13- Characterization of Invertible Matrices VU

Lecture 13
Characterizations of Invertible Matrices

This chapter involves a few techniques of solving the system of n linear equations in n
unknowns and transformation associated with a matrix.
Solving Linear Systems by Matrix Inversion
Theorem
Let A be an n × n invertible matrix. For any b ∈ Rn , the equation Ax = b …..(1) has the
unique solution i.e. x = A-1b.
Proof
Since A is invertible and b∈ Rn be any vector. Then, we must have a matrix A−1b which
is a solution of eq. (1) .i.e. Ax = A (A –1b) = Ib = b.
Uniqueness
For uniqueness, we assume that there is another solution u. Indeed, it is a solution of
eq.(1) so it must be u = A-1b, it means x = A-1b = u. This shows that u = x.
Theorem
Let A and B be the square matrices such that AB = I. Then, A and B are invertible with
B = A-1 and A = B-1
Example 1
Solve the system of linear equations
2 x1 + x2 + x3 =
1
5 x1 + x2 + 3 x3 =
3
x1 + 4 x3 =
6

Solution
Consider the linear system
2 x1 + x2 + x3 =
1
5 x1 + x2 + 3 x3 =
3
x1 + 4 x3 =
6

The Matrix form of system is Ax = b, where

©Virtual University Of Pakistan 143

13- Characterization of Invertible Matrices VU

2 1 1  x1  1 
= 
A 5 1 = 
3 , x = x2  , b 3
 
1 0 4   x3  6 

Here, det (A) = 2(4) – 1 (20 – 3) + 1(0 – 1) = 8 – 17 – 1 = -10 ≠ 0

So, A is invertible. Now, we apply the inversion algorithm:

2 1 1 1 0 0
5 1 3 0 1 0 

1 0 4 0 0 1 

1 0 4 0 0 1
5 1 3 0 1 0 

 2 1 1 1 0 0 
1 0 4 0 0 1
0 1 −17 0 1 −5 , − 5 R1 + R2 , − 2 R1 + R3

0 1 −7 1 0 −2 
1 0 4 0 0 1
0 1 −17 0 1 −5 , − R2 + R3 ,

0 0 10 1 −1 3 
 
1 0 4 0 0 1
  R3
0 1 −17 0 1 −5 ,
10
 1 −1 3
0 0 1 
 10 10 10 

 −2 2 −1 
1 0 0 5 5 5
 
0 1 0 17 −7 1 
, − 4 R3 + R1 , −17 R3 + R2
 10 10 10 
 
0 0 1 1 −1 3 
 10 10 10 

©Virtual University Of Pakistan 144

13- Characterization of Invertible Matrices VU

 −2 2 −1 
5 5 5
 
−1  17 −7 1 
Hence, A =
 10 10 10 
 
1 −1 3 
 10 10 10 
 −2 2 −1   −2 
5 5 5  1  5
   
−7 1   
b A=  1 
−1 −1 17
Thus, the solution of the linear system is=
x A= =
3
 10 10 10    5
  6  
1 −1 3    8 
 10 10 10   5 
 x1 + 2 x2 + 3x3 = 5

Example # 2 Solve the system 2 x1 + 5 x2 + 3 x3 =
3 by inversion method.

 x1 + 8 x3 =
17
 x1 + 2 x2 + 3 x3 = 5

Solution Consider the linear system  2 x1 + 5 x2 + 3 x3 =
3

 x1 + 8 x3 =
17

This system can be written in matrix form as Ax = b, where

 1 2 3  x1  5
= 
A 2 5 = 
3 , x = x2  , b  3 

1 0 8  x3  17 
Here, [det (A)] = 40 – 2 (16 – 3) + 3(0 – 5) = 40 – 26 – 15 = -1 ≠ 0
Therefore, A is invertible.

Now, we apply the inversion algorithm:

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 1 −3 −2 1 0  −2 R1 + R2 , −1R1 + R3

0 −2 5 −1 0 1 

©Virtual University Of Pakistan 145

13- Characterization of Invertible Matrices VU

1 2 3 1 0 0 
0 1 −3 −2 1 0  2R2 + R3

0 0 −1 −5 2 1 
1 2 3 1 0 0
0 1 −3 −2 1 0  −1R3

0 0 1 5 −2 −1
1 2 0 −14 6 3 
0 1 0 13 −5 −3 3R3 + R2 , −3R3 + R1

0 0 1 5 −2 −1
1 0 0 −40 16 9 
0 1 0 13 −5 −3 −2R2 + R1

0 0 1 5 −2 −1
 −40 16 9 
Hence, A=  13 −5 −3
−1

 5 −2 −1
 −40 16 9   5   1 
Thus, the solution of the linear system is x =A b = 13 −5 −3  3  =−
−1  
     1
 5 −2 −1 17   2 
Thus x1 =1, x2 = −1, x3 =2.

Note: It is only applicable when the number of equations = number of unknown and fails
if given matrix is not invertible.
Example 3
Solve the system of linear equation
x1 + 6 x2 + 4 x3 =
2
2 x1 + 4 x2 − x3 =
3
− x1 + 2 x2 + 5 x3 =
3

Solution
The matrix coefficient
1 6 4
=A  2 4 −1
 −1 2 5 

det (A) = 1(20 + 2) – 6(10 – 1) + 4 (4+4)

©Virtual University Of Pakistan 146

 13- Characterization of Invertible Matrices VU

= 22 – 54 + 32
= 0
Thus, A is not invertible. Hence, the inversion method fails.
Solution for the system of a n × n Homogeneous linear equations with
an Invertible Coefficient Matrix:-
Let see if the system is considered as homogeneous then what does the above theorem
say?
Theorem:-
Let Ax = 0 be a homogeneous linear system of n equations and n unknowns. Then, the
coefficient matrix A is invertible iff this system has only a trivial solution.
Example 4
State whether the following system of linear equation has a solution or not?

2 x1 + x2 + 3 x3 =
0
x1 − 3 x2 + x3 =
0
x1 − 4 x3 =
0

Solution
We see
2 1 3 
A 1 −3 1  is an invertible matrix (det (A) ≠ 0 )
=
1 0 −4 

Thus, this homogeneous linear system has only the trivial solution.

Example 5

 x1 + x2 + x3 = 8

Solve  2 x2 + 3 x3 = 24
 5x + 5x + x = 8
 1 2 3

Solution This system can be written in matrix form as Ax = b, where

1 1 1  x1  8 
= 
A 0 2 = 
3 , x = x2  , b  24 

5 5 1  x3  8 

©Virtual University Of Pakistan 147

13- Characterization of Invertible Matrices VU

Here [det (A)] = 1(2 – 15) – 1(0 – 15) + 1(0 – 10) = -13 + 15 – 10 = -8 ≠ 0
Therefore, A is invertible.

Now, we apply the inversion algorithm:

A I3
1 1 1 1 0 0
0 2 3 0 1 0 

5 5 1 0 0 1 
1 1 1 1 0 0
0 2 3 0 1 0  −5R1 + R3

0 0 −4 −5 0 1 
1 1 1 1 0 0
 3 1  1
0 1 0 0  R2
 2 2  2
0 0 −4 −5 0 1 

 1 1 
1 0 − 2 1 −
2
0
 
0 1 3 0
1
0 −1R2 + R1
 2 2 
0 0 −4 −5 0 1 

 
 1 1 
1 0 − 2 1 −
2
0 
 
0 1 3 0
1
0 
1
− R3
 2 2  4
 5 1 
0 0 1 0 −
 4 4 
 1 1 
1 0 − 2 1 −
2
0 
 
0 1 0 −
15 1 3  3
− R3 + R2
 8 2 8  2
 
0 0 1 5 1
0 − 
 4 4 
 13 1 1
1 0 0 − − 
8 2 8
 
0 1 0 −
15 1 3  1
R3 + R1
 8 2 8  2
 
0 0 1 5 1
0 −
 4 4 

©Virtual University Of Pakistan 148

13- Characterization of Invertible Matrices VU

 13 1 1
 8 − − 
2 8
 
Hence, −1
A = −  15 1 3 
 8 2 8 
 
 5 0 −
1
 4 4 
Thus, the solution of the linear system is
 13 1 1
 8 − −
2 8  8  0 
 
−1  15 1 3    
x= A b= − 24 =0
 8 2 8    
  8  8 
 5 1    
0 −
 4 4 
Thus= x1 0,= x2 0,= x3 8 .

Theorem (Invertible Matrix Theorem) Let A be a square n × n matrix. Then the

following statements are equivalent. (Means if any one holds then all are true).
(a) A is an invertible matrix.
(b) A is row equivalent to the n × n identity matrix.
(c) A has n pivot positions.
(d) The equation Ax = 0 has only the trivial solution.
(e) The columns of A form a linearly independent set.
(f) The linear transformation x → Ax is one-to-one.
(g) The equation Ax = b has at least one solution for each b in Rn.
(h) The columns of A span Rn.
(i) The linear transformation x → Ax maps Rn onto Rn.
(j) There is a n × n matrix C such that CA = I.
(k) There is a n × n matrix D such that AD = I.
(l) AT is an invertible matrix.

Example 6
1 0 −4 
Show that the matrix A = 1 1 5  is invertible by using Invertible Matrix Theorem
0 1 2 

©Virtual University Of Pakistan 149

 13- Characterization of Invertible Matrices VU

Solution
By row equivalent,
1 0 −4 
=A 1 1 5  , − R1 + R2

0 1 2 
1 0 −4 
0 1 9  , − R2 + R3

0 1 2 
1 0 −4 
0 1 9 
 
0 0 −7 

It shows that A has three pivot positions and hence is invertible, by the Invertible Matrix
Theorem (c).
Example 7

Use the Invertible Matrix Theorem to decide if A is invertible, where

 1 0 −2 
= A  3 1 −2 
 −5 −1 9 
Solution
1 0 −2  1 0 −2 
A = 0 1 4   0 1 4  Here, A has three pivot positions and hence is
0 −1 −1 0 0 3 
invertible by the Invertible Matrix Theorem (c).

Example 7 Find At and show that At is an invertible matrix.

1 2 1 0
2 4 1 1 
A=
1 0 1 1
 
0 1 1 1
Solution

1 2 1 0
2 4 0 1 
A =
t

1 1 1 1
 
0 1 1 1
©Virtual University Of Pakistan 150
13- Characterization of Invertible Matrices VU

Now, by row equivalent of A,

R2 − 2 R1 , R3 − R1 R23
1 2 1 0  1 2 1 0  1 2 1 0
2 4 1 1  0 0 −1 1  0 −2 0 1 
A=  
1 0 1 1  0 −2 0 1  0 0 −1 1
     
0 1 1 1  0 1 1 1  0 1 1 1

1
− R2 R4 − R2 (−1) R −R +R
2 3 3 4
1 2 0  1
1 2 1 0  1 2 1 0
1 2 0  
1
 1  −1   −1 
0 1  0 1 0 −  0 1 0 0 1 0
1 0 −   2  2  2
 2    
0 0 −1 1  0 −1  0 −1 
0 −1 1  
0 0 1 0 1
    
 
1  0
3   3  5
0 1 1
 0 1 0 0 1 0 0 0
2   2   2 

Here A has 4 pivot positions so by Invertible Matrix Theorem (c) A is invertible. Thus,
by (l) At is invertible.
Example 8

Use the Invertible Matrix Theorem to decide if A is invertible, where

1 2 1 0 
 2 4 1 2
A= 
1 0 1 2 
 
0 1 1 2

Solution
R2 − 2 R1 , R3 − R1 R23
1 2 1 0  1 2 1 0  1 2 1 0
2 4 1 2  0 0 −1 2  0 −2 0 2 
A=  
1 0 1 2  0 −2 0 2  0 0 −1 2
     
0 1 1 2 0 1 1 2  0 1 1 2

©Virtual University Of Pakistan 151

 13- Characterization of Invertible Matrices VU

1
− R2 R4 − R2 (−1) R −R +R
2 3 3 4
1 2 10  1 2 0  1
1 2 1 0  1 2 1 0
0 1 0−1 0 1 −1 0
0 1 0 −1 0 1 0 −1
   
0 0 −1 2  0 0 −1 2  0 0 1 −2  0 0 1 −2 
       
0 1 1 2  0 0 1 3  0 0 1 3  0 0 0 5

Here, A has 4 pivot positions and hence is invertible by Invertible Matrix Theorem (c).

Solving Multiple Linear Systems with a Common Coefficient Matrix

This technique is used in solving a sequence of linear systems
=
Ax1 b=
1 , Ax2 b2 ,  =
, Axk bk (1)

where coefficient matrix A remains same and off course if it is invertible, then we have a
sequence of solutions. i.e.
−1
=x1 A= b1 , x2 A−1b2 , =
 , xk A−1bk .

Find a Matrix from Linear Transformation

We can find a matrix corresponding to every transformation. In this section we will learn
how to find a matrix attached with a linear transformation.
Example 9
Let L be the linear transformation from R2 to P2 (Polynomials of order 2) defined by
T(x, y) = x y t + (x + y)t2
Find the matrix representing T with respect to the standard bases.
Solution
Let A = {(1,0),(0,1)} be the basis of R2, then
T(1,0) = t2 = (0,0,1) (This triple represents the coefficients of polynomial t2)
i.e. t 2 =
0.1 + 0.t + 1.t 2
Similarly, T (0, 1) = t2 = (0,0,1). Hence, the matrix is given by
0 0
 
A=  0 0
 
1 1

Now, we will proceed with a more complicated example.

©Virtual University Of Pakistan 152

13- Characterization of Invertible Matrices VU

Example 10
Let T be the linear transformation from R2 to R2 such that T(x, y) = (x , y + 2x). Find a
matrix A for T.
Solution
This matrix is found by finding T (1, 0) = (1, 2) and T (0, 1) = (0, 1)The matrix
1 0
is A =  .
2 1
Important Note
It should be clear that the Invertible Matrix Theorem applies only to square matrices. For
example, if the columns of a 4 × 3 matrix are linearly independent, we cannot use the
Invertible Matrix Theorem to conclude any thing about the existence or nonexistence of
solutions to equations of the form Ax = b.
Invertible Linear Transformations
Recall that matrix multiplication corresponds to composition of linear transformations.
When a matrix A is invertible, the equation A-1Ax = x can be viewed as a statement about
linear transformations. See Figure 2.

Multiplication

by A

x• • Ax

Multiplication

by A-1
Figure 2 A-1 transforms Ax back to x

Definition
A linear transformation T : R → R be a linear Transformation and A be a standard
n n

matrix for T. Then, T is invertible if and only if A is an invertible matrix in that case
linear transformation S given by S(x) = A-1x is a unique function satisfying (1) and (2)
S (T ( x)) = x ∀ x ∈ R n (1)
T ( S ( x)) = x ∀ x ∈ R n
(2)

©Virtual University Of Pakistan 153

13- Characterization of Invertible Matrices VU

Important Note
If the inverse of a linear transformation exits then it is unique.
Proposition
Let T : R n → R m be linear transformation, given as=
T ( x) Ax , ∀ x ∈ R n , where A is
a m × n matrix. The mapping T is invertible if the system y = Ax has a unique solution.
Case 1:
If m < n , then the system Ax = y has either no solution or infinitely many solutions, for
any y in Rm. Therefore, y = Ax is non-invertible.
Case 2:
If m = n , then the system Ax = y has a unique solution if and only if Rank (A) = n.
Case 3:
If m > n , then the transformation y = Ax is non-invertible because we can find a vector
y in Rm such that Ax = y is inconsistent.

©Virtual University Of Pakistan 154

13- Characterization of Invertible Matrices VU

Exercises

1.Solve the system of linear equations by inverse matrix method.

x1 + x2 + 4 x3 =
2
2 x2 + 3x3 =
4
5 x1 + x2 − x3 =
3

1 2  5   −2  3  1 
2. =
Let A  =  , b1 =  , b2  =  =
, b3   , and b4   .
3 8  7   −2  5  7 
(a) Find A-1 and use it to solve the equations Ax = b1, Ax = b2, Ax = b3, Ax = b4.
(b) Solve the four equations in part (a) by row reducing the augmented matrix
[A b1 b2 b3 b4 ] .

3. (a) Solve the two systems of linear equations

x1 + 2 x2 + x3 =
−1
x1 + 3 x2 + 2 x3 =
3
x2 + 2 x3 =
4
&
x1 + 2 x2 + x3 =
0
x1 + 3 x2 + 2 x3 =
0
x2 + 2 x3 =
4
by row reduction.
(b) Write the systems in (a) as Ax = b1 and Ax = b2, and then solve each of them by the
method of inversion.
Determine which of the matrices in exercises 4 to 10 are invertible?

5 0 3 2 3 4 5 −9 3 
 −4 16 
4.   5. 7 0 2  6.  2 3 4  7. 0 3 4 
 3 −9  9 0 1   2 3 4  1 0 3 

©Virtual University Of Pakistan 155

13- Characterization of Invertible Matrices VU

1 3 0 −1 1 0 0 0  7 −6 −4 1 
 0 1 −2 −1 2 5 0 0   −5 1 0 −2 
8.   9.  10.  
 −2 −6 3 2  3 6 8 0 10 11 7 −3
     
 3 5 8 −3 4 7 9 10  19 9 7 1 
5 4 3 6 3
7 6 5 9 5 

11.  8 6 4 10 4
 
 9 −8 9 −5 8
10 8 7 −9 7 
12. Suppose that A and B are n × n matrices and the equation ABx = 0 has a nontrivial
solution. What can you say about the matrix AB?
13. What can we say about a one-to-one linear transformation T from Rn into Rn?
14. Let T : R 2 → R 2 be a linear transformation given as T ( x) = 5 x , then find a matrix A of
linear transformation T.
In exercises 15 and 16, T is a linear transformation from R2 into R2. Show that T is
invertible.
15. T ( x1 , x2 ) =(−5 x1 + 9 x2 , 4 x1 − 7 x2 )

16. T ( x1 , x2 ) = (6 x1 − 8 x2 , −5 x1 + 7 x2 )

17. Let T : R → R be a linear transformation and let A be the standard matrix for T.
n n

Then T is invertible if and only if A is an invertible matrix.

©Virtual University Of Pakistan 156