Page 1

WEEK # 03
1.5 Elementary Matrices & a method to find A 1
1.6 More on Linear Systems and Invertible Matrices

Objectives:
1) To Understand the concept of Elementary Matrices
2) To learn row equivalent matrices
3) To find the Inverse of a matrix using row operations
(specially for matrix with order 3  3 )
4) Use properties of inverse matrices
5) Use an inverse matrix to solve a system of linear equations

FEW IMPORTANT DEFINITIONS:

Elementary matrix: A Matrix E is called an elementary matrix if it can be
obtained from an identity matrix by performing a single elementary row
operation.
Row Equivalent matrix: Matrices A and B are said to be row equivalent if either
can be obtained from the other by a sequence of elementary row operations.

Singular matrix: A matrix that does not have an inverse is called a singular matrix.

Note: For singular matrix determinant will be zero.

Non-Singular matrix: A matrix that does have an inverse is called a non-singular

matrix.

Note: For non-singular matrix determinant will be non-zero.

 Page 2

ELEMENTARY MATRIX
Definition : An n  n matrix is an elementary matrix when it can be
obtained from the identity matrix I n by a single elementary row operation.
1 0 0 
Example : 0 0 1 is an elementary matrix
0 1 0 
as it can be obtained from I 3 by a single operation.
1 0 0  1 0 0 
e.g. 0 1 0  R  R gives 0 0 1
  3 2  
0 0 1 0 1 0 
1 0 0  1 0 0 
And 0 1 0  5 R2  R2 gives 0 5 0  so it is also Elementary matrix.
0 0 1 0 0 1 
1 0 0  1 0 0 
and 0 1 0   3R3  R3 gives 0 1 0  so it is also Elementary matrix.
0 0 1 0 0  3

Which of the following are not Elementary matrices ?

1 0 0  0 1 0  1 0 0 
1)  0 2 0  2) 1 0 0  3)  0 1 0 
 0 0 3   0 0 1  0 0 0 
1 0 0  1 0 0 0 
1 0 0 
4)   5)  0  1 0  6)  0 1 0 0 
0 1 0   0 0 1   0 0 1 0 
Answer : TRY YOURSELF
 Page 3

Definition of Row Equivlent matrices

Matrices A & B are said to be row Equivalent if either can be
obtained from the other by a sequence of elementary row operations.

 2 3   1 8 
A  R1  R2  R1
 1 5   1 5 
 1 8 
  R2  R1  R2 
 0 13 
 1 8   R2 
    R2 
 0 1   13 
 1 0 
  R1  8 R2  R1
 0 1 
 I2
Thus Matrix A and I 2 in this example are row equivalent as
I 2 is possible after a series of row operations .
 Page 4

1 2 3  1 2 3 
A   2 5    R2  2 R1  R2
3   0 1 3 
1 0 R  R1  R3
 8   0 2 5  3
1 2 3 
  0 1  3  R3  2 R2  R3
 0 0 1 
 
1 2 3  1 2 0 
R1  3R3  R1
  0 1  3  () R3  R3   0 1 0 
 
0 0 1   0 0 1  R2  3R3  R2
   
 B
Thus Matrix A and B in this example are row equivalent as
B is possible after a series of row operations.

1 2 3  1 2 3 
P   2 5   R2  2 R1  R2
3   0 1  3 
1 0 8   0 2 5  R3  R1  R3
   
1 2 3 
  0 1  3  R3  2 R2  R3
 0 0 1 
 
1 2 3 
  0 1  3  () R3  R3
 
0 0 1 
 Q
Thus Matrix P and Q in this example are row equivalent as
B is possible after a series of row operations.
 Page 5

 1 6 4 
P   2 4  1 

 1 2 5 
 
 1 6 4 
  0 8  9  R2  2 R1  R2
 R R R
 0 8 9  3 1 3
 
 1 6 4 

  0 8  9  R R R
 3 2 3
 0 0 0 
 
 R
Thus Matrix P and R in this example are row equivalent as
R is possible after a series of row operations.
 Page 6

INVERSE (RECAP OF WEEK 02)

In real arithmetic every nonzero number ‘a’ has a reciprocal with
the property
a  a 1  a 1  a  1
Then number is sometimes called the multiplicative inverse of ‘a’.
Definition of Inverse of a Matrix
An n  n matrix A is invertible (or non  singular ) when there exist
an n  n matrix B such that :
AB  BA  I n
where I n is called the identity matrix of order n.
The matrix B is the multiplicative inverse (or only inverse) of A.
Note : 1 : If no matrix B exist then inverse of matrix is not possible
and hence A will called a Singular Matrix.
Note : 2 Non square matrices do not have inverses.

 2 5  3 5 
L et A    an d B   
 1 3 1 2
 2 5   3 5   1 0
C learly A B       I2
  1 3  1 2  0 1
3 5  2 5   1 0
BA       I2
 1 2   1 3  0 1
T hus A and B are invertible and each i s a n inverse of other
 Page 7

RECAP-Continue…..

Formula to find inverse of a 2  2 matrix

a b
The matrix A    is invertible if and only if ad  bc  0,
 c d 
in which case the inverse is given by the formula
1  d b 
A1   
ad  bc  c a 

NOTE : ad  bc is called determinant of matri x A

 2 4
Q . If A  , find A 1 ?
 3 1
a b 
Sol . C om paring it w ith A   
c d 
a   2, b  4, c  3, d  1
C learly ad  bc  (  2)(1)  (4)( 3)   2  12   14  0
A s ad  bc  0 ( M eans determ inant is non  zero )
M eans inverse of A exi st
 1 2
1  d b  1  1 4    14 7
A 1       OR  
ad  b c  c a   14   3 2   3 1
 
 14 7
Page 8
 Page 9

RECAP-Continue…..

 1 4  1
Q. If A  , find A ?
 2 1
a b
Sol. Comparing it with A   
 c d 
a  1, b  4, c  2, d  1
Clearly ad  bc  (1)(1)  (4)(2)  1  8  9  0
As ad  bc  0  inverse of A exist
OR As ad  bc  0 ( Means determinant is non  zero
hence inverse of A exist )
 1 4
 
1  d  b  1  1 4  9 9
A1       OR  
ad  bc  c a  9  2 1   2 1
 
 9 9

 5 1 1
Q. If A    , find A (if possible) ?
 10 2 
a b
Sol. Comparing it with A   
c d 
we have : a  5, b  1, c  10, d  2
Clearly ad  bc  (5)(2)  (1)(10)  10  10  0
As ad  bc  0  inverse of A Does not exist
No inverse possible for the given matrix.
 Page 10

RECAP-Continue…..

Another Method to find inverse of a matrix

 2 4 1
Q . If A    , fin d A ?
 3 1 
S o l . T o fin d th e in v e r s e o f A , le t ' s s o lv e A X  I 2 fo r X .
 2 4 x y  1 0
     
 3 1  z w 0 1
  2x  4z  2 y  4w  1 0
     
 3x  z 3y  w  0 1
  2x  4z  1 ,  2 y  4w  0
3 x  z  0, 3 y  w  1
1 2 3 1
S o lv in g w e g e t x   , y  , z  , w 
14 7 14 7
 1 2
 
1 14 7
S o th e in v e r s e o f A is X  A   
 3 1
 
 14 7 
 Page 11

RECAP-Continue…..

 1 4
L et ' s T ry a n o th er  If A  , fin d A 1 ?
 1 3 
 Page 12

 k 1
Q. If Find k so that matrix A    is singular ?
 4 1

RECAP-Continue…..
 Page 13

 2 6 
Q. If Find  so that matrix A    is non  singular ?
  3 

Recap done
 Page 14

Topic of this week continue…..

Finding the Inverse of a Matrix by Gauss  Jordan Elimination

If we have matrix A of order ' n ' then follow the following steps :
Step 1 : Write n  2n matrix that consist of A on the left and the n  n
identity matrix I n on the right to obtain [ A I n ]. This process
is called adjoining matrix I n to matrix A.
Step 2 : If possible, row reduce A to I n using elementary operations( ERO)
on the entire matrix  A I n .The result will be the matrix  I n A1  .
If this is not possible, then A is non  invertible (or singular ).
Step 3: Check your work to verify AA1  I n  A1 A
 Page 15

Ex. Find A1 of given matrix by Gauss Jordan elimination

 2 3
if A   
 1 5 
 2 3 | 1 0
Sol .  A | I n    
 1 5 | 0 1 
 1 8 |1 1 

1 5 | 0 1  R1  R2  R1
 
 1 8 |1 1
 R2  R1  R2 
 0 13 | 1 2 
 1 8 |1 1 
 R2 
 2    R2 
 0 1 | 1  13 
 13 13 
 5 3
 1 0 |  
13 13
   R1  8 R2  R1
 0 1 | 1 2
 13 13 
 5 3

 13 13 
1
Thus matrix A is invertible and its inverse is A   
 1 2 
 13 13 
NOTE : You can confirm this by showing that AA1  I n  A 1 A
 Page 16

 1 2 3
Find A1 of A   2 5 3  by Gauss Jordan elimination
1 0 8 
 
The given matrix can be written as :
1 2 3  1 0 0
  2 5 3  0 1 0

1 0 8  0 0 1 

1 2 3  1 0 0
  0  R2  2R1  R2
1 3  2 1 0 
0 R  R  R3
 2 5  1 0 1  3 1
1 2 3  1 0 0
  0 1 3 

2 1 0  R3  2R2  R3
0 0 1  5 2 1 

1 2 3  1 0 0 
  0 1  3  2 1 0  (1) R3  R3
 0 0 1  5 2 1
 
 1 2 0  14 6 3 
  R1  3R3  R1
  0 1 0  13 5  3 
 0 0 1  5 2 1  R2  3R3  R2
 
 1 0 0  40 16 9   40 16 9 
  0 1 0  13 5  3  R1  2R2  R1 Thus A1   13 5  3 
 0 0 1  5 2 1   5 2 1 
   
 Page 17

 2 1 0
Find inverse of A   1 
1 0  by using row operations.
 3 0  1 

 Page 18

 3 3  2 
Ex. Find inverse of A   2 
4 3  by using row operations
 0 5 4 

 Page 19

NOTE: Often it will not be known in advance if a given matrix is invertible or not.

In certain cases it will be impossible to reduce a matrix A to Identity matrix by

elementary row operations. This will be signaled by a row of zeros appearing on
the left side of the partition at some stage of the inversion algorithm. If this
happens, we can stop the computation and will conclude that A is not invertible.

 1 6 4
Find the inverse of matrix A   2 
4  1 if possible.
 1 2 5 

 Page 20

Solving two linear equations at Once

Solve the systems:

a) x  2 y  3z  4 b) x  2 y  3z  1
2 x  5 y  3z  5 2 x  5 y  3z  6
x  8z  9 x  8 z  6
1 2 3  4  1 1 2 3  4  1 
    R2  2 R1  R2
Sol .  2 5 3  5  6  0 1 3  3  4 
1 0 8  9   6  0  2 5  5   7  R3  R1  R3
  
1 2 3  4  1
 
 0 1  3   3  4  R3  2 R2  R3
  1   1  1 
0 0
1 2 3  4  1
  R
 0 1  3   3  4  3  R3
0 0 1
 1  1   1 
1 2 0  1  4
  R1  3 R3  R1
 0 1 0  0  1
0  R2  3 R3  R2
 0 1  1  1
1 0 0  1  2
 
 0 1 0  0  1  R1  2 R2  R1
0 0 1  1   1 

It follows from the last two columns that the solution of the system for
(a) x = 1, y = 0, z = 1
(b) x = 2, y = 1, z = -1
 Page 21

Determining Consistency by Elimination

Question For the matrix : x  2k y  1

3x  2 y  2
a ) Write : The augmented matrix of given matrix :
b ) Discuss the conditions on k for which system will have :
(1) no solution
(2) exactly one solution
(3) infinitely many solutions
 Page 22

Question For the matrix : x  2a y  1

3x  2 y  b
a ) Write : The augmented matrix of given matrix :
b ) For conditions on a & b the system will have :
(1) no solution
(2) exactly one solution
(3) infinitely many solutions
 Page 23

Question For the matrix : x 2y  1

2 ax  2 y  b
a ) Write : The augmented matrix of given matrix :
b ) For conditions on a & b the system will have :
(1) no solution
(2) exactly one solution
(3) infinitely many solutions
 Page 24

Q uestion : What conditions must b1 , b2 and b3 satisfy in order for the

system of equations x1  x2  2 x3  b1
x1  x3  b2
2 x1  x2  3 x3  b3 to be consistent ?
Sol . The augmented matrix is
1 1 2  b1 
 
 1 0 1  b2  which can be reduced to row ec helon form
2 1 3  b3 

1 1 2  b1 
  R2  R1  R2
  0 1  1  b2  b1 
 0 1 R3  2 R1  R3
  1  b3  2b1 
1 1 2  b1 
 
 0 1 1  b1  b2   R2  R2
 0 1  1  b3  2b1 

1 1 2  b1 
 
 0 1 1  b1  b2  R2  R3  R3
0 0 0  b3  b2  b1 

It is clear from third row in the matrix that the system has
a solution (infinite ) if and only if b1 , b2 and b3 satisfy the condition
b3  b2  b1  0 OR b3  b2  b1
More convenient way to express another way AX  b is consistent
if and only if b is MATRIX of the form :
 b1 
b   b2  where b1 & b2 are ar bitrary.
 b1  b2 
 Page 25

Ex. What conditions must b1 , b2 and b3 satisfy in order for the

system of equations x1  2 x2  3 x3  b1
2 x1  5 x2  3 x3  b2
x1  8 x3  b3 to be consistent ?
Sol. The augmented matrix is
 1 2 3  b1 
 
  2 5 3  b2  which can be reduced to row echelo n form
 
 1 0 8  b3 
 1 2 3  b1 
  R2  2 R1  R2
  0 1  3  b2  2b1 
 0  2 5  b  b  R3  R1  R3
 3 1 

 1 2 3  b1 
 
  0 1  3  b2  2b1  R3  2 R2  R3
 0 0  1  b  2b  5b 
 3 2 1

1 2 3  b1 
 
  0 1  3  b2  2b1   R3  R3
 0 0 1  5b  2b  b 
 1 2 3

 1 2 0   14b1  6b2  3b3 

  R1  3 R3  R1
 0 1 0  13b1  5b2  3b3 
0 0 1  5b  2 b  b  R2  3 R3  R2
 1 2 3 

 1 0 0   40b1  16b2  9b3 

 
 0 1 0  13b1  5b2  3b3  R1  2 R2  R1
 0 0 1  5b  2b  b 
 1 2 3 
In this case there are no restrictions on b1 , b2 and b3 so the system
has the unique / one solution.
Hence x1  40b1  16b2  9b3
x2  13b1  5b2  3b3
x3  5b1  2b2  b3 for all values of b1 , b2 and b3
 Page 26

Question For the matrix :

 x1  x2  x3  1
x1  3 x2  3 x3  1
2 x1  ax2  bx3  c
a ) Write : The augmented matrix of given matrix :
b ) Reduce the augmented matrix of the given system
to an upper triangular matrix.
c ) Discuss the conditions for which syst em will have :
(1) no solution
(2) exactly one solution
(3) infinitely many solutions
Page 27
 Page 28

EXAMPLE:
1 3 1
a ) For the matrix A   2 0 6  find A1 if exist .
1 3 2 
b) Use the information of A1 to solve the system :
 1
2 AT X   2 
 4 
 Page 29

1 3 1 1 0 0
P a r t ( a ) A   2 0 6 0 1 0 
1 3 2 0 0 1 
1 3 1  1 0 0
R 2  2 R1  R 2
  0 6 4   2 1 0 
R 3  R1  R 3
 0 0 1   1 0 1 
1 3 1  1 0 0
 2 1 1  R2
 0 1    0  R2
 3 3 6   6
0 0 1   1 0 1 

 1 
1 0 3  0 0
2
 
2 1 1
 0 1    0  R1  3 R 2  R1
 3 3 6 
0 0 1 1 0 1 

 
 1 
1 0 0  3 3
2
  R1  3 R 3  R1
1 1 2
 0 1 0   2
 3 6 3 R2  R3  R 2
0 0 1 1 0 1  3

 
 1 
 3  3
2
 
1 1 2
H ence A  1   
 3 6 3
 1 0 1 

 
 Page 30

 1
Par t ( b ) Given 2 A T X    3  ...........(1)
 4 

Sinc e A is invertible , so is 2 A T hence  2 A T   2 1  A 1  .

1 T

Thus (1) beco mes

T
 1 
 3  3
 1  1 2  1
 
X   2 AT   3   2 1 A 1 T   3   2 1   1  1 2  3 
  
1

     3 6 3  
 4   4   1 0  4 
 1 
 

 1 
 3  1
3
  1
1 1 1
  0    3 
2 2 6 
   4 
 2
3 1
 3 
 
 0
 
  0 
 1
 
 2
 Page 31

Part (a) Consider the following linear equations :

x1  2 x2  5 x4  1
x1  2 x2  2 x3  3x4  1
2 x1  4 x2  2 x3  8 x4  2
Use Gauss Jordan elimination to solve.
Sol. The augmented matrix is :
1  2 0 5 1
 1  2 2 3 1
 2  4 2 8 2 
1  2 0 5 1 
  R2  R1  R2
  0 0 2  2 0 
R3  2 R1  R3
0 0 2  2 0 
1  2 0 5 1 
R
 0 0 1  1 0  2  R2
2
0 0 2  2 0 
1  2 0 5 1 
 0 0 1  1 0  R3  2 R2  R3
0 0 0 0 0 
x1  2 x2  5 x4  1 x1  2 x2  5 x4  1
 
x3  x4  0 x3  x4
 x1  2s  5r  1, x2  s, x3  r , x4  r
 Page 32

Part ( B):Write the general solution of the homogeneous system :

x1  2 x2  5 x4  0
x1  2 x2  2 x3  3x4  0
2 x1  4 x2  2 x3  8 x4  0
Sol. Following steps from part (a ) above, we will have :
1  2 0 5 0 
 0 0 1  1 0 
0 0 0 0 0 
Then general solution is : x1  2 x2  5 x4  0
x3  x4  0
 x1  2 s  5r , x2  s, x3  r , x4  r
 Page 33

a ) The following system of equations can be written in the form Ax  B

2x  y  z  2
x y 0
x  z4 Find A1.
b) Use A1 to solve AT x  B
 2  1  1 1 0 0 
 
Sol. 1  1 0  0 1 0 
1 0 1  0 0 1 

1  1 0  0 1 0 
 
  2  1  1 1 0 0  R1  R2
1 0 1  0 0 1 

1  1 0  0 1 0 
  R2  2 R1  R2
0 1  1 1  2 0 
0 R3  R1  R3
 1 1  0  1 1 
1  1 0  0 1 0 
 
0 1  1  1  2 0  R3  R2  R3
0 0 2   1  1 1 

 
1 1 0 0 1 0 
  1
0 1 1  1  2 0    R3  R3
 2
1 1 1
0 0 1 
 2 2 2
 
1 1 0 0 1 0 
 
1 3 1
0 1 0  R2  R3  R2
 2 2 2
 1 1 1
0 0 1 
 2 2 2
 1 1 1  1 1 1
1 0 0   2 2
2 2 2 2
   
1 3 1  1 3 1
0 1 0  R1  R2  R1  Hence A  
1

 2 2 2   2 2 2
   
 0 1 1 1    1 1 1
0 1  
 2 2 2   2 2 2
 Page 34

b) Use A1 to solve AT x  B

Sol. A x  B
T
 x  A 1 T
B
T
 1 1 1
 2 2 2
   2   1 
1 3 1    
x   0   1
 2 2 2
   4   3
   
  1 1 1

 2 2 2
Page 35
 Page 36

3  7
The matrix A    can be reduced down to the identity in
1 0 
the following three steps :
0  7 1 0 1 0
A RowOp1 1 0   RowOp2   RowOp3   I2
  0  7 0 1
a) What are three elementary row operations used ?

b) What are three Inverse elementary row operations ?

c)Write A1 as a product of Elementary Matrices . Justify your answer.

Donot multiplymatrices in the final answer.
 Page 37

d )Write A as a product of Elementary Matrices . Justify your answer.

Do not multiply matrices in the final answer .
 Page 38

1 0 0
The matrix A   3 1 0  can be reduced down to the identity in
0 0 2 

the following three steps :
1 0 0 1 0 0 1 0 0
A  RowOp1  0 1 0   RowOp 2  0 1 0   RowOp 3  0 1 0   I
0 0 2 0 0 2  0 0 1 
     
a ) What are three elementary row operations used ?

b) What are three Inverse elementary row operations ?

 Page 39

c) Write A as a product of Elementary Matrices . Justify at each step,

the derivation of the formula you are using .Leave your answer
as the product. Do not multiply matrix.
Sol.
 Page 40

 0 3 1 
The matrix A  1 0 0 can be reduced down to the identity in
 0 6 1 
 
the following three steps :
 0 3 0  1 0 0   1 0 0  1 0 0
A RowOp1 1 0 0 RowOp2  0 3 0 RowOp3  0 3 0  RowOp4  0 1 0   I
 0 6 1   0 6 1  0 0 1  0 0 1 
       
a) What are four elementary row operations used ?

b) What are four Inverse elementary row operations ?

 Page 41

c) Write A1 as a product of Elementary Matrices . Justify your answer.

Do not multiply matrices in the final answer .
Sol.

d ) Write A as a product of Elementary Matrices . Justify your answer .

Do not multiply matrices in the final answer.
 Page 42

Practice Ques tions

 AB 
1
Find the inverse of following if exist & hence find for
the following :
 1 4  0 3
1) A   and B 
 2 3   2 3 
 
 1 0  1 3
2) A   and B  
 2 3   2 3


 1  1 1  4  1 3
3) A   2 3 1 and B   3 3 1 
 2 1 1   2 1 1 

 AB  ,  AT 
1
and  3 A 
1 1
TRY : Find for the following :

1  1 4 1  0 3
1) A   and B  
 2 3   2 3 


 1 0  1 3
2) A 1   and B 1

 2 3   2 3
 
 1  1 1  4  1 3
3) A 1   2 3 1 and B 1   3 3 1 
 2 1 1   2 1 1 
Page 43
 Page 44

References:
Anton, H. (2015). Selected Chapters and Solutions from Elementary Linear Algebra (11th ed.).
ISBN: 978-1-118-43441-3