MATRIX INVERSION AND EIGEN

VALUE PROBLEMS
 CROUT’S METHOD
To find the inverse of the square matrix A. Decompose A in to
A=LU where L is a lower triangular matrix and U is an unit upper
triangular matrix. A-1=(LU)-1 =U-1 L-1 .
If L is lower triangular, then L-1 is also lower triangular. Also if U is
L
upper triangular, then U-1 is also upper triangular. SinceL L-1=I we can
find the lower triangular matrix L-1 such that L L-1=I .
Since U U-1 = I ,we can find the upper triangular matrix U-1 such that
U U-1 = I . Having known L-1 and U-1 , we get A-1 =U-1 L-1 .
Decomposition of A in to A= LU .
Let us consider
 a11 a12 a13   l11 0 0  1 u12 u13 
 
A   a21 a22 a23  , L  l21 l22 0  andU  0 1 u23 
 a31 a32 a33  l31 l32 l33  0 0 1 
  l11 0 0  1 u12 u13   a11 a12 a13 
l l22 0  0 1 u23    a21 a22 a23 
 21
l31 l32 l33  0 0 1   a31 a32 a33 
l11 l11u12 l11u13   a11 a12 a13 
l l u  l l u  l u   a a a 
 21 21 12 22 21 13 22 23   21 22 23 

l31 l31u12  l32 l31u13  l32u23  l33   a31 a32 a33 

Equating coefficients and simplifying, we have,

a12 a
l11  a11 , l21  a21 , l31  a31 , u12  , u13  13 , l22  a22  l21u12 ,
a11 a11
a23  l21u13
l32  a32  l31u12 , u23  , l33  a33  l31u13  l32u23. .
l22
EXAMPLE FOR FINDING INVERSE OF A MATRIX BY CROUT’S METHOD.
Find the inverse of the matrix A by using crout’s method where
1 2 3
A   0 1 4
 2 2 0 
SOLUTION:
Let us decompose A as LU.
 l11 0 0  1 u12 u13   1 2 3
l l22 0  0 1 u23    0 1 4
 21
l31 l32 l33  0 0 1   2 2 0 

Equating the corresponding elements both sides,

a12 a
l11  1, l21  0, l31  2, u12   2, u13  13  3, l22  a22  l21u12  1
l11 l11
a23  u13l21 40
l32  a32  u12l31  2  4  2, u23    4,
l22 1
l33  a33  l31u13  l32u23  2.

 1 0 0  1 2 3 
L
 0 1 0  , U  0
 1 4 


 2 2 2 
 
0 0 1 

To find L-1 , LL-1 =I. Therefore

 1 0 0   x11 0 0  1 0 0
0 1 0   x21 x22 0   0 1 0 

 2 2 2   x31 x32 x33  0 0 1 
x11  1, x22  1, x33  1/ 2, x21  0, x11  x21  x31  0, x22  x32  0,
x31  1, x32  1.
1 0 0 
L1   0 1 0 
 1 1 1/ 2 
To find U-1 ,we have U U-1 =[Link],
1 2 3  1 c12 c13  1 0 0
0 1 4  0 1 c  0 1 0
  32   

0 0 1 0 0 1   
0 0 1

Equating elements,
c12  2  0, c13  2c23  3  0, c23  4  0, c12  2, c23  4, c13  5.
1 2 5
U 1  0 1 4
 

0 0 1

Since A=LU,A-1=(LU)-1=U-1L-1

1 2 5   1 0 0   4 3 5 / 2
A1  0 1 4  0 1 0    4 3 2 
0 0 1   1 1 1/ 2  1 1 1/ 2 

PROBLEMS FOR PRACTICE:

Find the inverse of the following matrices using crout’s method.

.  1 1 1 
 4 4 4 
3 2 4  
(1).  1  9 11 5 
2 1  ........................ ANS :
 8 8 8 
 5  5
1  1 
3
 7 

 8 8 8 
 1 3 1 
 10 10 5 
1 0 1  
7 2 
(2)  
5  ..................... ANS : 
21
3 4
 10 20 5 

0 6 7   
 9 3 1 

 10 10 5 
 INVERSION OF A MATRIX
GAUSS ELIMINATION METHOD
Consider a non singular matrix A of order [Link] X is the inverse of
A then AX=I, where I is the unit matrix of order [Link]
 a11 a12 a13   x11 x12 x13 
A   a21 a22 a23  andX   x21 x22 x23 
 a31 a32 a33   x31 x32 x33 
Therefore AX=I reduces to
 a11 a12 a13   x11 x12 x31  1 0 0
a a22 a23   x21 x22 x23   0 1 0
 21
 a31 a32 a33   x31 x32 x33  0 0 1 

This equation is equivalent to the nine equations given below:

 a11 a12 a13   x11  1 
 
 a21 a22 a23   x21    0 .................................................(1)
a a33   x31   0
 31 a32
 a11 a12 a13   x12  0
 
 a21 a22 a23   x22   1  ...................................................(2)
a a33   x32  0
 31 a32
  a11 a12 a13   x13  0
 
 a21 a22 a23   x23   0 .....................................................(3)
a a33   x33  1
 31 a32
From equations (2),(3)and (4), we can solve for the vectors

 x11   x12   x13 

 x  ,  x  and  x 
 21   22   23 
 x31   x32   x33 

using Gauss elimination method. The solution set of each systems

(2),(3),(4) will be the corresponding column of the inverse matrix X.
{solution of system (1) by Gauss elimination method.
Now the system (1) is equivalent to AX=B, where
 a11 a12 a13   x11  1 
 
A   a21 a22 a23  , X   x21  andB  0
a a33 
 31 a32  x31  0

Now our aim is to reduce the augmented matrix (A:B) to an upper

triangular matrix
  a11 a12 a13 1
( A : B)   a21 a22 a23 0 
 a31 a32 a33 0 
 ai1
Now we multiply the first row by and add to the ith row of
a11

(A:B) where i=2,[Link] this all elements in the first column of (A:B)
except a11 are made zero. Now (A:B) takes the form
 a11 a12 a13 1 
( A : B)   0 b22 b23 c 
 0 b32 b33 d 
Now take the pivot b22 . Now considering b22 as pivot we will
make b32 in the second column of (A:B) as zero. Now multiply the
second row by b32 and add to the corresponding element of the
b22

3rd row. Now b32 is reduced to zero. Now (A:B) takes the form
  a11 a12 a13 1
( A : B)   0 b22 b23 c 
 0 0 c33 e 

Now we have the following three equations.

a11 x11  a12 x21  a13 x31  1
b22 x21  b23 x31  c
c32 x31  e

By back substitution method, we get the values of x11,x21,x31. }

Similarly we can solve the system of equations (2)&(3) and get the
values of x12,x22,x32,x13,x23,x33 .By this way we find the inverse of a
given square matrix A by Gauss elimination method.
0 1 1
EXAMPLE A  5 2 0 
QN: By Gaussian elimination find the inverse of  3 1 4 
 SOLUTION : Now( A : I )is
0 1 1 1 0 0 
( A : I )  1 2 0 0 1 0 
3 1 4 0 0 1 

Since the element a11 is zero we will interchange the first and the
second row. The reduced system is
1 2 0 0 1 0
0 1 1 1 0 0 

3 1 4 0 0 1 
By performing R3+(-3)R1,we get
1 2 0 0 1 0
0 1 1 1 0 0 
 
0 7 4 0 3 1 
By performing R3+7R2 , we get
1 2 0 0 1 0
0 1 1 1 0 0
 

0 0 3 7 3 1

7
x31 
x11  2 x21  0  3
 4
x21  x31  1   x21 
 3
3 x31  7  8
x11 
3
x12  2 x22  1 x32  1

x22  x32  0   x22  1
3x32  3   x12  1
1
x33 
x13  2 x23  0  3
 1
x23  x33  0   x23 
 3
3 x33  1 
2
x13 
3
  8 2 
 3 1
3 
 
Hence A1 
4
1
1 
 3 3 
 
 7 1
1 

 3 3 
GAUSS JORDAN METHOD
This method is a modification of the above Gauss elimination
method. In this method, we consider the augmented matrix
(A:B) to be  a11 a12 a13 1 0 0
( A : I )   a21 a22 a23 0 1 0 
 a31 a32 a33 0 0 1 
The coefficient matrix A can be brought to a unit matrix by
making all the elements above and below the leading diagonal are
zeros and all the leading diagonal elements are 1’s. Then (A:I)
takes the form
1 0 0 a b c 
( A : I )  0 1 0 d e f 
0 0 1 g h i 
Now the inverse of A is
a b c
A1   d e f 
 g h i 

EXAMPLE
Find the inverse of the matrix A using Gauss Jordan method where
2 1 1
A   3 2 3
1 4 9 
SOLUTION:
Consider (A:I) to be 2 1 1 1 0 0
( A : I )   3 2 3 0 1 0 
1 4 9 0 0 1 
Add multiples of first row to the other rows so that the elements
in the second row first column and the third row first column
become zero.
  
2 1 1 1 0 0 3
  R2  R2  R1
0 1 3 3
0
2
 1
 2 2 2  1
  3
R  R3  R1
7 17 1 2
0  0 1
 2 2 2 

Make the element in the third row and the second column as
zero by using the element in the second row and the second
column. 2 1 1 1 0 0
 1 3 3 
0 1 0  R3  R3  7 R2
 2 2 2 
 0 0 2 10 7 1 
 
Make the element in the second row and the third column zero
by using the element in the third row and the third column .
2 1 1 1 0 0
 1 17 3  3
0 0 6 R2  R2  R3
 2 4 4 4
0 0 2 10 7 1 

Make the element in the first row and third column to
zero by using the element in the third row and third
column.
 7 1
2 6
1 0 2 2
 
0 1 17 3 1
0 6 R1  R1  R3
 2 4 4 2
0 0 2 10 7 1 

 
Make the element in the first row and second column to zero by
using the element in the second row and second column.

2 0 0 6 5 1
 1 17 3  1
0 0 6 R1  R1  R3
 2 4 4 2
0 0 2 10 7 1 

Make all the diagonal elements in the coefficient matrix to unity.
Then (A:I) becomes,
  5 1 
 3  R1
2 2 R 
1 0 0  1
2
0 17 3
1 0 12 R2  R2 * 2
 2 2 
0 0 1
7 1  R3  R3  1 
 5
 2 2   2 

Hence the inverse of the given matrix A by Gauss Jordan

method is given by,  5 1 

 3 2 2 
 
17
 12
3 
A1
 2 2 
 7 1 
 5
 2 2 
PROBLEMS FOR PRACTICE:
• Find the inverse of the following matrices by both
Gauss elimination and Gauss Jordan methods:
3 10 1 
4 4 4
2 4 3  
1
(1)  0 1  ............................ ANS : 
1 
1 2
2 2
 2 1 1 1 
2
 1
 2 2 
1 3 7 
4 4 4 
3 1 2  
1 1
(2)  2 
3  ............................. ANS : 
5 
5
4 4 4 
 1 1  1 13 
2
 5
 4 4 4 
 EIGEN VALUES AND EIGEN VECTORS
If A is any square matrix of order ‘n’ with elements aij we can find
a column matrix X such that AX=λ X (or) (A- λ I)X=[Link] matrix
equation represents ‘n’ homogeneous linear equations,
( a11   ) x1  a12 x2  .........................  a1n xn  0
a21 x1  ( a22   ) x2  .........................  a1n xn  0
.
.
.
an1 x1  an 2 x2  .........................  ( ann   ) xn  0

which will have a non trivial solution only if the coefficient

determinant vanishes.
a11   a12 a1n
a21 a22   a2 n
 0.....................(2)

an1 an 2 ann  
On expansion it gives an nth degree equation in λ called
the characteristic equation of the matrix [Link] roots λi ,
(i=1,2,………..n) are called the eigen values (or) latent roots
and corresponding to each eigen value, equation (2) will have non
zero solution X=[x1 x2 xn ]T which is known as the eigen
vector.
POWER METHOD
In many engineering problems it is required to compute the
numerically largest eigen value and the corresponding eigen
[Link] such cases, the following iterative method can be used. If
X1,X2,……………Xn be the eigen vectors corresponding to the eigen
values λ1 , λ2 , λn then an arbitrary column vector can be
written as, X  k1 X1  k2 X 2  ..................  kn X n .
Then
AX  k1 ( AX 1 )  k2 ( AX 2 )  ...................  kn ( AX n ).
AX  k1 (1 X 1 )  k2 (2 X 2 )  ...................  kn (n X n ).
A2 X  k1 (12 X 1 )  k2 (2 2 X 2 )  ...................  kn (n 2 X n ).
Then Ar X  k11r X 1  k2 2 r X 2  ..................  kn n r X n
  2 
r
 n  
r

A X  1  k1 X 1  k2    .................  kn   
r r

  1   1  

As r increases, Ar X =λ1r k1X1 , since

1  2  ...................................... n .
Therefore, everytime we multiply the column vector by A, it
becomes nearer to the eigen vector X1 . Then we make the largest
component of the resulting column vector unity to avoid the
the factor k1 .
Thus start with a column vector X which is as near the solution
as possible and evaluate AX which is written as  (1) X (1)
after normalisation.
.This gives the first approximation λ(1) to the eigen value
and X(1) to the eigen vector. Similarly we evaluate
AX(1) = λ(2) X(2) which gives the second approximation. We repeat
this process till X(r) - X(r-1) becomes negligible. Then λ(r) will be the
largest eigen value and X(r) ,the corresponding eigen vector.
This iterative procedure for finding the dominant eigen value of a
matrix is known as Rayleigh’s power method.
1
-1 -1
NOTE: Rewriting AX=λX as X= λA X, we have A X=  X. If we use
this equation the above method yields the smallest eigen value.

EXAMPLE: Obtain by power method, the numerically dominant

eigen value and eigen vector of the matrix

 15 4 3
A   10 12 6
 20 4 2
 SOLUTION
Let the initial approximation to the eigen vector be X=[1 1 1]T .Then
 15 4 3 1  8   0.444
AX  10 12 6 1   4   18  0.222   18 X (1)
 20 4 2 1  18  1 
So,the first approximation to the eigen value is -18 and the
corresponding eigen vector is [-0.444 0.222 1]T .Now
 15 4 3 0.444  1 
AX (1)  10 12 6  0.222   10.548  .105    (2) X (2)
20 4 2  1  0.736
The second approximation to the eigen value is -10.548 and the
eigen vector is [1 -0.105 -0.736]T . Repeating the above process,
 0.93
AX (2)  18.948  0.361   (3) X (3)
 1 
  1 
AX (3)  18.394  0.415    (4) X (4)
 0.981
 0.995
AX (4)  19.698  0.462    (5) X (5)
 1 
 1 
AX (5)  19.773  0.480    (6) X (6)
 0.999 

 0.997 
AX (6)  19.922  0.49    (7) X (7).
 1 
 1 
AX (7)  19.956  0.495    (8) X (8).
 0.999 
Since λ(7) = λ(8) and X(7) =X(8) approximately the dominant
eigen value and the corresponding eigen vector are given by
 1   1 
 (8) X (8)  19.956 0.495 (or )20 0.5
0.999  1 
PROBLEMS FOR PRACTICE:
Find the dominant eigen values and the corresponding eigen
vectors of the following matrices using power method.
1 1 0   0.5
(1)  1 2 1 ....................... ANS :3,  1 
 0 1 1   0.5
 1 1 2  0.25 
(2)  0 3 1 ........................ ANS :5,  0.5
 0 0 5   1 
 JACOBI METHOD

Let A be a given real symmetric matrix. Its eigen values are real
and there exists a real orthogonal matrix B such that B-1AB is a
diagonal matrix D. Jacobi method consists of diagonalising A by
applying a series of orthogonal transformations B1,B2,………..Br such
that their product B satisfies the equation B-1AB=D. For this
purpose,we choose the numerically largest non-diagonal element aij
and form a 2*2 sub matrix,
 aii aij 
A1   whereaij  a ji
 a ji a jj 
which can be easily diagonalised.
• Consider an orthogonal matrix, cos   sin  
B1   
 sin  cos  

• so that B1-1 =B1T .

Then,
1  cos  sin    aii aij  cos   sin  
B A1 B1   
1
  sin  cos    a ji a jj   sin  cos  
 1 
 aii cos 2
  a jj sin 2
  aij sin 2 ai j cos 2  (a jj  aii ) sin 2 
B11 A1 B1   2
 .........(1)
 a cos 2  1 (a  a ) sin 2 aii sin 2   a jj cos 2   aij sin 2 
 ij 2
jj ii

Now this matrix will reduce to the diagonal form,if
1
aij cos 2  ( a jj  aii ) sin 2  0
2
2aij
tan 2  .......................(2)
aii  a jj

This equation gives four values of Ө , but to get the least possible
rotation , we choose -π/4 ≤ Ө ≤ π/4 .Thus (1) reduces to a diagonal
matrix.
As a next step, the largest non diagonal element (in magnitude) in
the new rotated matrix is found and the above procedure is
repeated using the orthogonal matrix B2
In this way , a series of a series of such transformations are
performed so as to annihilate the non diagonal [Link]
making ‘r’ transformations we obtain
Br 11 Br 11...................B11 AB1...............Br 1 Br  B 1 AB
As r→∞, B-1 AB approaches a diagonal matrix whose diagonal
elements are the eigen values of A. Also the corresponding
columns of B=B1B2B3…………Br are the eigen vectors of A.
EXAMPLE: Using jacobi’s method, find all the eigen values and
eigen vectors of the matrix   1 2 2
 
A 2 3 2
 
 2 2 1 
SOLUTION: Here the largest non diagonal elements is a13=a31=2.
Also a11=1,a33=1.
2a13 2* 2 
tan 2    . 
a11  a33 1  1 4
Then
 1 1 
0 
 cos  0  sin    2 2
 
B1   0 1 0    0 1 0  & B11  BT
 sin  0 sin    1 1 
 0 
 2 2 

The first approximation gives

 1 1   1 1 
 2 0  2  0  
2 1 2
 2 2 3 2 0 
   
1
D1  B1 AB1   0 1 0  2 3 2  0 1 0   2 3 0 
 1 1   2 
1   1 1   0 0 1
 0  2 
 2 2   2 2 
Now the largest non diagonal element is a21=a12=2,a11=3,a22=3.
2a12 4 
tan 2    . 
a11  a22 0 4
Then  1 1 
  0
 cos   sin  0  2 2

 1 
B2   0
1
 sin  cos  
  2 2
0

 0 0   0
1 
 0 1

 


The second transformation gives

 1 1   1 1 
 2 0   0
2 2 2
  3 2 0   5 0 0
 1   
0   2 3 0   0   0 1 0 
1 1 1
B2 1 D1 B2   
2 2 2 2
  0 0 1   0 0 1
 0 0 1   0 0 1 
   

Hence the eigen values of the given matrix are 5,1,-1 and the
corresponding eigen vectors are the columns of
 1 1   1 1 1 
 1 1   0   
 0  2 
2
2 2 2 2
2  
  1 1   1 1 
B  B1 B2   0 1 0  0   0 
 1 1 
2 2
 
2 2

 0  0 0 1  1 1 1 
 2     2 
2 
2
 2
DISADVANTAGES:
The element annihilated by a transformation, may not
remain zero during the subsequent transformations.
This method can be used only for symmetric matrices.
PROBLEMS FOR PRACTICE:
(1).Obtain using Jacobi’s method,all the eigen values and eigen
vectors of the matrix
2.536, 0.006,1.469
1 1 0.5 
0.531 0.721 0.444 
(1). A   1 1 0.25 ...................... ANS :  
 0.461 0.686 0.562 
0.5 0.25 2 
 0.71 0.094 0.698 
4, 2, 6
 1 1 
5 0 1   2 0 2 
(2). A  0 2 0  ............................. ANS : 
 0 1 0 
1 0 5   1
 1 
0
 2 2 