MATRICES

1
 Definition:
Matrix is an array of objects arranged in m rows and n columns

 5 1 2 30.3 2
A   4  1  0.2 1 4
 3  1 5 0 7

m = 3 (no of rows) ; n = 5 (no of columns)

a11  5 Each entry: aij a24  1

i denotes row j denotes column

The size of a matrix is denoted as m x n

The size of a matrix A is 3 x 5
Is this a square matrix?

2
 Matrix A  aij  and B  bij  are equal if

i) If A and B have same no. of rows and no. of columns

i) If aij  bij for each i and j.

 3 2  a   3 8 
Eg: What is the value of a-c?    
 6 4   3c 4 

Ans: a = 6 ; c = 2

3
Only same sizes of matrices can perform addition / subtraction:

 3 2  3 8  3  3 2  8  6 10
 6 5    0 4   6  0    
     5 4   6 9 

Product of a Matrix and a Scalar:

 3 2 4  3 4  2  12 8 
4     
 6 5   4  6 4  5   24 20 

4
 Product of Matrices:

A B
(m x n) (p x q)
AB Only true if n = p
(m x q)

Eg: A B
(3 x 2) (2 x 4)
AB
(3 x 4)

5
 1 2
 1 7
1   1 0 1 4
Find AB if A 2   ; B   0
 4 5 9   6 3

2X 3 3 X2
Row 1 of A * Column 1 of B Row 1 of A * Column 2 of B
1 2
1  1  1  (0  4)  (1  6) (1  7)  (0  0)  (1  3)
AB =  (4  1)  (5  4)  (9  6) (4  7)  (5  0)  (9  3) 
(2 X 2) 2  
Row 2 of A * Column 1 of B Row 2 of A * Column 2 of B

AB =  7  4

70 55 

6
Matrix Transposition: Interchanging rows and columns
 5 1 2 30.3 2
Eg: A   4  1  0.2 1 4
 3  1 5 0 7 Row 1 becomes column 1
Row 2 becomes column 2
 5 4 3
 1 1  1
Row 3 becomes column 3

AT   2  0.2 5
 
30.3 1 0
 2 4 7 

7
 Determinant of a square matrix: |A|:
a b
i) Determinant of a 2X2 matrix: |A|: If A   then A  ad  bc
c d 

3  4
Eg : A    then A  (3  2)  (4  1)  10
1 2 

ii) Determinant of a 3X3 matrix: |A|:

a b c
Eg : A  d e f  then A  a (ei  fh )  b(di  fg )  c(dh  eg )
 g h i 

 1 2 0
Eg : A   1 1 0 then A  1(1  3  0  4)  2(1  3  0  2)  0(1  4  1  2)  9
 2 4 3

8
 Echelon Matrices (pg 13)
If the number of zeros preceding the distinguished element
increases row by row = Echelon matrix

Distinguished element =first nonzero entry of a row

Examples:
2 3 2 0 4 5  6
  Distinguished
2 zeros 0 0 7 1 3 2 0 
1.  element
5 zeros 0 0 0 0 0 6 2 
 
0 0 0 
 0 0 0 0

Increases row by row

9
 Echelon Matrices (pg 13)
Examples: Distinguished
element
1 2 3
 
2 zeros 0 0 4
2. 
0 0 0 Distinguished
 
0 0  element
 0

1 zero 0 1 3 0 0 4 0
 
3 zeros 0 0 0 1 0 3 0
3. 
4 zeros 0 0 0 0 1 2 0
 
0 1 
6 zeros  0 0 0 0 0

10
 Row Reduced Echelon Matrices (pg 13)
If the distinguished elements are:
a) The only nonzero entries in their respective columns
b) Each equals 1

Examples: 0 1 3 0 0 4 0
 
0 0 0 1 0 3 0
0 0 0 0 1 2 0
 
0 1 
 0 0 0 0 0

1 0 0
 
0 1 0
0 0 1
 
0 0 
 0
11
 (pg 14)

Every matrix has a unique row reduced echelon matrix.

But an echelon matrix of a given matrix is not unique

Matrix Echelon Matrix

Gaussian elimination

Matrix Row reduced Echelon Matrix

Gauss-Jordan elimination

12
 Elementary Row Operations

We can perform 3 kinds of operations:

a) Exchange the ith row and the jth row : Ri  R j .

Eg : Perform R1  R3

R1 2 3 1  0 0 1 
   
Ans : R2 4 1  6  R1  R3  4 1  6 
R3 0 0 1  2 3 1 
 

13
 Elementary Row Operations

b) Multiply the ith row by a nonzeroscalar k : Ri  kRi , k  0.

Eg : Perform R2  3R2

5 6 1  5 6 1 
   
Ans :  2 1  3  R2  3R2  6 3  9 
1 9 2  1 9 2 
   
-3X3
2X3 1X3

14
 Elementary Row Operations

Replace the ith row by a k times the jth row plus the ith row :
Ri  Ri  kR j , k  0.

Eg : Perform R1  R1  3R2

1 0 0 1 3 0
   
Ans :  0 1 0  R1  R1  3R2  0 1 0 
0 0 1 0 0 1
   
1+(3*0) 0+(3*0)
0+(3*1)

15
 MATRIX INVERSE A-1

• Invertible Matrices
A square matrix A (nxn) is said to be invertible if there exists
a unique matrix A-1 (nxn) such that A A-1 = A-1 A= I n
(I is an identity matrix)

1 0
Note: I2   
 0 1 
1 0 0
I 3  0 1 0
0 0 1

16
• If A has an inverse, then A is called a nonsingular matrix;
it means
det A  A  0

Properties of inverse matrix

i)  AB1  B 1 A1 ii) A 

1 1
A

T
iii) If A is the transpose of A, then A   T 1
 .
 A 1 T

iv) A 
n 1
  , n  0,1,2,...
 A 1 n

iii) If any nonzero scalar k , then kA

1

1 1
k
 
A .

17
Finding inverse by Gauss-Jordan elimination (pg 16)

A / In  Elementary Row Operations

 I n / B
1
A

18
 6 1
Eg: Find the inverse of A    by Gauss-Jordan elimination:
 3 2
6 0
Ans :  A I 
11
  
3 20 1 
R2  R2  (1 / 2) R1
6 1 1 0
 
 0 3 / 2 1/ 2 1 
 
R2  (2 / 3) R2

6 1 1 0   2 / 9 1/ 9 
1
   A   
 0 1  1 / 3 2 / 3  1/ 3 2 / 3 
 
R1  R1  R2
 6 0 4 / 3  2 / 3

 0 1 1/ 3 2 / 3 

 1 0 2 / 9 1/ 9 
 
  I A1
 0 1 1/ 3 2 / 3  
  R  (1 / 6) R  
1 1

19
  1 1 0
 
Eg: Find the inverse of A    1 3 4  by Gauss-Jordan elimination:
 0 4 3
 
 1 1 01 0 0 1 1 01 0 0
   
Ans : A I     1 3 40 1 0  R2  R2  R1  0 4 41 1 0
 0 4 30 0 1  0 4 30 0 1 
 

1 1 0 1 0 0 1 0 1 3 / 4  1/ 4 0
  R1  R1  1R2  
R2  (1 / 4) R2  0 1 1 1/ 4 1/ 4 0 0 1 1 1/ 4 1/ 4 0
0  
1  3 2
1 
R R 4 R
 4 3 0 0 3
0 0 1 1 1

 1 0  1 3 / 4  1/ 4 0  1 0 0 7 / 4 3 / 4  1
  R1  R1  R3  
R3  (1) R3  0 1 1 1 / 4 1 / 4 0 0 1 0  3/ 4  3/ 4 1 
0 0 1 1  
 1 2 3
 1
R R 2 1R
 1 0 0 1 1 1

 7/4 3/ 4  1
 
A 1    3/ 4  3/ 4 1
 1  1
 1

20
The inverse for 2X2 square matrix can also be found by: (pg 15)

a b  1 1  d  b

If A   
 ; then A   
c d  ad  bc   c a 

6 1
A Eg: Find the inverse of A    :
 3 2

1 1  2  1  2 / 9  1 / 9 
A      
12  3   3 6    1 / 3 2 / 3 

21
The inverse for 3X3 and larger square matrices can be found
by using the classical adjoint method too: (pg 18)

A1 
1
adjoint A
det A

adjoint A  (cofactorA)T
A

a b c
  e f d f d e
If A   d e f ; A a b c
g h i g i g h
 h i 

22
 Finding cofactor A https://youtu.be/EcI4E15ElK0
 1 1 0
 
Eg: Find the cofactor of A    1 3 4 
 0 4 3
 
 a11  a12 a13 
 
Eg: Let’s say cofactor A    a21 a22  a23 
 a  a32 a33 
 31
 1 1 0
3 4  
a11  refer to row 1 and column 1 of matrix A;  7 A    1 3 4 
4 3  0 4 3
 

1 4 
1 1 0
a12  refer to row 1 and column 2 of matrix A; -  3 A   1 3 
0 3  4
 0 4 3
 
1 3  1 1 0
a13  refer to row 1 and column 3 of matrix A;  4 A    1 3 4 
0 4  
 0 4 3
  23
  1 1 0
 
Eg: Find the inverse of A    1 3 4  by cofactor /adjoint method:
 0 4 3
 
3 4 1 4
Ans : det A  A  1 1  0  4
4 3 0 3
 a11  a12 a13    7 3  4 
   
cofactor A    a21 a22  a23     3 3  4 
 a  a32 a33   4  4 4 
 31

 7 3 4 
 
adjoint A  (cofactorA)T  3 3  4
 4  4 4 
 

 7 3 4   7/ 4 3 / 4  1
   
1
A 
1
adjoint A   3 3  4     3 / 4  3 / 4 1 
1
det A 4  
 4  4 4   1 1  1

24
 System of Linear Equations
a11 x1  ...  a1n xn  b1
a21 x1  ...  a2 n xn  b2

am1 x1  ...  amn xn  bm Augmented
Matrix

 a11  a1n   x1   b1   a11  a1n b1 

       
 .  .   x2   b2  ~  .  . . 
A ; X   ;B  A
.  .    .  . . 
       
a     a  amn bm 
 m1  amn   xn   bn   m1

25
 Solving system of linear equations by
Gaussian eliminations
In solving a system of linear equations by Gaussian eliminations,
we can perform all the 3 kinds of elementary row operations to the
augmented matrix:
Eg: Solve using Gaussian Elimination:
x1  2 x2  1
x1  x2  5

~ 1 2  1 Augmented
Let A    Matrix
1 1 5
Next, Perform the Elementary Row operations

26
Elementary Row Operations
~ 1 2  1
A   
1 1 5
R2  R2  R1
 1 2  1
 
 0 1 6 
Row reduced echelon matrix
R2   R2
 1 2 1 
 
 0 1  6 By back substitution:
x1  11
R1  R1  2R2
x2  6
 1 0 11 
  Unique solution exist
 0 1  6
27
 Solve the linear system by Gaussian elimination: (pg 11)

3x1  2 x2  2 x3  5 x4  8
0.6 x1  1.5 x2  1.5 x3  5.4 x4  2.7
1.2 x1  0.3x2  0.3x3  2.4 x4  2.1

 3 2 2 5 8 
~   Augmented
Let A   0.6 1.5 1.5  5.4 2.7 
 1.2  0.3  0.3 2.4 2.1  Matrix
 

28
Elementary Row Operations
 3 2 2 5 8  By back substitution:
~   From second row:
Let A   0.6 1.5 1.5  5.4 2.7 
 1.2  0.3  0.3 2.4 2.1  1.1x2  1.1x3  4.4 x4  1.1
 
R2  R2  0.2 R1 Simplify : x2  1  x3  4 x4
R3  R3  0.4R1 From first row:
3 2 2 5 8  3x1  2 x2  2 x3  5 x4  8
 
 0 1.1 1.1  4.4 1.1  Substitute x2 into this :
 0  1.1  1.1 4.4  1.1
  x1  2  x4
R3  R3  R2
x3 and x4 Remains unknown
3 2 2 5 8 
  Infinitely many solutions
 0 1.1 1.1  4.4 1.1
0 0 0 0 
 0
29
 Solving the system of linear equations by Cramer’s Rule :
(pg 20)
Ai
The solution to the system of linear equation AX=B is : xi 
A
Where Ai is the matrix obtained by replacing the entries
In the i-th column of A by entries in the matrix B
2 x1  6 x2  4
Eg:
x1  3 x2  8
 2 6  x1    4
A    ; X   ; B   
 1 3  x2    8
  4 6  2  4
A1    A2   
  8 3 1  8
Replace the second column
Replace the first column of A with B
of A with B

30
 Eg: Solve the system by using Cramer’s Rule: ( pg 20)

x1  3 x2  4 x3  1
 x1  x2  3 x3  14
x2  3 x3  5
 1  3  4  x1  1
     
Let A    1 1  3  ; X   x2 ; B  14 
0 1  3  x  5
  3  

 1  3  4  1 1  4  1 3 1 
     
 A1  14 1  3  ; A2    1 14  3 ; A3    1 1 14 
 5 1  3  0 5  3 0 1 5 
    

Replace with B

31
  1  3  4  x1  1
     
Let A    1 1  3  ; X   x2 ; B  14 
0 1  3  x  5
  3  

 1  3  4  1 1  4  1 3 1 
     
 A1  14 1  3  ; A2    1 14  3 ; A3    1 1 14 
 5 1  3  0 5  3 0 1 5 
    

A1 117
x1    9
A 13
A2 10
x2  
A 13

A3 25
x3  
A 13
32
Rank of matrix A is the number of nonzero rows in echelon matrix of A.
(pg 21)
1 1 4 2
 
Let A   0 1 3 2  . Find rank (A).
 3  2 15 8 
 

First find echelon matrix by performing row operations.

33
 Revision: Echelon Matrices (pg 13)
If the number of zeros preceding the first nonzero entry of a row
(distinguished element)
increases row by row
Examples: 2 3 2 0 4 5  6
  Distinguished
0 0 7 1 3 2 0 
1.  element
0 0 0 0 0 6 2 
 
Distinguished 0 0 0 
 0 0 0 0
element
1 2 3 0 1 3 0 0 4 0
   
0 0 4 0 0 0 1 0 3 0
2. 
0
3. 

0 0
 0 0 0 0 1 2 0
0  
 0 0  0 1 
 0 0 0 0 0

34
 1 1 4 2
 
A  0 1 3 2
 3  2 15 8 
 
R3  R3  3R1
1 1 4 2
 
 0 1 3 2
 0 1 3 2
 
R3  R3  R2
There are 2 nonzero
rows in echelon matrix 
1 1 4 2
 
 0 1 3 2 Echelon matrix
Therefore, rank (A)=2 0 0 0 0
 

35
 T
rank ( A ) = rank (A ) pg (22)

A linear system of m equations in n unknowns has solutions if and

~
only if rank (A ) = rank ( A ) . Furthermore,
Augmented matrix
~
If rank (A ) = rank ( A ) = n, then the system has one solution.

~
If rank (A ) = rank ( A ) < n, then the system has many solutions.

36
 Eg: Does this system have solution? ( pg 22)
2 x1  3 x 2  6
4 x1  6 x 2  18

 2  3 ~ 2  3 6 
Let A    Let A   
 4  6  4  6 18 
R2  R2  2R1 R2  R2  2R1

 2  3  2  3 6
  Echelon matrix   Echelon matrix
0 0   0 0 6
There are 1 nonzero There are 2 nonzero
row in echelon matrix rows in echelon matrix
~
Therefore, rank (A)=1 Therefore, rank (A)= 2

~
Therefore no solution since rank (A )  rank ( A ) .

37
 Eg: Does this system have solution?
x  3 y  1
4 x  3 y  11
 1  3 ~  1  3  1
Let A    Let A   
4 3   4 3 11 
R2  R2  4R1
R2  R2  4R1
 1  3  1  3  1
  Echelon matrix   Echelon matrix
 0 15   0 15 15 
There are 2 nonzero There are 2 nonzero
rows in echelon matrix rows in echelon matrix
~
Therefore, rank (A)=2 A
Therefore, rank ( )= 2
~
rank (A ) = rank ( A) =2 and n=2
Therefore this system has one solution.
38