2.

2 The Inverse of a matrix

Motivation

b
Reclnumbers: aX = b ⇒ X = a

Slow motion: aX = b

Suppose a−1 is some number for which a−1 a = 1.

aX = b
a−1(aX) = a−1(b)
(a−1a)X = a−1b
1X = a−1b
b
X = a−1b → a

1
matrices: AX = B

Suppose A−1 is another matrix where

A−1A = I.

AX = B

A−1(AX) = A−1(B)

(A−1A)X = A−1B

IX = A−1B

X = A−1B

2
Definintion Suppose A is an n×n matrix
(square).

If there is another n×n matrix A−1 such

that

A−1A = I AND AA−1 = I

then A is called invertible with inverse A−1.

If such a matric A−1 does not exist, then A

is called

not invertible or singular

3
How to Find A−1 (if it exists)

EX (Slow motion)
" #
1 2
A=
3 4

Is there a matrix A−1 where AA−1 = A−1 A=I ?

AA−1 = I

" # " #
1 2 h i 1 0
C~1 C~2 =
3 4 0 1

" # " # " # " #

1 2 1 1 2 0
C~1 = C~2 =
3 4 0 3 4 1
" . # " . #
1 2 .. 1 1 2 .. 0
. .
3 4 .. 0 3 4 .. 1

"& .
.. #
−3 1 2 . 1 0
.
3 4 .. 0 1
" .. #
1 2 . 1 0
.
0 −2 .. −3 1 −3R1 + R2

4
 " .. #
1 0 . −2 1 R2 + R1
.
0 −2 .. −3 1
" . #
1 0 .. −2 1
.
0 1 .. 3 −1 2 2
−1
2 R2
" # " #
−2 1
C~1 = 3
C~1 = −1
2 2

So
" #
h i −2 1
A−1 = C~1 C~2 = 3 −1
2 2

Summary:
" #
1 2
A=
3 4
" . #
1 2 .. 1 0 h i
..
. A . I
3 4 .. 0 1
" .. #
1 0 . −2 1 h i
. −1
.
. I . A
0 1 .. 32 −1
2
" #
−2 1
A−1 = 3 −1
2 2

5
  
1 2 0
 
EX A= −2 1 1 


0 5 1

Find A−1 if it exists.

AA−1 = I
 
1 2 0 | 1 0 0
 
2  −2 1 1 | 0 1 0 
 
0 5 1 | 0 0 1
 
1 2 0 | 1 0 0
 
−1 0 5 1 | 2 1 0 2R1 + R2
 
0 5 1 | 0 0 1
 
1 2 0 | 1 0 0
 
 0 5 1 | 2 1 0  −R2 + R3
 
0 0 0 | −2 −1 1

This matrix tells us that there is no matrix

A−1 where AA−1 = I.
A is not invertible (singular)

6
A Formula for 2×2 Matrix Inverses

   −cR1 −ca −cb −c 0

a b a b ... 1 0 aR2 ac ad 0 9
A=  
c d c d ... 0 1 −−− −−− −−− −−
Add 0 ad − bc −c 9

 
a b .
.. 1 0
 
.
0 ad − bc .. −c a

For A−1 to exist, we need ad − bc 6= 0

 
.
a b .. 1 0
 
.
0 1 .. ad−bc ad−bc
−c a

⇓ One more step to row reduce

 
1 0 ... d −b
 ad−bc ad−bc 
0 1 ... −c
ad−bc
a
ad−bc

7
  
d −b
A −1
= ad−bc ad−bc 
−c a
ad−bc ad−bc
OR

A
   
d −b a b
A−1 = 1    
ad−bc
−c a c d

(1) Reverse diagonal entries

(2) Change sign of the nondiagonal entries

1
(3) Multiple by ad−bc
 
1 2
OR A =  
2 6

det(A) = ad − bc = (1)(6) − (2)(2) = 2

   
6 −2 3 −1
A−1 = 12  = 
1
−2 1 −1 2

8
Using A−1 to Solve ~ = ~b
AX

EX

X1 + 3X2 = 1
2X1 + 4X2 = 6
" # " # " #
1 3 X1 1
=
2 4 X2 6
" # " # " # " # " #
1 3 1 3 X1 1 3 1
( )−1 = ( )−1
2 4 2 4 X2 2 4 6
" # " # " # " #
1 0 X1 1 4 −3 1
= (1)(4)−(2)(3)
0 1 X2 −2 1 6
" # " # " #
X1 1 −14 7
= −2 = )
X2 4 −2

The soln. X1 = 7, X2 = −2

9
Properties of Matrix Inverses

~ = ~b is consistent with
(1) If A is invertible , then AX
exactly one solution .

~ = A−1~b
X

(2) If A is invertible with inverse A−1 , then A−1 is also

invertible (A−1 )−1 = A

AA−1 = I
A−1 is the inverse of A
A is the inverse of A−1

(3) If A,B are two invertible n × n matrices, then AB

is another invertible matrix with

(AB)−1 = B −1 A−1

(AB)(?) = I

(AB)(B −1 A−1 ) = A(BB −1 )A−1

= A(I)A−1
= AA−1
= I

10
 (?)(AB) = I
(B −1 A−1 )(AB) = B −1 (A−1 A)B
= B −1 (I)B
= B −1 B
= I
So:

(AB)−1 = B −1 A−1

(4) If A is invertible, then so is AT and (AT )−1 = (A−1 )T

Check:

AT (A−1 )T = (A−1 A)T

= (I)T
= I
(A−1 )T AT = (AA−1 )T = (I)T = I

Page. of transpose (AB)T = B T AT

11