0.0461
= 0.62 −
3.5810

=0.6071

Put n=2

f (x2 )
x3 = x2 −
f ' (x2 )

f (0.6071)
= 0.6071 −
f ' (0.6071)

= 0.6071 − 0

=0.6071

Hence the root is 0.607

2. By using Newton-Raphson’s method, find the root of 2 x − 3x − 6 = 0 .

3

Solution:

Let f (x ) = 2 x3 − 3x − 6

f (0) = −6  0
f (1) = −7  0
f (2) = 4  0

Therefore the root lies between 1 and 2.

Newton-Raphson’s formula is
 2𝑥 3 − 3𝑥 − 6
𝑥−( )
6𝑥 2 − 3

f (xn )
xn +1 = xn −
f (xn )

f ' (x ) = 6 x 2 − 3

Assume x0 = 2

f (xn )
xn +1 = xn −
f (xn )

Put n=0

f (x0 )
x1 = x0 −
f ' (x0 )

f (2)
=2−
f ' (2)

4
=2−
21

=1.8095

Put n=1

f (x1 )
x2 = x1 −
f ' (x1 )

f (1.8095)
= 1.8095 −
f ' (1.8095)

0.4212
= 2−
16.6457

=1.9747

Put n=2
 f (x2 )
x3 = x2 −
f ' (x2 )

f (1.9747 )
= 1.9747 −
f ' (1.9747 )

3.4763
= 1.9747 −
20.3966

=1.8043

Put n=3

f (x3 )
x4 = x3 −
f ' (x3 )

f (1.8043)
= 1.8043 −
f ' (1.8043)

0.3349
= 1.8043 −
16.5330

=1.7840

Put n=4

f (x4 )
x5 = x4 −
f ' (x4 )

f (1.7840)
= 1.7840 −
f ' (1.7840)

0.0037
= 1.7840 −
16.0959

=1.7838

Hence the root is 1.784

2𝑥 3 − 3𝑥 − 6
𝑥−( )
6𝑥 2 − 3

X0=1.5
X1=1.85714

X2=1.78711

X3=1.78378

X4=1.78377

X5=1.78377

Correct to 4 decimal places approx soln is 1.7838.

3. Using Newton-Raphson’s method, find the real root of x log10 x = 1.2 .

Correct to four decimals.

Solution:

Let f (x ) = x log10 x − 1.2

f (1) = −1.2  0
f (2) = −0.59794  0
f (3) = 0.23136  0

Therefore the root lies between 2 and 3.

f (x ) = xlog e x log10 e − 1.2

f (x ) = x0.43429 log e x − 1.2

 1
f ' (x ) = x 0.43429  + log e x 0.43429
 x

f ' (x ) = 0.43429 + log10 x

Newton-Raphson’s formula is

f (xn )
xn +1 = xn −
f (xn )
Assume x0 = 3

f (xn )
xn +1 = xn −
f (xn )

Put n=0

f (x0 )
x1 = x0 −
f ' (x0 )

f (3)
= 3−
f ' (3)

0.23136
= 3−
0.91141

=2.74615

Put n=1

f (x1 )
x2 = x1 −
f ' (x1 )

f (2.74615)
= 2.74615 −
f ' (2.74615)

0.00480
= 2.74615 −
0.87301

=2.74065

Put n=2

f (x2 )
x3 = x2 −
f ' (x2 )

f (2.74065)
= 2.74065 −
f ' (2.74065)

= 2.74065 − 0

=2.74065

Hence the root is 2.7407

4. Using Newton-Raphson’s method, find the real root of xe − 2 = 0 .
x

Solution:

Let f (x ) = xex − 2

f (0 ) = −2  0
f (1) = 0.7183  0

Therefore the root lies between 0 and 1.

f ' (x ) = xex + e x

Newton-Raphson’s formula is

f (xn )
xn +1 = xn −
f (xn )

Assume x0 = 1

f (xn )
xn +1 = xn −
f (xn )

Put n=0

f (x0 )
x1 = x0 −
f ' (x0 )

f (1)
= 1−
f ' (1)

0.7183
=1−
5.4366

=0.8679

Put n=1

f (x1 )
x2 = x1 −
f ' (x1 )
 f (0.8679)
= 0.8679 −
f ' (0.8679)

0.0673
= 0.8679 −
4.4492

=0.8528

Put n=2

f (x2 )
x3 = x2 −
f ' (x2 )

f (0.8528)
= 0.8528 −
f ' (0.8528)

0.0008
= 0.8528 −
4.3471

=0.8526

Hence the root is 0.853

5. Find an iteration formula of N where N is the positive number and hence find
12

Solution:

Let x = N

x2 = N

x2 − N = 0

f (x ) = x 2 − N …………….(1)

f l (x ) = 2 x
f (xn )
xn +1 = xn −
f (xn )

= xn −
x2
n −N 
2 xn
 2 xn2 − xn2 + N
=
2 xn

xn2 + N
=
2 xn

xn2 N
= +
2 xn 2 xn

xn N
xn + 1 = +
2 2 xn
……………(2)

This is the required iterative formula.

To find square root of N where N=12

Put N=12 in (1) and (2)

f (x ) = x 2 − 12

xn 12
xn + 1 = +
2 2 xn

f (0) = −12  0
f (1) = −11  0
f (2 ) = −8  0
f (3) = −3  0
f (4 ) = 4  0

Therefore the root lies between 3 and 4.

Assume x0=3.5

Put n=0

x0 12
x1 = +
2 2 x0

3 12
= +
2 2(3)
 3
= +2
2

=3.5

x1 12
x2 = +
2 2 x1

3.5 12
= +
2 2(3.5)

=1.75+1.7143

=3.4643

x2 12
x3 = +
2 2 x2

3.4643 12
= +
2 2(3.4643)

=1.7322+1.7320

=3.4643

Hence the root is 3.464

X1=3.464

X2=3.464

6. Find an iteration formula to find the reciprocal of given number N and hence find
the value of 1/31

Solution:
 1
Let x =
N

1
N=
x

1
N− =0
x

f (x ) = N −
1
…………….(1)
x

f l (x ) =
1
x2
f (xn )
xn +1 = xn −
f (xn )

 1
N − 
= xn − 
xn 
1
xn2

xn2
= xn − xn2 N +
xn

= xn − xn2 N + xn

xn+1 = 2 xn − xn2 N
……………(2)

This is the required iterative formula.

To find 1/31

Put N=31 in (1) and (2)

f (x ) = 31−
1
x

xn +1 = 2 xn − 31xn2

1
Assume x0 =
30
Put n=0

x1 = 2 x0 − 31x02

2  1 
= − 31 
30  900 

=0.0322

Put n=1

x2 = 2 x1 − 31x12

= 2(0.0322 ) − 31(0.0322 )
2

=0.0323

Hence the root is 0.032

Practice problems

1. Using Newton-Raphson’s method, find the real root of x log10 x = 12.34 .

Ans:11.595

2. By using Newton-Raphson’s method, find the root of x 4 − x − 10 = 0

Ans:1.856

3. Find an iteration formula of N where N is the positive number and hence find
11

Ans:3.3170

Solution of a system of non-linear equations by Newton-Raphson method

To solve the system of equations f (x, y ) = 0, g (x, y ) = 0 in two unknowns x and y .

Method 1

1. Determine the functions f x , f y , g x , g y ,

2. Let the starting values of x and y be x0 and y 0 .

3. Using these values, find the values of f x , f y , g x , g y

fx fy gf y − fg y fg − gf x
4. Also find the values of J = = fxgy − f y gx , h = ,k= x
gx gy J J

5. Now set x1 = x0 + h, y1 = y 0 + k .

6. Using x1 and y1 , repeat steps 3 and 4 to get x2 and y2 .

7. Repeat till desired accuracy is achieved.

Method 2

Consider the formulae:

f ( xi , y i )
xi +1 = xi −
f x ( xi , y i )

g ( xi , y i )
y i +1 = y i −
g y ( xi , y i )

Iterate using the latest available values of x and y , till we get the desired accuracy.

Example1 (Method 1)

Using Newton-Raphson method, solve: x 2 + y − 11 = 0, y 2 + x − 7 = 0 starting with the approximate

solution x0 = 3.5 and y 0 = −1.5 .

Answer:

f (x, y ) = x 2 + y − 11
g ( x, y ) = y 2 + x − 7

f x = 2x
fy =1
gx = 1
g y = 2y
 fx fy
J= = f x g y − f y g x = (2 x )(2 y ) − (1) (1) = 4 xy
gx gy
gf y − fg y
h=
J
fg − gf x
k= x
J

We find the required values in the following table:

To find x1 , y1 To find x2 , y 2 To find x3 , y 3

x 0 = 3 .5 x1 = 3.5682 x 2 = 3.5845
y 0 = −1 . 5 y1 = −1.8864 y 2 = −1.8485
f = x 2 + y − 11 -0.25 -0.1543 0.0001

g = y2 + x − 7 -1.25 0.1267 0.0015

f x = 2x 7 7.1364 7.1690
fy =1 1 1 1
gx = 1 1 1 1
gy = 2y -3 -3.7728 -3.6970
J = 4 xy -22 -27.9242 27.5038
gf y − fg y 0.0682 0.0163 -0.0001
h=
J
fg − gf x -0.3864 0.0379 0.0004
k= x
J
xn+1 = xn + h 3.5682 3.5845 3.5844
y n+1 = y n + k -1.8864 -1.8485 -1.8481

Thus, x = 3.584, y = −1.848

Example 2 (Method 2)

Using Newton-Raphson method, solve:

f (x, y ) = y 2 + 4 x 2 + 2 xy − y − 2 = 0
g (x, y ) = y 2 + 2 x 2 + 3xy − 3 = 0

starting with x0 = 0.4 and y 0 = 0.9

Answer.
 f (x, y ) = y 2 + 4 x 2 + 2 xy − y − 2 = 0
g (x, y ) = y 2 + 2 x 2 + 3xy − 3 = 0

f x = 8x + 2 y
g y = 2 y + 3x

y 2 + 4 x 2 + 2 xy − y − 2
xi +1 = x −
8x + 2 y

y 2 + 2 x 2 + 3xy − 3
yi +1 = yi −
2 y + 3x

We iterate using these formulae, using the latest available values.

x y
0.4 0.9
0.5460 0.9348
0.5216 0.9695
0.5101 0.9857
0.5048 0.9932
0.5023 0.9967
0.5011 0.9984
0.5005 0.9993
0.5002 0.9997
0.5001 0.9999
0.5 1
0.5 1
The required solution is x = 0.5, y = 1

3. Using Newton-Raphson method, solve: x 2 + y 2 = 16, x 2 − y 2 = 4 starting with the approximate

(
solution 2 2 ,2 2 . )
4. Using Newton-Raphson method, solve:

f (x, y ) = y 2 + 4 x 2 + 2 xy − y − 2 = 0
starting with x0 = 0.4 and y 0 = 0.9
g (x, y ) = y 2 + 2 x 2 + 3xy − 3 = 0

5. Solve x 2 − y 2 = y, x 2 + y 2 = x, x0 = 0.8, y 0 = 0.4 , by Newton Raphson method.

6. Given (x0 , y 0 ) = (2, 0.25), compute (x1 , y1 ), (x2 , y 2 ), (x3 , y3 ) by solving
x 2 − 2 x − y + 0.5 = 0, x 2 + 4 y 2 = 4 .
Solution of a system of linear algebraic equations

Direct methods – Gauss elimination method, Gauss Jordan method

Iterative methods – Gauss Jacobi method, Gauss Seidel method

Gauss elimination method:

1. Write the given system of equations in matrix form as AX = B .

2. Form the augmented matrix

c

 
3. Reduce A B to an upper triangular matrix by elementary row transformations.

4. Rewrite in the form of equations and solve by back substitution.

Gauss Jordan method:

1. Write the given system of equations in matrix form as AX = B .

2. Form the augmented matrix A B 

 
3. Reduce A B to a diagonal matrix (or identity matrix) by elementary row
transformations.

4. Rewrite in the form of equations and solve directly.

Problems:

1. Apply Gauss Elimination and Gauss Jordan method to find the solution of the
following system 10 x + y + z = 12 , 2 x + 10 y + z = 13 , x + y + 5 z = 7

Solution:

Gauss Elimination method

The given system of equations can be written in matrix form Ax=B as

10 1 1  x  12 
    
 2 10 1  y  = 13 
 1 1 5  z   7 
    

The augmented matrix is

10 1 1 12 
 
( A, B ) =  2 10 1 13 
1 1 5 7
 

1 1 5 7
 
~  2 10 1 13  by R1  R3
10 1 1 12 
 

1 1 5 7 
  R2 → R2 − 2 R1
~ 0 8 − 9 − 1  by 
 0 − 9 − 49 − 58  R3 → R3 − 10 R1
 

1 1 5 7 
 
~ 0 8 −9 − 1  by R3 → 8R3 + 9 R2
 0 0 − 473 − 473 
 

This can be rewritten as

x + y + 5 z = 7 ..............(1)
8 y − 9 z = −1 ..............( 2)
− 473 z = −473 .........(3)

Solving using back-substitution.

(3)  −473 z = −473

z=1

( 2)  8 y − 9 z = −1

8 y − 9(1) = −1

y=1

(1)  x + y + 5 z = 7
x + 1 + 5(1) = 7

x=1

Hence the solution is x = 1, y = 1, z = 1 .

Gauss Jordan method

1 1 5 7 
 
~ 0 8 −9 −1 
 0 0 − 473 − 473 
 

1 1 5 7   R2
 9 1  R2 → 8
~ 0 1 − −  by 
 8 8  R → R3
0 0 1 1   3 − 473

1 1 5 7
   9
~  0 1 0 1  by R2 → R2 + R3
0 0 1 1  8
 

1 1 5 6
 
~  0 1 0 1  by R1 → R1 − R2
0 0 1 1
 

 1 0 0 1
 
~  0 1 0 1 by R1 → R1 − 5R3
 0 0 1 1
 

Hence the solution is x = 1, y = 1, z = 1 .

2. Apply Gauss Elimination and Gauss Jordan method to find the solution of the
following system x + 2 y + z = 3 , 2 x + 3 y + 3z = 10 , 3 x − y + 2 z = 13

Solution:

Gauss Elimination method

The given system of equations can be written in matrix form Ax=B as

 1 2 1  x   3 
    
 2 3 3  y  = 10 
 3 − 1 2  z  13 
    

The augmented matrix is

1 2 1 3 
 
( A, B ) =  2 3 3 10 
 3 − 1 2 13 
 

1 2 1 3
  R2 = R2 − 2 R1
~  0 − 1 1 4  by 
 0 − 7 −1 4  R3 = R3 − 3R1
 

1 2 1 3 
 
~ 0 −1 1 4  by R3 → R3 − 7 R2
 0 0 − 8 − 24 
 

This can be rewritten as

x + 2 y + z = 3 ..............(1)
− y + z = 4 ..............( 2)
− 8 z = −24 .........(3)

Solving using back-substitution.

(3)  −8 z = −24

z=3

( 2)  − y + z = 4

− y+3 = 4

y=-1

(1)  x + 2 y + z = 3

x + 2(− 1) + 3 = 3

x=2

Hence the solution is x = 2, y = −1, z = 3 .

Gauss Jordan method

1 2 1 3 
 
~  0 −1 1 4 
 0 0 − 8 − 24 
 

1 2 1 3 
 
~  0 1 − 1 − 4  by R2 → R2 (− 1)
 0 0 − 8 − 24 
 

1 0 3 11 
 
~  0 1 − 1 − 4  by R1 → R1 − 2 R2
 0 0 − 8 − 24 
 

 1 0 3 11 
   R
~  0 1 − 1 − 4  by R3 → 3
0 0 1  −8
 3 

1 0 0 2 
  R1 → R1 − 3R3
~  0 1 0 − 1 by 
0 0 1 3  R2 → R2 + R3
 

Hence the solution is x = 2, y = −1, z = 3 .

3. Solve the system of equation by Gauss Jordan method

x + y + z + w = 2 , 2 x − y + 2 z − w = −5 , 3x + 2 y + 3z + 4w = 7 , x − 2 y − 3z + 2w = 5

Solution:

The given system of equations can be written in matrix form Ax=B as

1 1 1 1  x   2 
    
 2 − 1 2 − 1 y   − 5 
3 2 =
3 4  z   7 
    
 1 − 2 − 3 2  w   5 
    

The augmented matrix is

 1 1 1 1 2 
 
2 −1 2 −1 − 5
( A, B ) = 
3 2 3 4 7 
 
1 − 2 − 3 2 5 


1 1 1 1 2 
   R2 → R2 − 2 R1
0 − 3 0 − 3 − 9 
~   by  R3 → R3 − 3R1
0 −1 0 1 1 R → R − R
   4
0 − 3 − 4 1 3  4 1


1 1 1 1 2
 
0 1 0 1 3  R
~   by  R2 → 3
0 −1 0 1 1  −3
 
1 − 3 − 4 
 1 3

1 0 1 0 − 1
   R1 → R1 − R2
0 1 0 1 3 
~   by  R3 → R3 + R2
0 0 0 2 4  R → R + 3R
   4
0 0 −4  4 2
 4 12 

1 0 1 0 − 1
 
0 1 0 1 3
~  by R3  R4
0 0 −4 4 12 
 
0 4 
 0 0 2

1 0 1 0 −1  R
  R3 → 3
0 1 0 1 3 
 −4
~   by 
0 0 1 −1 −3  R → R4
 
0 2   4 2
 0 0 1

1 0 0 1 2 
 
0 1 0 1 3 
~  by R1 → R1 − R3
0 0 1 −1 − 3
 
0 2 
 0 0 1
 1 0 0 0 0
   R1 → R1 − R4
0 1 0 0 1 
~   by  R2 → R2 − R4
0 0 1 0 −1 R → R + R
   3
0 2  3 4
 0 0 1

Hence the solution is x = 0, y = 1, z = 1, w = 2 .

Practice Problems:

Solve by Gauss Elimination and Gauss Jordan method

1. x + 2 y + z = 4 , 3x − y + 2 z = −3 , x + 2 y + 4 z = 7
Ans: x = -1 , y = 2 , z = 1

2. 2 x + y + z = 10 , 3x + 2 y + 3z = 18 , x + 4 y + 9 z = 16
Ans: x = 7 , y = -9 , z = 5

3. x + y + z = 9 , 2 x − 3 y + 4 z = 13 , 3x + 4 y + 5 z = 40
Ans: x = 1 , y = 3 , z = 5

4. 2 x − y + 3z = 8 , − x + 2 y + z = 4 , 3x + y − 4 z = 0
Ans: x = 2 , y = 2 , z = 2

5. Solve the system of equation by Gauss Jordan method

5x1 + x2 + x3 + x4 = 4 , x1 + 7 x2 + x3 + x4 = 12 , x1 + x2 + 6 x3 + x4 = −5 , x1 + x2 + x3 + 4 x4 = −6
Ans: x1 =1 ,x2 = 2 , x3 = -1 , x4=-2

Inverse of a matrix using Gauss Jordan method:

We can use Gauss Jordan method to find the inverse of a non-singular matrix:

1. Form the augmented matrix ( A, I ) where I is the identity matrix of the same order as A .

2. Transform ( A, I ) into the form (I, B ) using elementary row operations.

3. Then A −1 = B .
  1 1 2
 
1. Find the inverse of the matrix  1 2 3  by Gauss Jordan method.
2 3 1
 

 1 1 2
 
Let A=  1 2 3 
2 3 1
 

 1 1 2 1 0 0
 
(A / I ) =  1 2 3 0 1 0
 2 3 1 0 0 1
 

1 1 2 1 0 0
  R2 → R2 − R1
~  0 1 1 − 1 1 0  by 
0 1 − 3 − 2 0 1 R3 → R3 − 2 R1
 

1 0 1 2 −1 0
  R1 → R1 − R2
~  0 1 1 − 1 1 0  by 
0 0 − 4 −1 −1 1 R3 → R3 − R2
 

 
1 0 1 2 −1 0 
~ 0 1 1 −1 1 0  by R3 → − R3
1
 1 1 1  4
0 0 1 − 
 4 4 4

 7 5 1 
1 0 0 − 
 4 4 4 
 5 3 1   R1 → R1 − R3
~ 0 1 0 − by 
 4 4 4   R2 → R2 − R3
 1 1 1
0 0 1 − 
 4 4 4

 7 5 1 
 − 
 4 4 4 
Hence A =  −
−1 5 3 1 
 4 4 4 
 1 1 1
 − 
 4 4 4
  1 2 − 1
 
2. Find the inverse of the matrix  4 1 0  by Gauss Jordan method.
 2 −1 3 
 

Solution

 1 2 − 1
 
Let A =  4 1 0
 2 −1 3 
 

 1 2 − 1 1 0 0
 
(A / I ) =  4 1 0 0 1 0 
 2 − 1 3 0 0 1
 

 1 2 −1 1 0 0 
  R2 → R2 − 4 R1
~  0 − 7 4 − 4 1 0  by 
0 − 5 5 − 2 0 1 R3 → R3 − 2 R1
 

1 2 −1 1 0 0
 −4 4 −1   R → R2 / − 7
~ 0 1 0  by  2
 7 7 7  
 0 − 5 5 −2 0 1 

 1 −1 2 
1 0 0
 7 7 7 
−4 −1  R1 → R1 − 2 R2
~ 0 1 0  by 
4
 7 7 7   R3 → R3 + 5R2
 15 6 −5 
0 0 1
 7 7 7 

 1 −1 2 
1 0 0
 7 7 7  
−4 −1 
~ 0 1 
4
0 by  7
 7 7 7  R 3 → R3 *
 
6 −1 7   15
0 0 1 
 15 3 15 
  1 1 −1 
1 0 0 −   1
 5 3 15 
 R1 → R1 − R3
 4 −1 4   7
~ 0 1 0 by 
 5 3 15 
 R → R + 4 R
6 −1 7   2 2
7
3
0 0 1 
 15 3 15 

 −1 1 −1 
 
 5 3 15 
−1 4 
=
4
A −1
 5 3 15 
 2 −1 7 
 
 5 3 15 

 1 0 − 4
 
3. Find the inverse of the matrix  0 − 1 2  by Gauss Jordan method.
 −1 2 1 


Solution

 1 0 − 4
 
Let A =  0 − 1 2 
 −1 2 1 


1 0 − 4 1 0 0
 
(A / I ) =  0 − 1 2 0 1 0 
−1 2 1 0 0 1 


 1 0 − 4 1 0 0
 
~  0 −1 2 0 1 0 
 −1 2 1 0 0 1 


1 0 − 4 1 0 0
 
~ 0 −1 2 0 1 0  by R3 → R3 + R1
0 2 − 3 1 0 1 


1 0 − 4 1 0 0
 
~ 0 1 − 2 0 − 1 0  by R2 * (− 1)
0 2 − 3 1 0 1 

 1 0 − 4 1 0 0
 
~ 0 1 − 2 0 − 1 0  by R3 → R3 − 2 R2
0 0 1 1 2 1 


1 0 0 5 8 4
  R1 → R1 + 4 R3
~ 0 1 0 2 3 2  by 
0 0 1 1 2 1  R2 → R2 + 2 R3


 5 8 4
−1
 
A =  2 3 2
1 2 1
 

Practice Problems:

 1 1 3 
 
1. Find the inverse of  1 3 − 3  using Gauss Jordan method
− 2 − 4 4 
 

 7 −3 − 3
 
 2 2 2 
−1
Ans A −1 =  0 
1
 2 2 
 −1 1 
 0 
 2 2 

4 1 2
 
2. Find the inverse of  2 3 − 1 using Gauss Jordan method
1 − 2 2 
 

 −1 1 − 1
 
 5 3 15 
−1
Ans A = 
−1 4 4 
 5 3 15 
 2 −1 7 
 
 5 3 15 

0 1 1
 
3. Find the inverse of  1 2 0  using Gauss Jordan method
 3 −1 4 
 
  8 2 
 −1 
 3 3 
−4 − 1
Ans A −1 =  1
 3 3 
 7 1 
 −1 
 3 3 

Iterative methods (Indirect methods)

1. Gauss-Jacobi method

2. Gauss-Seidel method

Sufficient condition:

Each equation of the system must possess one large coefficient and the large coefficient must be attached
to a different unknown in each equation. This condition will be satisfied if the large coefficients are along
the leading diagonal of the coefficient matrix.

The system of equations

a11 x1 + a12 x 2 + a13 x3 = b1

a 21 x1 + a 22 x 2 + a 23 x3 = b2 can be solved by iterative method if:
a31 x1 + a32 x 2 + a33 x3 = b3

a11  a12 + a13

a22  a21 + a23 , i.e., the system is diagonally dominant
a33  a31 + a32

Gauss-Jacobi method

a1 x + b1 y + c1 z = d1
Consider a 2 x + b2 y + c 2 z = d 2 , which is diagonally dominant
a3 x + b3 y + c3 z = d 3

Then rewrite the equations as

 d1 − b1 y − c1 z
x=
a1
d 2 − a2 x − c2 z
y=
b2
d 3 − a3 x − b3 y
z=
c3

Use initial values x (0 ) , y (0 ) , z (0 ) and get x (1) , y (1) , z (1) . Use these values and get x (2 ) , y (2 ) , z (2 ) .
Proceed till we get the desired accuracy.

Gauss-Seidel method

Here we use initial values y (0 ) , z (0 ) and get x (1) . Then use x (1) , z (0 ) to get y (1) . Use x (1) , y (1) to get z (1) .
Continue by using the latest available values every time, till we get the desired accuracy.

Note 1: The current values of the unknowns at each stage of iteration are used to find the forthcoming
values , hence Gauss-Seidel method converges faster than Gauss-Jacobi method. The convergence in
Gauss-Seidel method is nearly two times that of Gauss-Jacobi method.

Note 2: Iteration method is a self-correcting method. i.e., any error made in computation is corrected in
the subsequent iterations.

Problems

1. Solve by Gauss-Jacobi method:

10 x + 2 y + z = 9
x + 10 y − z = −22
− 2 x + 3 y + 10 z = 22

Answer:

10  2 + 1
10  1 + 1
10  2 + 3

The given system of equations is diagonally dominant. So we rewrite the equations as

x=
1
(9 − 2 y − z )
10
y = (− 22 − x + z )
1
10
z = (22 + 2 x − 3 y )
1
10
We start with initial values x = 0, y = 0, z = 0 and iterate, using the values obtained in the previous step

Iteration 1 1 1

x =  (9 − 2 y − z ) y =  (− 22 − x + z ) z =  (22 + 2 x − 3 y )
 10   10   10 
Initial 0 0 0
1 0.9 -2.2 2.2
2 1.1200 -2.0700 3.0400
3 1.0100 -2.0080 3.0450
4 0.9971 -1.9965 3.0044
5 0.9989 -1.9993 2.9984
6 1.0000 -2.0000 2.9996
7 1.0000 -2.0000 3.0000
The required solution is x = 1, y = −2, z = 3

Iteration 1 1 1

x =  (9 − 2 y − z ) y =  (− 22 − x + z ) z =  (22 + 2 x − 3 y )
 10   10   10 
Initial 0 0 0
1 0.9 -2.2 2.2
2 1.12000 -2.07000 3.04000
3 1.01000 -2.00800 3.02640
4 0.99896 -1.99650 3.00440
5 0.998861 -1.99945 2.99837
 6 1.00005 -2.00004 2.99961
7 1.00005 -2.00004 2.99961

X=1,y=-2,z=3

2. Solve using Gauss Seidel method

8 x − 3 y + 2 z = 20
6 x + 3 y + 12 z = 35
4 x + 11 y − z = 33

Answer: The given system of equations is not diagonally dominant. So we rearrange the given system
of equation in such a way that it is diagonally dominant.

8 x − 3 y + 2 z = 20
4 x + 11 y − z = 33 .
6 x + 3 y + 12 z = 35

83+2
11  4 + 1
12  6 + 3

Then we rewrite them as:

x=
1
(20 + 3 y − 2 z )
8
y = (33 − 4 x + z )
1
11
z = (35 − 6 x − 3 y )
1
12

We start with initial values y = 0, z = 0 and iterate, using the latest available values everytime.

Iteration
x=
1
(20 + 3 y − 2 z ) y=
1
(33 − 4 x + z ) z=
1
(35 − 6 x − 3 y )
8 11 12
Initial - 0 0
1 2.5 2.091 1.1439
2 2.9244 2.0674 0.9376
3 3.0409 1.9795 0.9013
4 3.0170 1.9848 0.9120
5 3.0163 1.9861 0.9120
6 3.0168 1.9859 0.9118
7 3.0168 1.9859 0.9118
Hence the required solution is x = 3.017, y = 1.986, z = 0.912

3. Solve by Gauss Jacobi method and Gauss Seidel method:

x + y + 54 z = 110
6 x + 15 y + 2 z = 72
27 x + 6 y − z = 85

Solution:

27𝑥 + 6𝑦 − 𝑧 = 85

6𝑥 + 15𝑦 + 2𝑧 = 72

𝑥 + 𝑦 + 54𝑧 = 110

|27| > |6| + |1|

x=
1
(85 − 6 y + z )
27
y = (72 − 6 x − 2 z )
1
15
z = (110 − x − y )
1
54
 Iteration
x=
1
(85 − 6 y + z ) y=
1
(72 − 6 x − 2 z ) z=
1
(110 − x − y )
27 15 54
Initial 0 0
1 3.1481 Sub x=3.1481,z=0 Sub
y=3.5408 x=3.1481,y=3.5408
Z=1.91317
2 Sub Sub Sub
y=3.5408,z=191317 x=2.4322,z=1.91317 x=2.4322,y=3.5720
X=2.4322 Y=3.5720 Z=1.92585
3 2.4257 3.57294 1.92595
4 2.4255 3.5730 1.9260
5 2.4255 3.5730 1.9260
6
7
8
9
10

X=2.426,y=3.573,z=1.926

Gauss Jacobi method:

The given system of equations is not diagonally dominant. So we rearrange the given system of equation
in such a way that it is diagonally dominant.

27 x + 6 y − z = 85
6 x + 15 y + 2 z = 72
x + y + 54 z = 110

Then we rewrite them as

x=
1
(85 − 6 y + z )
27
y = (72 − 6 x − 2 z )
1
15
z = (110 − x − y )
1
54

We start with initial values x = 0, y = 0, z = 0 and iterate, using the values obtained in the previous step
 Iteration
x=
1
(85 − 6 y + z ) y=
1
(72 − 6 x − 2 z ) z=
1
(110 − x − y )
27 15 54
Initial 0 0 0
1 3.1481 4.8000 2.0370
2 2.1569 3.2692 1.8899
3 2.4917 3.6854 1.9366
4 2.4009 3.5451 1.9226
5 2.4316 3.5833 1.9269
6 2.4232 3.5704 1.9257
7 2.4260 3.5740 1.9260
8 2.4253 3.5728 1.9259
9 2.4255 3.5731 1.9260
10 2.4255 3.5731 1.9260

The required solution is x = 2.426, y = 3.573, z = 1.926

Guass seidel method

We start with initial values y = 0, z = 0 and iterate, using the latest available values everytime.

Iteration
x=
1
(85 − 6 y + z ) y=
1
(72 − 6 x − 2 z ) z=
1
(110 − x − y )
27 15 54
Initial - 0 0
1 3.1481 3.5408 1.9132
2 2.4322 3.5720 1.9260
3 2.4257 3.5729 1.9260
4 2.4255 3.5730 1.9260
Hence the required solution is x = 2.426, y = 3.573, z = 1.926

Practice Problems

Solve, using Gauss Jacobi and Gauss Seidel method:

1. 28 x + 4 y − z = 32, x + 3 y + 10 z = 24, 2 x + 17 y + 4 z = 35
Ans: x= 0.994 , y = 1.507 , z = 1.849

2. 10 x − 5 y − 2 z = 3 ,4 x − 10 y + 3z = −3 , x + 6 y + 10 z = −3
Ans: x= 0.342 , y =0.285 , z = -0.505
 3. 6 x + 3 y + 12 z = 36, 8 x − 3 y + 2 z = 20, 4 x + 11y − z = 33
Ans: x= 3 , y = 2 , z = 1

4. 8 x − 3 y + 2 z = 20, 4 x + 11y − z = 33, 6 x + 3 y + 12 z = 35

Ans: x= 3.017 , y =1.986 , z = 0.912

Power method to find the numerically largest eigenvalue and corresponding eigenvector

1. Let A be a matrix and X 1 be an arbitrary vector chosen as the initial approximation of the required
eigenvector (with a component as 1) .

2. Find AX 1 and write it as AX 1 = 1 X 2 (by taking out the largest component 1 of X 1 ). Now 1 is
the approximate eigen value and X 2 is the corresponding approximate eigenvector.

3. Form the sequence of iterations AX 1 = 1 X 2 , AX 2 = 2 X 3 ,AX 3 = 3 X 4 ……..

4. This sequence converges to give the dominant eigenvalue and the corresponding eigenvector.

Note:

1. To find numerically smallest eigen value of A obtained the dominant eigen value 1

of A and then find B= A- 1 I and find the dominant eigen value of B

2. Then the second eigen value of A is equal to the dominant eigen value of B+ 1
1 0 0 1
       
3. Then we select the initial eigen vector be  0  1 0 1
0 0 1 1
  ,   ,   ,  

Problems

1. Use power method to find the dominant eigenvalue and the corresponding eigenvector of the

 4 2
matrix A =   . Also find the other eigenvalue.
1 3

Solution:

 4 2
A =  
Given 1 3
 1
Let the initial vector be X 1 =   . Then
 0

 4 2  1   4   1 
AX 1 =    =   = 4 
 1 3  0   1   0.25 

 1 
The first approximation to the eigenvalue is 4 and the corresponding eigenvector is  
 0.25 

 1 
Let X 2 =  
 0.25 

 4 2  1   4.5   1 
AX 2 =    =   = 4.5 
 1 3  0.25  1.75   0.39 

 4 2  1   4.78   1 
AX 3 =    =   = 4.78 
 1 3  0.39   2.17   0.45 

 4 2  1   4.9   1 
AX 4 =    =   = 4.9 
 1 3  0.45   2.35   0.48 

 4 2  1   4.96   1 
AX 5 =    =   = 4.96 
 1 3  0.48   2.44   0.49 

 4 2  1   4.98   1 
AX 6 =    =   = 4.98 
 1 3  0.49   2.47   0.49 

 4 2  1   4.98   1 
AX 7 =    =   = 4.98 
 1 3  0.49   2.47   0.49 

 1 
Hence the dominant eigenvalue of A is 1 =4.98 and the corresponding eigenvector is  
 0.49 

Sum of eigen values = Trace of the matrix.

4.98 + 2 = 7  2 = 2.02

 25 1 2 
 
2. Find the numerically largest eigenvalue of A =  1 3 0  and the corresponding eigen vector.
 2 0 − 4
 
Solution:

 25 1 2 
 
Given A =  1 3 0 
 2 0 − 4
 

1
 
Let X 1 =  0 
0
 

 25 1 2  1   25   1 
      
AX1 =  1 3 0  0  =  1  = 25 0.04 
 2 0 − 4  0   2   0.08 
      

 25 1 2  1   25.2   1 
      
AX 2 =  1 3 0  0.04  = 1.1200  = 25.2 0.0444 
 2 0 − 4  0.08  1.6800   0.0667 
      

 25 1 2  1   25.1778   1 
      
AX 3 =  1 3 0  0.0444  =  1.1332  = 25.1778 0.0450 
 2 0 − 4  0.0667   1.7332   0.0688 
      

 25 1 2  1   25.1826   1 
      
AX 4 =  1 3 0  0.0450  =  1.1350  = 25.1826 0.0451
 2 0 − 4  0.0688   1.7248   0.0685 
      

 25 1 2  1   25.1821  1 
      
AX 5 =  1 3 0  0.0451 =  1.1353  = 25.1821 0.0451
 2 0 − 4  0.0685   1.7260   0.0685 
      

 25 1 2  1   25.1821  1 
      
AX 6 =  1 3 0  0.0451 =  1.1353  = 25.1821 0.0451
 2 0 − 4  0.0685   1.7260   0.0685 
      

 1 
 
Hence the dominant eigenvalue of A is 1 = 25.1821 and the corresponding eigenvector is  0.0451 
 0.0685 
 
 1 6 1
 
3. Find the dominant eigenvalue and the corresponding eigenvector of A =  1 2 0  by power method.
 0 0 3
 
Find the other eigenvalue also.

Solution:

1 6 1
 
Given A =  1 2 0 
 0 0 3
 

1
 
Let X 1 =  0 
0
 

 1 6 1  1   1   1 
      
AX1 =  1 2 0  0  =  1  = 1 1 
 0 0 3  0   0   0 
      

 1 6 1  1   7   1 
      
AX 2 =  1 2 0  1  =  3  = 7 0.4286 
 0 0 3  0   0   0 
      

 1 6 1  1   3.5716   1 
      
AX 3 =  1 2 0  0.4286  =  1.8572  = 3.5716 0.52 
 0 0 3  0   0   0 
      

 1 6 1  1   4.12   1 
      
AX 4 =  1 2 0  0.52  =  2.04  = 4.12 0.4951
 0 0 3  0   0   0 
      

 1 6 1  1   3.9706   1 
      
AX 5 =  1 2 0  0.4951 =  1.9902  = 3.9702 0.5012 
 0 0 3  0   0   0 
      
  1 6 1  1   4.0072   1 
      
AX 6 =  1 2 0  0.5012  =  2.0024  = 4.0072 0.4997 
 0 0 3  0   0   0 
      
 1 6 1  1   3.9982   1 
      
AX 7 =  1 2 0  0.4997  =  1.9994  = 3.9982 0.5001
 0 0 3  0   0   0 
      

 1 6 1  1   4.0006   1 
      
AX 8 =  1 2 0  0.5001 =  2.0002  = 4.0006 0.5000 
 0 0 3  0   0   0 
      

 1 6 1  1   4   1 
      
AX 9 =  1 2 0  0.5  =  2  = 4 0.5 
 0 0 3  0   0   0 
      

 1 
 
Hence the dominant eigenvalue of A is 1 = 4 and the corresponding eigenvector is  0.5 
 0 
 

To find the least eigenvalue B is B = A − 1I

1 6 1  4 0 0
   
B = 1 2 0 −  0 4 0
 0 0 3  0 0 4
   

− 3 6 1
 
B= 1 −2 0 
 0 0 − 1


1
 
Let Y1 =  0 
 0
 

− 3 6 1  1   − 3   1 
      
BY1 =  1 − 2 0  0  =  1  = −3 − 0.3333 
 0 0 − 1 0   0   
  0 
 − 3 6 1  1   − 4.9998   1 
      
BY2 =  1 − 2 0  − 0.3333  =  1.6666  = −4.9998 − 0.3333 
 0 0 − 1     
 0   0   0 

− 3 6 1  1   − 4.9998   1 
      
BY3 =  1 − 2 0  − 0.3333  =  1.6666  = −4.9998 − 0.3333 
 0 0 − 1     
 0   0   0 

Hence the dominant eigenvalue =-5

Second eigenvalue of A = Dominant eigenvalue of B + 1

= -5+4

= -1

Sum of the eigen value = Trace A

1 + 2 + 3 = 1 + 2 + 3

4 − 1 + 3 = 6
3 = 3

Therefore the eigenvalues are 4 , -1 and 3

Practice problems

5 0 1
 
1. Using power method, find all the eigenvalues of A =  0 − 2 0 
1 0 5
 

Using power method, find the dominant eigenvalue and the corresponding eigenvector of

2.
 2 −1 1 
 
A =  −1 2 −1
 1 −1 2 
 
 5 0 1 1 5 1 1
[0 −2 0] [0]=[0]=5[ 0 ], x2=[ 0 ]
1 0 5 0 1 0.2 0.2
5 0 1 1 5.2 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.2[ 0 ], x3=[ 0 ]
1 0 5 0.2 2 0.384 0.384
5 0 1 1 5.384 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.384[ 0 ],x4=[ 0 ]
1 0 5 0.384 2.92 0.54 0.54
5 0 1 1 5.54 1 1
[0 −2 0] [ 0 ]=[ 0 ]= 5.54[ 0 ], x5=[ 0 ]
1 0 5 0.54 3.7 0.67 0.67
5 0 1 1 5.67 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.67[ 0 ], x6=[ 0 ]
1 0 5 0.67 4.35 0.77 0.77
5 0 1 1 5.77 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.77[ 0 ], x7=[ 0 ]
1 0 5 0.77 4.85 0.84 0.84
5 0 1 1 5.84 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.84[ 0 ], x8=[ 0 ]
1 0 5 0.84 5.2 0.89 0.89
5 0 1 1 5.89 1 1
[0 −2 0] [ 0 ]=[ 0 ]= 5.89[ 0 ], x9=[ 0 ]
1 0 5 0.89 5.45 0.92 0.92
5 0 1 1 5.92 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.92[ 0 ], x10=[ 0 ]
1 0 5 0.92 5.6 0.95 0.95
5 0 1 1 5.92 1 1
[0 −2 0] [ 0 ]=[ 0 ]=5.92[ 0 ], x11=[ 0 ]
1 0 5 0.94 5.6 0.95 0.95
1
λ=5.92, [ 0 ]=x
0.95