Finding inverse using LU decomposition (section 4.6.

1)

Example. Find the inverse of the following upper triangular matrix:

 
2 4 6
 
 
U =  0 −1 −8 
 
0 0 96

Solution.

[U |I] =
   
2 4 6 1 0 0 2 4 6 1 0 0
 ( 1 )  −6R3 add to
−→ R1
   
 0 −1 −8 0 1 0  R3  0 −1 -8 0 1 0 
  96   8R add to
3 −→ R2
0 0 96 0 0 1 1
0 0 1 0 0 96

   
2 4 0 1 0 − 16
1
2 4 0 1 0 − 16
1
   
    add to
 0 -1 0 0 1 1  (−R2 )  0 1 0 0 −1 − 12
1  − 4R2 −→ R1
 12   
1 1
0 0 1 0 0 96
0 0 1 0 0 96

   
2 13 1 13
0 0 1 4 1 0 0 2
 48  1  2 96 
   
 0 1 0 0 −1 − 12
1  ( R1 )  0 1 0 0 −1 − 12
1  = [ identity | U −1 ]
  2  
1 1
0 0 1 0 0 96 0 0 1 0 0 96

and as soon as we get the identity matrix on the left-hand part, then the matrix on the
right-hand part is the inverse to the upper triangular matrix:
 
1 13
2
 2 96 
 
U −1 =  0 −1 − 12
1 
 
1
0 0 96

1
Example. Find the inverse of the following lower triangular matrix:
 
2 0 0
 
 
L =  8 −7 0 
 
4 9 −27

Solution.

[L|I] =
   
2 1
0 0 1 0 0 1 0 0 0 0
  (1 )  2  −8R1 add to
−→ R2
   
 8 −7 0 0 1 0  R1  8 −7 0 0 1 0 
  2   −4R add to
1 −→ R3
4 9 −27 0 0 1 4 9 −27 0 0 1

   
1 1
1 0 0 1 0 0 0 0 0 0
 ( 1 ) 2 2 
    add to
 0 -7 0 −4 1 0  − R2  0 1 0 4
− 17 0  − 19 R2 −→ R3
  7  7 
0 9 −27 −2 0 1 0 9 −27 −2 0 1

   
1 1
 1 0 0 2 0 0 ( ) 1 0 0 2 0 0

  1  
 0 1 − 17 0  −1
 − 27 R3  0  = [ identity | L ]
4
 0 0 1 0 4
− 71

7
  7 
0 0 -27 − 50 9 0 0 1 50
− 21
1
− 27
1
7 7 1 189

So:
 
1
0 0
 2 
−1  
L = 4
− 17 0 
 7 
50
189 − 21
1
− 27
1

Note. If the upper triangular matrix or lower triangular matrix has 1 all over the main
diagonal, then there is no need to apply the row operations to get the inverse, you only need
to change the signs of the off-diagonal elements.

2
Example.
 −1  
1 0 0 1 0 0
   
   
 3
1 0  = −2
3
1 0 
 2   
3 14 1 −3 −14 1

 −1  
1 4
− 56 1 − 43 5
 3    6
   
 0 1 3  = 0 1 − 37 
 7   
0 0 1 0 0 1

Note. The inverse operator has the following property:

A = BC ⇒ A−1 = C −1 B −1

Example. Find the inverse of the matrix A that has the LU decomposition:
  
1 0 0 2 4 6
  
  
A= 3
1 0   0 −1 −8 
 2  
3 14 1 0 0 96

Solution. Using our findings in the first example , we can write:

 −1  −1   
1 13
2 4 6 1 0 0 2 1 0 0
     2 96  
      3 
A−1 =  0 −1 −8   3
1 0  =  0 −1 − 12
1   −2 1 0 
   2    
0 0 96 3 14 1 0 0 1
96 −3 −14 1

 
− 32
93 5 13
 48 
96
 1 
=  7 1
− 12 
 4 6 
− 32
1
− 48
7 1
96

3
So here is twp-step procedure to find the inverse of a matrix A:
Step 1.. Find the LU decomposition A = LU (Gaussian form or the Crout form whichever
you are told to find)
Step 2.. Find the inverse of A−1 = U −1 L−1 by inverting the matrices U and L.