52 Chapter 1 Systems of Linear Equations and Matrices

Working with Technology 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, . . .

T1. Let A be the matrix the terms of which are commonly denoted as
 1 1
0 2 3
1  F0 , F1 , F2 , F3 , . . . , Fn , . . .
A=
4 0 1
5
1 1
After the initial terms F0 = 0 and F1 = 1, each term is the sum of
6 7
0 the previous two; that is,
Discuss the behavior of Ak as k increases indefinitely, that is, as
k → `. Fn = Fn−1 + Fn−2

T2. In each part use your technology utility to make a conjecture Confirm that if
   
about the form of An for positive integer powers of n. F2 F1 1 1
    Q= =
a 1 cos θ sin θ F1 F0 1 0
(a) A = (b) A =
0 a − sin θ cos θ then  
Fn+1 Fn
T3. The Fibonacci sequence (named for the Italian mathematician Q =n

Leonardo Fibonacci 1170–1250) is

Fn F0

1.5 Elementary Matrices and a Method for Finding A−1

In this section we will develop an algorithm for finding the inverse of a matrix, and we will
discuss some of the basic properties of invertible matrices.

In Section 1.1 we defined three elementary row operations on a matrix A:

1. Multiply a row by a nonzero constant c.
2. Interchange two rows.
3. Add a constant c times one row to another.
It should be evident that if we let B be the matrix that results from A by performing one
of the operations in this list, then the matrix A can be recovered from B by performing
the corresponding operation in the following list:
1. Multiply the same row by 1/c.
2. Interchange the same two rows.
3. If B resulted by adding c times row ri of A to row rj , then add −c times rj to ri .
It follows that if B is obtained from A by performing a sequence of elementary row
operations, then there is a second sequence of elementary row operations, which when
applied to B recovers A (Exercise 33). Accordingly, we make the following definition.

DEFINITION 1 Matrices A and B are said to be row equivalent if either (hence each)
can be obtained from the other by a sequence of elementary row operations.

Our next goal is to show how matrix multiplication can be used to carry out an
elementary row operation.

DEFINITION 2 A matrix E is called an elementary matrix if it can be obtained from

an identity matrix by performing a single elementary row operation.
 1.5 Elementary Matrices and a Method for Finding A−1 53

E X A M P L E 1 Elementary Matrices and Row Operations

Listed below are four elementary matrices and the operations that produce them.
 
1 0 0 0    
  1 0 3 1 0 0
1 0 0 0 0 1
     
  0 1 0 0 1 0
0 −3 0 0 1 0
0 0 1 0 0 1
0 1 0 0

æ
Multiply the Interchange the Add 3 times Multiply the
second row of second and fourth the third row of first row of
I2 by −3. rows of I4 . I3 to the first row. I3 by 1.

The following theorem, whose proof is left as an exercise, shows that when a matrix A
is multiplied on the left by an elementary matrix E , the effect is to perform an elementary
row operation on A.

THEOREM 1.5.1 Row Operations by Matrix Multiplication

If the elementary matrix E results from performing a certain row operation on Im and
if A is an m × n matrix, then the product EA is the matrix that results when this same
row operation is performed on A.

E X A M P L E 2 Using Elementary Matrices

Consider the matrix
 
1 0 2 3
 
A = 2 −1 3 6
1 4 4 0

and consider the elementary matrix

 
1 0 0
 
E = 0 1 0
3 0 1

which results from adding 3 times the first row of I3 to the third row. The product EA is
Theorem 1.5.1 will be a use-
 
1 0 2 3
ful tool for developing new re-  
sults about matrices, but as a EA = 2 −1 3 6
practical matter it is usually 4 4 10 9
preferable to perform row op-
erations directly. which is precisely the matrix that results when we add 3 times the first row of A to the
third row.

We know from the discussion at the beginning of this section that if E is an elementary
matrix that results from performing an elementary row operation on an identity matrix
I , then there is a second elementary row operation, which when applied to E produces
I back again. Table 1 lists these operations. The operations on the right side of the table
are called the inverse operations of the corresponding operations on the left.
54 Chapter 1 Systems of Linear Equations and Matrices

Table 1

Row Operation on I Row Operation on E

That Produces E That Reproduces I

Multiply row i by c  = 0 Multiply row i by 1/c

Interchange rows i and j Interchange rows i and j

Add c time row i to row j Add −c times row i to row j

E X A M P L E 3 Row Operations and Inverse Row Operations

In each of the following, an elementary row operation is applied to the 2 × 2 identity
matrix to obtain an elementary matrix E , then E is restored to the identity matrix by
applying the inverse row operation.
     
1 0 1 0 1 0
−→ −→
0 1 0 7 0 1

æ
Multiply the second Multiply the second
row by 7. row by 17 .

     
1 0 0 1 1 0
−→ −→
0 1 1 0 0 1
æ

æ
Interchange the first Interchange the first
and second rows. and second rows.
     
1 0 1 5 1 0
−→ −→
0 1 0 1 0 1
æ

Add 5 times the Add −5 times the

second row to the second row to the
first. first.

The next theorem is a key result about invertibility of elementary matrices. It will be
a building block for many results that follow.

THEOREM 1.5.2 Every elementary matrix is invertible, and the inverse is also an ele-
mentary matrix.

Proof If E is an elementary matrix, then E results by performing some row operation

on I . Let E0 be the matrix that results when the inverse of this operation is performed
on I . Applying Theorem 1.5.1 and using the fact that inverse row operations cancel the
effect of each other, it follows that
E0 E = I and EE0 = I
Thus, the elementary matrix E0 is the inverse of E .

Equivalence Theorem One of our objectives as we progress through this text is to show how seemingly diverse
ideas in linear algebra are related. The following theorem, which relates results we
have obtained about invertibility of matrices, homogeneous linear systems, reduced row
 1.5 Elementary Matrices and a Method for Finding A−1 55

echelon forms, and elementary matrices, is our first step in that direction. As we study
new topics, more statements will be added to this theorem.

THEOREM 1.5.3 Equivalent Statements

If A is an n × n matrix, then the following statements are equivalent, that is, all true or
all false.
(a) A is invertible.
(b) Ax = 0 has only the trivial solution.
(c) The reduced row echelon form of A is In .
(d ) A is expressible as a product of elementary matrices.

Proof We will prove the equivalence by establishing the chain of implications:

The following figure illustrates
(a ) ⇒ (b ) ⇒ (c ) ⇒ (d ) ⇒ (a ).
visually that from the se-
quence of implications
(a) ⇒ (b) Assume A is invertible and let x0 be any solution of Ax= 0. Multiplying both
(a ) ⇒ (b ) ⇒ (c ) ⇒ (d ) ⇒ (a )
sides of this equation by the matrix A−1 gives A−1 (Ax0 ) = A−1 0, or (A−1 A)x0 = 0, or
we can conclude that I x0 = 0, or x0 = 0. Thus, Ax = 0 has only the trivial solution.
(d ) ⇒ (c ) ⇒ (b ) ⇒ (a )
(b) ⇒ (c) Let Ax = 0 be the matrix form of the system
and hence that
(a ) ⇔ (b ) ⇔ (c ) ⇔ (d ) a11 x1 + a12 x2 + · · · + a1n xn = 0
(see Appendix A).
a21 x1 + a22 x2 + · · · + a2n xn = 0
.. .. .. .. (1)
(a) . . . .
an1 x1 + an2 x2 + · · · + ann xn = 0
and assume that the system has only the trivial solution. If we solve by Gauss–Jordan
(d) (b)
elimination, then the system of equations corresponding to the reduced row echelon
form of the augmented matrix will be

(c) x1 =0
x2 =0
.. (2)
.
xn = 0
Thus the augmented matrix
 
a11 a12 ··· a 1n 0
a a22 ··· a2 n 0
 21 
 . .. .. .. 
 .. . . .
an1 a n2 ··· ann 0

for (1) can be reduced to the augmented matrix

 
1 0 0 ··· 0 0
 
0 1 0 ··· 0 0
 
0 0 1 ··· 0 0
 
 .. .. .. .. .. 
. . . . .
0 0 0 ··· 1 0
56 Chapter 1 Systems of Linear Equations and Matrices

for (2) by a sequence of elementary row operations. If we disregard the last column (all
zeros) in each of these matrices, we can conclude that the reduced row echelon form of
A is In .

(c) ⇒ (d ) Assume that the reduced row echelon form of A is In , so that A can be reduced
to In by a finite sequence of elementary row operations. By Theorem 1.5.1, each of these
operations can be accomplished by multiplying on the left by an appropriate elementary
matrix. Thus we can find elementary matrices E1 , E2 , . . . , Ek such that

E k · · · E 2 E 1 A = In (3)

By Theorem 1.5.2, E1 , E2 , . . . , Ek are invertible. Multiplying both sides of Equation (3)

on the left successively by Ek−1 , . . . , E2−1 , E1−1 we obtain

A = E1−1 E2−1 · · · Ek−1 In = E1−1 E2−1 · · · Ek−1 (4)

By Theorem 1.5.2, this equation expresses A as a product of elementary matrices.

(d ) ⇒ (a) If A is a product of elementary matrices, then from Theorems 1.4.7 and 1.5.2,
the matrix A is a product of invertible matrices and hence is invertible.

A Method for Inverting As a first application of Theorem 1.5.3, we will develop a procedure (or algorithm) that
Matrices can be used to tell whether a given matrix is invertible, and if so, produce its inverse. To
derive this algorithm, assume for the moment, that A is an invertible n × n matrix. In
Equation (3), the elementary matrices execute a sequence of row operations that reduce
A to In . If we multiply both sides of this equation on the right by A−1 and simplify, we
obtain
A −1 = E k · · · E 2 E 1 In
But this equation tells us that the same sequence of row operations that reduces A to In
will transform In to A−1 . Thus, we have established the following result.

Inversion Algorithm To find the inverse of an invertible matrix A, find a sequence of

elementary row operations that reduces A to the identity and then perform that same
sequence of operations on In to obtain A−1 .

A simple method for carrying out this procedure is given in the following example.

E X A M P L E 4 Using Row Operations to Find A−1

Find the inverse of
 
1 2 3
 
A = 2 5 3
1 0 8

Solution We want to reduce A to the identity matrix by row operations and simultane-
ously apply these operations to I to produce A−1 . To accomplish this we will adjoin the
identity matrix to the right side of A, thereby producing a partitioned matrix of the form

[A | I ]
 1.5 Elementary Matrices and a Method for Finding A−1 57

Then we will apply row operations to this matrix until the left side is reduced to I ; these
operations will convert the right side to A−1 , so the final matrix will have the form
[I | A−1 ]
The computations are as follows:
 
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 We added −2 times the first
row to the second and −1 times
0 −2 5 −1 0 1 the first row to the third.

 
1 2 3 1 0 0
 
0 1 −3 −2 1 0 We added 2 times the
second row to the third.
0 0 −1 −5 2 1
 
1 2 3 1 0 0
 
0 1 −3 −2 1 0 We multiplied the
third row by −1.
0 0 1 5 −2 −1
 
1 2 0 −14 6 3
 
0 1 0 13 −5 − 3 We added 3 times the third
row to the second and −3 times
0 0 1 5 −2 −1 the third row to the first.

 
1 0 0 −40 16 9
 
0 1 0 13 −5 − 3 We added −2 times the
second row to the first.
0 0 1 5 −2 −1
Thus,  
−40 16 9
 
A−1 =  13 −5 − 3
5 −2 −1

Often it will not be known in advance if a given n × n matrix A is invertible. However,

if it is not, then by parts (a) and (c) of Theorem 1.5.3 it will be impossible to reduce A
to In by elementary row operations. This will be signaled by a row of zeros appearing
on the left side of the partition at some stage of the inversion algorithm. If this occurs,
then you can stop the computations and conclude that A is not invertible.

E X A M P L E 5 Showing That a Matrix Is Not Invertible

Consider the matrix  
1 6 4
 
A= 2 4 − 1
−1 2 5
58 Chapter 1 Systems of Linear Equations and Matrices

Applying the procedure of Example 4 yields

 
1 6 4 1 0 0
 
 2 4 −1 0 1 0
−1 2 5 0 0 1
 
1 6 4 1 0 0
 
 0 −8 −9 −2 1 0 We added −2 times the first
row to the second and added
0 8 9 1 0 1 the first row to the third.
 
1 6 4 1 0 0
 
 0 −8 −9 −2 1 0 We added the second
row to the third.
0 0 0 −1 1 1
Since we have obtained a row of zeros on the left side, A is not invertible.

E X A M P L E 6 Analyzing Homogeneous Systems

Use Theorem 1.5.3 to determine whether the given homogeneous system has nontrivial
solutions.
(a) x1 + 2 x2 + 3x3 = 0 (b) x1 + 6x2 + 4 x3 = 0
2 x1 + 5x2 + 3x3 = 0 2 x1 + 4 x2 − x3 = 0
x1 + 8x3 = 0 −x1 + 2x2 + 5x3 = 0
Solution From parts (a) and (b) of Theorem 1.5.3 a homogeneous linear system has
only the trivial solution if and only if its coefficient matrix is invertible. From Examples 4
and 5 the coefficient matrix of system (a) is invertible and that of system (b) is not. Thus,
system (a) has only the trivial solution while system (b) has nontrivial solutions.

Exercise Set 1.5

In Exercises 1–2, determine whether the given matrix is ele- In Exercises 3–4, find a row operation and the corresponding
mentary. elementary matrix that will restore the given elementary matrix to
    the identity matrix.
1 0 −5 1
1. (a) (b)  
−5 1 1 0   −7 0 0
1 −3  
  3. (a) (b)  0 1 0
  2 0 0 2 0 1
1 1 0 0 0 1
0 1 0 0
   
(c) 0 0 1 (d)    
0 0 1 0   0 0 1 0
0 0 0 1 0 0 0
0 0 0 1    1 0 0

(c)  0 1 0 (d)  
  1 0 0 0
  −5 0 1
0 0 1 0 0 0 1
1 0  
2. (a) √ (b) 0 1 0
0 3  
1 0 0   1 0 0
1 0  
    4. (a) (b) 0 1 0
1 0 0 −1 0 0 −3 1
    0 0 3
(c) 0 1 9 (d)  0 0 1
   
0 0 1 0 1 0 0 0 0 1 1 0 − 17 0
0 1 0 0 0 1 0 0
   
(c)   (d)  
0 0 1 0 0 0 1 0
1 0 0 0 0 0 0 1