MATRICES AND MATRIX OPERATIONS

Matrices

Definition1. An m  n − matrix A is rectangular array of m  n scalars

 a11 a12 a1n 


a a22 a2 n 
A =  21
 
 
 am1 am 2 amn  mn

1 3 25  1 
2 5 6 
Example 1. B = 5 7 10  , C =  , D =  2  , E = 1 212
  8 10 15  23  
1 2 −10 33  3  31

Scalars themselves will be identified with the corresponding 11 − matrices, e.g. a   a  ;

3   3  ; 5 + i  5 + i  2 − i   2 − i  , where i = −1  is the so-called imaginary

unit.

Definition 2. For any k 1, 2,..., m , the 1 n − matrix  ak 1 ak 2 akn  is called

 a1 j 
a 
the k − th row of A ; and for each j 1, 2,..., n , the m1 − matrix   is called the j-th
2j

 
 
 amj 
column of A .

1 2 4 5
7 8 9 10 
Example 2. A = 
 5 −3 7 9
 
2 5 8 3

We can show the second row as  7 8 9 10

4
9 
Also the third column as  
7 
 
8 
 Definition 3. An element akj of a matrix A is called the kj − th entry of A and is

denoted by  A kj := akj . Two matrices A and B are equal if they are of the same size, say

both are of m  n size, and  A kj =  B  kj for all k 1, 2,..., m ; j 1, 2,..., n .

Example 3. The matrices below are not equal, since their size are not same.

1 2 0  1 2 
4 6 0  4 6
  23   22

and

0 0
0 0
22 =   0 0   011 = 0 .
0 0 22 0 0
  32

Remark 1. Any m  n − matrix may be considered as a function of pairs ( k , j ) of

naturals, k 1, 2,..., m and j 1, 2,..., n with scalar values.

Definition 4. The identity matrix of size n  n is

1 0 0 0 0 
0 1 0 0 0 
 
I nn = 0 0 1 0 0 
 
 
0 0 0 0 1  nn

Similarly the function f ( k , j )  0 gives rise to zero matrices of an arbitrary size m  n .

1 0 0 0
1 0 0  0
1 0  0 
= 0 1 0  , I 22 =   1 0
=
0 1 
Example. I 33 , I 44
 0 0 1 0
0 0 1   
0 0 0 1

Definition 5. Any matrix of size n  n is called a square matrix.

Example 4. Every identity matrix is a square matrix. But a zero matrix doesn’t need to
be a square matrix, for example
 0 0 0 
23 =  
0 0 0 

is a zero matrix which is not a square matrix.

Some operations on matrices

We begin with matrix operations which do not depend on size of matrices.

Definition 6. The scalar multiplication of a matrix A by a scalar  is given by

 Akj =   Akj .

1 0
Example 5. Let A be 2  2 − matrix such as A =   and  = 2 . Then
3 5 

1 0  2.1 2.0  2 0 
2. A = 2.  = = 
3 5   2.3 2.5 6 10 

Definition 7. The transpose operation is the interchanging of rows of a matrix with its
columns: A → AT , where

 AT  =  A jk .
kj

T  2 4
 1 2 3  
T
2 4 6   
Example 6.  2     = 4 0 8  = 4 0 
  2 0 4    6 8 

Clearly, ATT = A for any matrix A .

The following several operations are depending on the size of involved matrices.

Definition 8. Let A and B be m  n matrices, then the sum A + B is an m  n − matrix

given by

 A + B kj =  Akj +  B kj for all k 1, 2,..., m and j 1, 2,..., n.

The introduced above matrix operations possess the following natural properties:

Prop 1. If the matrix operations below are well defined, then:

a) A + B = B + A (commutativity of addition);
b) ( A + B ) + C = A + ( B + C ) (associativity of addition);

c)  . ( A + B ) =  A +  B ;

d) (  A ) =   AT ;
T

e) ( A + B ) = AT + BT ;
T

f) 1 A = A and 0. Amn = mn .

g) A + O = O + A = A

0 1  1 2 1 0 
Example 7. A =   , B=  , C=  and  = 2 .
1 2  3 4  0 1 

 0 1  1 2 0 + 1 1 + 2  1 3 
A + B =  + = = 
 1 2  3 4 1 + 3 2 + 4   4 6 
a)    A+ B = B + A
 B + A = 1 2 0 1  1 + 0 2 + 1  1 3 
3 + = =

  4  1 2  3 + 1 4 + 2   4 6  

 0 1  1 2   1 0  1 3 1 0   1 + 1 3 + 0   2 3 
b) ( A + B) + C =   + +  = + = = 
  1 2  3 4    0 1   4 6   0 1   4 + 0 6 + 1   4 7 

0 1   1 2 1 0  0 1   2 2  2 3
A+ (B + C) =  + +  =  + = 
 1 2   3 4   0 1    1 2   3 5   4 7 

 ( A + B) + C = A + ( B + C )

 1 3   2 6  
 . ( A + B ) = 2.     =   
  4 6   8 12  
c)    .( A + B ) =  A +  B
0 1  1 2   0 2   2 4   2 6  
 A +  B = 2.   + 2. 3 4  =  2 4  +  6 8  = 8 12  
1 2         
T
 0 1    0 2
T
 0 2 0 1 
( 2  A) =  2.   = = = 2 = 2. AT
T
d)    
 1 2    2 4   2 4 1 2 
 1 4  
T
 0 1  1 2   1 3
T

( A + B) =   + = = 3 6  
T
 3 4   4 6
 1 2          A + B T = AT + BT
e)  ( )
 0 1   1 3  1 4  
AT + BT =   +  =  
 1 2   2 4  3 6  
0 1  0 0  0 1  0 0  0 1 
f) A+O =  + = = +  =O+ A= A
1 2   0 0  1 2   0 0  1 2 

Our next operation of the matrix multiplication is applicable if the number of columns
of the first factor is equal to the number of rows of the second factor.

Definition 9. For Amn and Bn p , the m  p − matrix A.B is given by

n
 AB ij =   Aik   B kj .
k =1

Theorem 1. If the matrix operations below are well defined, we have:

a) ( A.B ) .C = A. ( B.C ) (associativity of multiplication);

b) A. ( B + C ) = A.B + A.C (distributivity);

c) ( A + B ) .D = A.D + B.D (distributivity);

d)  . ( A.B ) = ( . A ) .B = A. ( .B ) .

The matrix multiplication is not commutative in general. For example

0 1  1 0 
A=  and B =  
0 0 0 0

Then

Am n B n  p = Cm p
0 1  1 0 
A=  ,B =  
0 0 0 0
0 1  1 0   0.1 + 1.0 0.0 + 1.0   0 0 
AB =   = = 
0 0   0 0   0.1 + 0.0 0.0 + 0.0   0 0 
  AB  BA
1 0  0 1  1.0 + 0.0 1.1 + 0.0  0 1  
BA =   = = 
0 0  0 0  0.0 + 0.0 0.1 + 0.0   0 0 
 1 2 5 6
Example 8. For the matrices A =   and B =   , find AB and BA .
3 4  7 8 

1 2 5 6 1.5 + 2.7 1.6 + 2.8   5 + 14 6 + 16  19 22 

Solution. AB =   = = = .
3 4 7 8  3.5 + 4.7 3.6 + 4.8 15 + 28 18 + 32  43 50 

 5 6 1 2 5.1 + 6.3 5.2 + 6.4   5 + 18 10 + 24  23 34 

BA =   = = = 
7 8  3 4  7.1 + 8.3 7.2 + 8.4  7 + 24 14 + 32   31 46 

2 4 1 0 
1 0 2  0 −1 2 
Example 9. A =   , B = 1 3  , C =   and D =  0 2 
 2 1 3  23   4 2 1  23
0 5  32 3 1  32

a) Find AB.
A23 .B32 = E22

 2 4
1 0 2
AB =   1 3  = 1.2 + 0.1 + 0.2 1.4 + 0.3 + 2.5 =  2 14 
    2.2 + 1.1 + 3.0 2.4 + 1.3 + 3.5  5 26
 2 1 3  23 0 5     
  32

 c11 c12 
(If we multiply A and B , we get 2  2 matrix. Let AB be AB = C =   .Then
c21 c22 
to get c11 we will multiply the first row of A with the first column of B , to get c12 we

will multiply the first row of A with the second column of B , to get c21 we wil

multiply the second row of A with the first column of B and to get c22 we will

multiply the second row of A with the second column of B .)

b) Find BA.
 2 4  2 + 8 0 + 4 4 + 12  10 4 16 
  1 0 2 
B32 A23 = F33 = 1 3    =  1 + 6 0 + 3 2 + 9  =  7 3 11 .
0 5  
2 1 3
0 + 10 0 + 5 0 + 15  10 5 15

c) Find AC.
A23C23 is not defined, since 3  2.

d) Show that ( A.B ) .C = A. ( B.C )

 2 4 
 1 0 2    0 −1 2   2 14  0 −1 2 
( A.B ) .C =    . 1 3   .  =
2 1  5 26  22  4 2 1  23
.
  2 1 3  0 5    4
  
  2.0 + 14.4 2 ( −1) + 14.2 2.2 + 14.1  56 26 18 
= = 
5.0 + 26.4 5. ( −1) + 26.2 5.2 + 26.1 104 47 36 

 2 4 
1 0 2    0 −1 2   56 26 18 
A. ( B.C ) =   .  1 3  .  =
 2 1 3   0 5   4 2 1   104 47 36 
  
e) Show that A. ( B + C ) = A.B + A.C .

  2 4  1 0   3 4 
1 0 2        1 0 2  
A. ( B + D ) =    1 3  + 0 2   =   . 1 5 
 2 1 3    0 5  3 1    2 1 3  3 6 
     

3 + 0 + 6 4 + 0 + 12   9 16 
= = 
 6 + 1 + 9 8 + 5 + 18  16 31
 9 16 
A.B + A.D =  
16 31

Theorem 2. If A.B is defined, then ( A.B ) = BT . AT .

T

2 4
1 0 2 
Example 10. A =   and B = 1 3 
 2 1 3  23  
0 5  32

Then

 2 4 
T


T
    =    
( A.B ) =  
1 0 2 2 14 2 5
 =
T
  1 3   5 26 14 26 
  2 1 3  0 5       

  
  ( A.B ) = B . A
T T T

1 2  
 2 1 0     2.1 + 1.0 + 0.2 4 + 1 + 0   2 5  
B .A = 
T T
 . 0 1  =  =
 4 3 5   2 3   4 + 0 + 10 8 + 3 + 15 14 26  
  

 x −1  2 y   2 0
Example 11 (Midterm). Find x , y , z and t if −2   + 3 = .
3 1   z 4   z t 

 x −1  2 y   2 0  −2 x 2   6 3 y   2 0
Solution. −2   + 3 =  + = 
 3 1   z 4   z t   −6 −2 3z 12   z t 
  −2 x + 6 2 + 3 y   2 0
 = 
 −6 + 3z −2 + 12  z t 

−2 x = −4  x = 2
−2 x + 6 = 2 
2 + 3 y = 0 2
 3 y = −2  y = − 2
  3  x = 2, y = − , z = 3, t = 10
−6 + 3z = z −6 = −2 z  z = 3 3
10 = t 
t = 10

1 2  2 −1
Example 12 (Midterm). Let A =   and B =  .
3 2   −3 4 
a) Find the solution of 2 A + 3B − 4I ,

b) For the given A and B show that ( AB ) = BT AT .

T

Solution. a)
1 2  2 −1 1 0  2 4  6 −3  4 0 
2 A + 3B − 4 I = 2   + 3  −4 = + − 
3 2   −3 4  0 1  6 4   −9 12  0 4

2 + 6 − 4 4 − 3 + 0   4 1 
= = 
 6 − 9 + 0 4 + 12 − 4   −3 12 

1 2   2 −1  2 − 6 −1 + 8   −4 7   −4 0 
b) AB =  = =  ( AB ) = 
T
     
3 2  −3 4   6 − 6 −3 + 8  0 5   7 5
1 3   2 −3
AT =   and BT =  
2 2  −1 4 
 2 −3 1 3  2 − 6 6 − 6   −4 0
 BT AT =   = = 
 −1 4   2 2   −1 + 8 −3 + 8  7 5 

 ( AB ) = BT AT .
T