Chapter Solving Systems

of Linear Equations
Algebraically and
with Matrices
t the end of Chapter 4, we encountered systems of
linear equations and saw that they could be solved
SOLVING SYSTEMS
by. determining the point(s) of intersection of the
fo) omMINT
ele) Wale) Sms)4 graphs of the equations. In the event that a single point
THE SUBSTITUTION of intersection occurred, that point was identified as
METHOD the solution;and the system was called consistent.
lf the lines were parallel (and hence no intersection
Se AVINGRS AUIS
would occur), the system was called inconsistent, and
fo) Kel) Wale) NS
UsinG THE METHOD we said. there was no solution. In the rather rare in-
OF ELIMINATION stance where the two lines had the same graph, the
system was called consistent and dependent, and the
SUBSTITUTIONS T HAT solution set consisted of all points lying on their mutual
LEAD TO SYSTEMS OF graph.
MINN ele) Wale) NS
That visual. method of solving systems possesses in-
TNWEsxe)
DlUleunle)\
herent weaknesses. First, a high-quality graph must be
TO MatTRICES
in your possession, with lines drawn that are extremely
UsinGc MatrICcESs TO precise. Even the slightest imperfection in the graph
SOLVE SYSTEMS OF of one or both of the lines could lead you to an incor-
MINT -V;a hele) Wale) iS rect determination of the solution. Second, even with
a high-quality graph in hand, if the solution point were
such that its coordinates were not integer valued, ex-
act identification of the solution could prove to be dif-
ficult or impossible.

These issues lead us to consideration of methods

for solution that are immune from the ambiguity of
graphical analysis, methods that are wholly algebraic
and computational. Such methods are free from
the requirement that a high-quality graph be in our
 Solving Systems of Linear Equations Algebraically and with Matrices 163

possession and are also immune from the problem of identifying

noninteger coordinates.

We will consider several computational techniques, including the

methods of substitution and elimination and the matrix method. We
will see that these approaches offer us protection from the vagaries
of the graphing method and that they employ mathematical tech-
niques already well known to us.

When you have successfully completed the materials of this chapter, you will be
able to:

@ Solve a system of linear eregerenatey er by citeritae store Mera] fietatlsteree

@ Solve a system of linear equations by the method of
elimination.

@ Use substitutions to convert systems of equations into linear

systems and solve them.

@ Understand the basic definitions and operations of matrices.

@ Solve a system of linear eYeqerhateyel-melsbelmiretiau leer

5 . 1 SO) AVINIE! SYSTEMS OF LINEAR EQUATIONS

BY THE SUBSTITUTION METHOD
Let’s consider a system oflinear equations whose solution would be difficult to
determine using the method ofgraphing. You will note that there is nothing in
the appearance of the equations that suggests that we are headed for trouble by
attempting to solve by graphing; the obstacles arise only when the graph is in
our possession:

Die di
6x + 12y= 11

Suppose we attempt to solve the system by graphing and see what happens. You
are already familiar with the procedure for producing the graphs, so we will
suppress that work and present the result for analysis (Figure 5.1). You will note
that the point of intersection is somewhat difficult to determine. The coordi-
nates are clearly not integer valued, so we could start to make conjectures about
their exact values and check our guesses in the original equations of the system,
but that is rather risky and inefficient. If we suppose that the coordinates of
164 Chapter 5

Figure 5.1 Graph ofa

System of Linear Equations Cengage
2014
©
Learning

‘ | MiesWsa e
the intersection are & >), for instance, how do we know that they are not
2 4

oe 2 Or some other fractions that are close to = and >» We could make
20-9 2 4
guesses all day long and never find the actual answers, and that situation is
clearly unsatisfactory.
Let's now approach the problem from a purely computational angle. There
are, as weve mentioned, several techniques from which to choose, and none is
universally better than any of the others, and all will work, so it’s really a mat-
ter of our choosing one to work with. The first of the methods we will call the
method of substitution.
Substitution proceeds in the following manner:
1. Choose one equation to work with and in that equation isolate one of
the variables.
Go to the equation not chosen in step 1 and replace the variable you
isolated in step 1 by the expression to which you found it equal when the
variable was isolated.
3. Solve the resulting equation.
4. Recalling the result of step 1, where one variable was isolated, substitute
the solution found instep 3 and compute the value of the missing variable.
5. Combine your results into an ordered pair and check to see if it verifies
each of the original equations of the system. If it does not, then a compu-
tational error must exist somewhere in your work—examine your steps
to find your error.
Those steps may sound challenging, but we'll be up to it. Let's recall the system
given earlier and apply the process.

BRXAMPLE 5.1 Solve the system by the method of substitution:

2x 3y—=3~
6x + 12y=11
 SOLUTION
There is no reason to pick any particular equation with which to work, so we'll
choose the first one, and in that equation we'll choose to isolate x. You'll see, as
you go along, that there are occasionally situations where it is to your advantage
to choose a particular equation to work with and in that equation a particular
variable to isolate, but typically it is irrelevant which one you choose.

Step One: Solve the first equation for x:

OETA RN:
It is imperative that you
LXE Y tS
go to the other equation in
CX ee oY
step 2! If you reuse the
x = (3 — 3y)/2
same equation you chose in The work above is the process described in step 1 of the substitution process.
step 1, you will find that you We selected an equation and a variable within it to isolate and accomplished
are led to a “dead end” in your that goal. Now we advance to step 2.
computations!
Step Two: In the other equation, 6x + 12y = 11, replace the x with the expres-
sion (3 — 3y)/2:
6[(3 — 3y)/2] + 12y=11
Step Three: Solve this equation. Note that, in the first term, the 6 reduces with
the 2, yielding
3(3 — 3y) + 12y= 11
9—9y+ l4y=11
9+ 3y= 11
Vee
2
Saha,

Step Four: Return to x = (3 — 3y)/2 and insert the value of y that we obtained
previously. This gives us
Bs Ns
Z
K=(G—2))2
]
==
2

Combining the results into an ordered pair, we find the solution is (apparently)
the point (5, +). It remains for us to check that this is correct, but it appears to
agree with the result suggested by the graph considered earlier.
Step Five: Remember that any check must be performed in the original equa-
tions we were given. Once we modified those equations, it was no longer clear
whether anything was correct because we may have made computational errors.
Therefore, we must check our answer in
ay aan
(owe ap Ail
166 Chapter 5

2
Inserting x = 1 and Lie find
2

{5 +3(2] =e sy

and

(3 - (4) =3+ co l1V

and thus we have confirmed that our solution is correct.

EXAMPLE 5.2 Solve the following system, using the method of substitution:
ANGI)
a= O
BPeweaa I

SOLUTION
As was the case in Example 5.1, choose one equation and isolate one variable within
that equation. Here, consideration of the alternatives leads us to suspect that isolat-
ing y in the first equation would be the easiest approach (do you see why?).
Step One:
AX a0
Dp ae Ona a 9.3
5 eS eh a3

Step Two: Turning to the other equation of the system, we make a substitu-
tion for y:
5x — 5(3 — 2x) = 18
Step Three: Solve the equation for x:
SX 1S. lO a8
5.15 Ss
15x= 1384-15
15x
= 33
33
x=—
15
tt
x=—
5

Step Four: Find the value of the other variable (y) by back-substitution:

i
11
Oe
 Solving Systems of Linear Equations Algebraically and with Matrices 167

wee 22
7 cer
eu
ee
Combining the values found in steps 3 and 4, we conclude that the solution to
11 if
the system is the point (=.=~2

Step Five: Check the solution. Inserting the values of x and y into the original

()o()
equations of the system, we find

44 14_
cea
30
—=6)

(2)
5
11 a
5 5
55 35
+= =18
5 5
11+7=187V
Since the solution checks in both original equations, we conclude that our an-
swer is correct. Note that we could gain further support for the correctness of
this answer by graphing the two equations of the system on the same plane,
and observing the point of intersection agrees with the solution point we found
computationally.

What would it look like, you might wonder, if a system were to have one of
the other possible outcomes, that is, if it were an inconsistent system or con-
sistent and dependent? We want to be sure that we can recognize those occur-
rences, and a few examples will help make the situation clear.

PXOAMPLE 5.3 Solve the system of equations, using the method of substitution:
PRY a WPS,
4X OV =o
SOLUTION
Again, without any particular reason, we will choose to isolate x in the first equa-
tion. Keep in mind that it generally does not matter too much which equation
we choose to work with and which letter in that equation we choose to isolate.
Step One: Solving the first equation for x, we find
2x — 3y=5
Phe —* BY nets)
x = (3y + 5)/2
168 Chapter 5

Step Two: Next, we substitute that expression in place of x in the second equa-
tion. This gives
A[(3y + 5)/2) — 6y =9
Step Three: Again, we can reduce the 4 with the 2 and obtain
2(3y + 5) — 6y =9
6y + 10 -6y =9
10=9
Notice that we have reached a rather unsettling point in our computations. The
variable y has completely disappeared from the equation, and we are left with a
false statement of equality (10 equal to 9 is just about as false as you can get!).
This is what we refer to as a contradiction, and it indicates to us that the equa-
tion we are attempting to solve has no solution. The no-solution case is also
If, in the process of solving
referred to as the inconsistent case, and hence we can conclude that the system
by substitution, you reach a
we are attempting to solve is inconsistent and has no solution.
point where your equation
Graphing the equations of the system, we find that the two lines do not in-
has become a contradiction,
tersect. Remember, the solution to a system of linear equations is equivalent to
then stop. Your system is
the point of intersection of the graphs of the equations of the system, and thus a
inconsistent and has no
system with no solutions would generate a pair of equations such as this, which
solution!
are parallel to one another.

EXAMPLE 5.4 Solve the following system of equations, using the method of substitution:
DX OV —a10
15x 24y— 930
SOLUTION
To shake things up a bit, we will choose to work with the second equation and
to isolate y in that equation. There is no particular reason for this; we have just
chosen to do something different.
Step One:
15x + 24y = 30
24y = 30 — 15x
y = (30 — 15x)/24
Note. y = (10 — 5x)/8
If, in the process of solving Step Two: Turning to the other equation, we substitute this expression for yand solve:
by substitution, you reach a
5x + 8[(10 — 5x)/8] = 10
point where your equation
has become an identity, Step Three: The 8 reduces with the 8, giving us
then stop. Your system Is
SX LOS a il)
consistent and dependent
1010
and has infinitely many
solutions! Express the
As in Example 5.2, we see the variable has subtracted out, but in this case, we
solution set using set-builder
are left with a mathematically true equality. This is what we refer to as a math-
notation. — ematical identity, and its appearance in the substitution process tells us that our
system is consistent and dependent and has infinitely many solutions. As was
 Solving Systems of Linear Equations Algebraically and with Matrices 169

done in the graphing method, we express this solution using set-builder nota-
tion and one of the original equations: {(x, y) |5x + 8y = 10}.

BX AMPUE*5(5 Solve the system of equations, using the method of substitution:

Ox aay nl
See et Oi
SOLUTION
In this case, we will make our first attempt to be shrewd in our choices. We ob-
serve that, in the second equation of the system, it would be very easy to isolate
the variable y, so we will make the choice to work with the second equation and
to solve that equation for y. We do not have to make that choice, but it certainly
appears that this would lead to the least amount of work for us.
Step One:
eae ay
Vira mn
Step Two: Turning to the other equation in the system, we replace the variable
y by this expression and solve:
Oa eA Ja OX jean
Step Three:
9x28 = 20x —s13
= ike 23 = 3s
rl Be Ais
15
==
11

This is not the most attractive solution we could imagine, but it is the one we are
stuck with, so we must live with the rather unappealing fraction. We substitute
this value into the expression we obtained when isolating y and find the following.
Step Four:

y=7-211
OT 7 eS
imei?
2
oT
Step Five: Our conclusion is that the solution to the system appears to be
Se 5 :
(S.=| It remains to check this in the original equations of the system to

verify that our work is correct:

9x+4y=13

o{ 4 =Ja
11 11
170 Chapter 5

Ci,
—=7\
1]
The solution satisfies both of the original equations of the system, and we con-
clude that we have verified that our solution is correct.

Before moving on to the next computational method of solution, we should

address the question of when it would be appropriate to choose to use this
method over the others we will see. We have already noted that all the methods
will work, but when would it be to our advantage to choose substitution?
Generally, substitution is best used either when it is easy to isolate one vari-
If it appears that one variable
able in one of the equations (as was the case in Example 5.4) or in a problem
is computationally easy
where one of the equations is given to us with a variable already isolated. In
to isolate in one of the
other situations, it generally requires a bit more work to use substitution than
equations in the system,
it would to use other methods. The amount of extra work will never be oner-
consider attempting to solve
ous, so you should not shun the method out of fear that it will be unnecessarily
by substitution!
complicated, but do be aware that in other situations, the alternative methods
we will present are likely to be more efficient.

Exercises
Solve the following systems using the method of substitution.
Lex == 7, 6. 4x+3y=13
8x+1l0y=16 5x- y=2
2 PON
Sy e0 7. 4X—3y=20
x-y=0 8x—-6y=40
a6. 2k V2 8 USES
2 =O
6x+8y=39 0.2x—-1.6y=—3.6
4, 2x-—y=10 O22 =e
x+5y=-6 4x+5y=3
>. “Seayas 10. 2x=7—-4y
2X3 y=3-2x
 Solving Systems of Linear Equations Algebraically and with Matrices 171

Tle :7”
—-x--yr=-
5 1 14. S 2x+5y=6
XG
vy=
5 3x—y+6=0
Yay wk)
3 Loo ke 2-40
2x—-5y=11 5x—-3y=10
3x+5y=4

y=3x
x-—2y+6=0

The following problems can be solved by constructing and solving a suitable system of linear
equations. Construct the appropriate system, solve using the method of substitution, and then
interpret the solution in the context of the problem.
16. A hot dog stand entrepreneur invests $8000 to open his business. The
owner knows that a hot dog and bun will cost him $1.10, and he intends
to sell his hot dogs for $1.85. How many hot dogs must he make and sell
in order to break even (obtain a profit of $0)? How many hot dogs must
he make and sell in order to amass a profit of $10,000?
17. You are offered two jobs as a computer salesperson. The first job offers
you a salary of $600 each week, plus 3% commission on sales. The sec-
ond job offers you a straight salary of 7% commission on sales. How
much would your sales have to be in order to make the straight commis-
sion salary the better offer?
18. A chemistry student wants to create a mixture that contains a 16% alco-
hol solution. She has two premixed ingredients, one of which is a 10%
alcohol solution and the other a 70% alcohol solution. If she wants to
obtain 50 liters of the 16% solution, how much of each of her ingredients
should she use?
19: Suppose the wheat supply in a particular country is approximated by
the model p = 7x — 400 (where p is in U.S. dollars), while the wheat
demand in that country is modeled by p = 510 — 3.5x. Here, p is the
price of wheat per metric ton, and x is the quantity of wheat in millions
of metric tons. Find the price at which the supply and demand are equal.
20. A widget salesman has been offered two possible compensation plans.
The first plan offers a monthly salary of $2000 plus a royalty of 10% of
the total of his monthly sales. The second plan offers a monthly salary of
$1000 plus a royalty of 20% ofthe total of his monthly sales. What must
his widget sales be in a month in order for the two compensation plans
to yield the same monthly salary?
ae A construction company needs to rent a dump truck and is comparing
two potential rental companies. Big Truck Rental offers the truck for $75
per day plus $0.40 per mile, while Tiny Truck Rental offers the same type
of truck for $105 per day plus $0.25 per mile. How far would the truck
have to be driven in order to have the costs of rental for both companies
be equal, and what would be the cost involved in such a rental?
172 Chapter 5

5.2 SOLVING SYSTEMS OF EQUATIONS USING

THE METHOD OF ELIMINATION
An alternative algebraic technique for solving systems of linear equations
is sometimes referred to as “annihilation” but is more commonly called the
method of elimination. It is particularly useful in the case where the equations
are given to you in the form Ax + By = C, but since any linear equation can
be rewritten in that form, the method applies to all systems of linear equations.
We will assume that the coefficients of the equations of the system have
integer values. If they are not integers, then for each equation having fractional
coefficients, multiplication of all terms in the particular equation by the least
common denominator of the fractions will clear the fractions to produce inte-
ger coefficients.

2 4 i ;
BAAMPLE ’5.6 Rewrite the equation a = Z y =8 with integer coefficients.

SOLUTION
The least common denominator of the fractions in the equation is 15, so we
multiply all the terms by that value and simplify:

15(2x +15{$y)= 15(8)

lOc e2y — 120

The method of elimination bears a superficial resemblance to the method

of substitution in that choices are being made, but the procedure is completely
different. For elimination, the series of steps we must follow are these:
1. Pick one variable to eliminate from the system.
2. Multiply one or both of the equations, as needed, by constants chosen in
such a way that the coefficients of the variable you chose to eliminate will
be equal but opposite in sign.
3. Add the equations of the system together, eliminating the variable you
chose in step 1.
4. Solve the resulting equation, producing one coordinate of the solution point.
5. If the solution you found in step 4 is an integer, substitute it into one of
the original equations of the system and solve to find the value of the
other variable. This is the second coordinate of the solution point. If the
solution you found in step 4 is not an integer, you may choose to repeat
the elimination process, this time opting to eliminate the other letter.
This, again, leads you to the second coordinate of the solution point.
6. Combine your results in an ordered pair and check to see that it satisfies
each of the original equations of the system. If it does not, then a compu-
tational error must exist somewhere in your work—examine your steps
to find your error.
 Solving Systems of Linear Equations Algebraically and with Matrices 173

PS ADLPLH 5.2 Solve the following system, using the method of elimination:
BS eo gentle ayae
ox toy — "8

SOLUTION
Step One: We will want to try to make our choice in step 1 shrewdly, but we
could make an arbitrary choice. The system was
Sees Bay os)
Dxat OY = 6
The objective in step 2 will be to make the coefficients of one variable in the
system be equal but opposite in sign. We will, essentially, want to change the co-
efficients of the variable into their least common multiple, with opposite signs.
We could do this with either variable, but consideration of our options reveals
that it would be slightly easier to work with y than with x.
Step Two: The least common multiple of the coefficients of y is 6, and one of
the equations already has that coefficient for its y term.
In order to match the coefficients of y (with opposite signs), we need to
multiply the first equation by 2:
2(8x — 3y = —25)
Diet OY 18
which produces
6x
OV ee)
Die) a
Step Three: If the equations are now added together, we obtain
21x = —42

Step Four: Solving the equation obtained in step 3, we find

a
Step Five: This value (being an integer) is substituted into one of the original
equations, and we solve for the other variable, y:
8(—2) — 3y = —25
16 —3y = 25
Shia)

WE
Step Six: Our solution appears to be the point (—2, 3). We confirm this by substi-
tution into the original equations of the system:
8(—2) — 3(3) = —25
16-9 = —25V
5( 2) 6(3) =8
—10+ 18 = 8V
174 Chapter 5

Since both of the equations are satisfied by the coordinates of the point (—2, 3),
we conclude that our solution is correct.

As was the case for substitution, we should be aware of how we would

recognize the inconsistent or the dependent cases. These will be determined in
step 3 when the equations are added together to annihilate the variable chosen
in step 1. The inconsistent case is identified by the appearance of a contradiction
at this stage and the dependent case by the appearance of an identity.
Remember that a contradiction would be a false equality. Both of the vari-
ables will be eliminated from the equation, and you will be confronted with an
equality of the form 0 = k, where k is some nonzero number. An identity is a
universally true equality, which in this situation will be 0 = 0. In either case, the
process will terminate after step 3 and the solution identified.

EXAMPLE
5.8 Solve the system of equations, using elimination:
2 Xe Vee 0)
4x — 10y = 15
SOLUTION
Step One: We choose to eliminate the letter x, since the least common mul-
tiple of the coefficients of the x terms is seen to be 4.
Step Two: To create a situation where the coefficients of x are equal but
opposite in sign, the first equation is multiplied by (—2):
(2)2x soy = 10)
Ax 1Oy = 15
which yields
=a Oy 20)
AX LO Vea
Step Three: When we add the equations together, note that we obtain
0=-5
This, being a contradiction, tells us that the system is inconsistent and has no
solution. .

EXAMPLE
5.9 Solve the system of equations by the elimination method:
DKS Veen
ZO Ve
SOLUTION
Step One: Again, we see that it would not be too difficult to eliminate x, since
the least common multiple of the x-coefficients is 21.
Step Two: Since we want to create a situation where the coefficients of x are
equal but opposite in sign, we multiply the first equation by (—3):
(—3)(7x — 3y = 2)
=21% + 9y' = =6
 Solving Systems of Linear Equations Algebraically and with Matrices 175

which yields
21x oy =O
ONE emesisVM)
Step Three: Adding the equations together, we obtain 0 = 0, an identity. The
conclusion is that the equations are consistent and dependent, and the system has
infinitely many solutions. The set of solutions, in set-builder notation, is {(x, y) |;
732)

ANonre
During the elimination process, if (after adding the equations together) you obtain a
contradiction of the form 0 =k (for some nonzero number k), stoo—the system has
no [Link] Is inconsistent. If (after adding the equations together) you obtain
the identity O = 0, stop—the system is consistent and dependent and has infinitely
many solutions (which you should express in set-builder notation).

Exercises
Answer each of the following, using complete sentences and proper grammar and spelling.
1. If the process of elimination should lead to an identity during the pro-
cess of solving by elimination, what conclusion can be drawn?
2. If the process of elimination should lead to a contradiction during the
process of solving by elimination, what conclusion can be drawn?
Solve the following systems of linear equations, using the method of elimination.
3. x-4y=3 oD: 2x+ y=9
3x+8y=5 3x—- y=16
4, 2x-y=l 10; c=
DV = 9
4x—-2y=2 x+3y=16

5A x= y= 3 Wale 4x-—3y=25
x+2y=5 —3x+8y=10

6. 2X=3yV =) ie: Lox

13 ye
—4x+6y=6 —6x + 6.9V==2

PGR =O 10 Ihe} 1l2x—-—3y=6

14x-6y=5 4x-y=2

S032 = D2 ye Ue 14. 2x+3y=5

0.8x+0.4y
=0 Cy 7

The following application problems can be solved through the use of a system of linear equa-
tions. Construct a system appropriate to each application example, solve the system using the
method of elimination, and interpret the solution in the context of the problem.
15. An airplane flies into a headwind and travels the 1000-mile distance be-
tween two cities in 2 hours and 10 minutes. On the return flight, the
176 Chapter 5

same distance is covered in exactly 2 hours. Find the airspeed of the

plane and the speed of the wind, assuming that both have a constant
magnitude.
16. A school is presenting a play where tickets for the event cost $5 for
adults and $2.50 for children; 548 total tickets are sold, generating rev-
enue of $2460. How many children’s tickets were sold, and how many
adult tickets were sold?
L?. How many liters of a 20% alcohol solution must be mixed with 40 liters
of a 50% alcohol solution in order to create a 30% alcohol solution?
18. Sterling silver is 92.5% pure silver. How many grams of sterling silver
must be mixed with a 90% silver alloy to obtain 500 grams of a 91%
silver alloy?
Ly. The demand and supply functions for a new type of cell phone are

p = 200 — 0.00001x
p = 80 + 0.00002x
where p is the price in dollars for the phone and x is the number of units
sold. Find the equilibrium point of the system, which is the value of p
and the value of x that will satisfy both of the equations in the system
simultaneously.
20. If a total of $18,000 is invested in a pair of bonds that pay 9% and 6%,
respectively, how should you divide your investment if you want to earn
$1410 in interest?
Al Suppose peanuts cost $2.50 per pound, and cashews cost $4.00 per
pound. If the two types of nuts are to be mixed together and sold as a
blend, how much of each type should be used to produce 25 pounds of
a mix that will cost $3.00 per pound?
22. Suppose Juan has $8.05 in nickels and quarters. If he has 29 more nickels
than he does quarters, how many of each type of coin does he have?
2s If two cities are 500 miles apart and Joan leaves city A at noon and travels
toward city B at 60 miles per hour while Sonya leaves city B at 1:00 p.m. and
heads toward city A at 50 miles per hour, when will they pass one another?
24. Two runners start from the same spot on the same quarter-mile oval
track. One runner runs at 6 miles per hour and the other at 8 miles per
hour. If they start at the same moment and run in the same direction,
when will the faster runner be one lap ahead of the slower runner?

5 : 3 SUBSTITUTIONS THAT LEAD To SYSTEMS

OF LINEAR EQUATIONS
Suppose an equation had the following form:
 Solving Systems of Linear Equations Algebraically and with Matrices 177

where h, k, and C are constants. In order to avoid a situation where the expression
becomes undefined, it’s clear that we can have neither x = 0 nor y = 0, and hence
neither the origin nor any quadrantal points can lie on the graph of the equation.
Under the assumption that neither variable can be 0, we can rewrite the
equation using the following substitutions: u = 1/x, v = 1/y. With these substi-
tutions, the equation becomes
hu +kv=C

which is a linear equation in the unknowns u and v.

Although not a common situation, it is possible that we could encounter a

system in which each equation has the form i 1 a = C, and in that event, the
AP BOL
previously mentioned substitutions would lead us to a system of linear equa-
tions that we could solve by one of our established methods. Naturally, since the
solutions to that new system will be in terms of u and v, we will have to translate
those in terms of x and y and maintain vigilance that the original expression is
not undefined for the solutions we obtain.
Let’s consider an example and see how the process would work.

EXAMPLE 5.10 Solve the following system of equations:

Seo
2
a
y's)
20
se Ae
SOLUTION
Using the substitutions u = 1/x, v = 1/y, the system becomes
OU OV ae
Mies sy AW
We are now free to choose a method with which to solve this system. Inspection
of the equations does not reveal a particular variable that would be easy to iso-
late, and consequently we steer clear of the method of substitution. Attempting
to solve by graphing might be possible, but if the point of intersection in the u,
v plane is not apparent, this could be wasted effort. Consequently, we choose to
use the method of elimination.
The coefficients of v are already opposite in sign, so the variable we choose
to eliminate in this situation will be v. The least common multiple of the v-
coefficients is 15, so the first equation will be multiplied by 3 and the second by 5:
Ops idBay VS weeks)
DOU eel Dy aN)
We add the equations together, producing
44u = 44
or
u= 1
178 Chapter 5

Since this result is an integer, we can easily find the value of v by inserting this
value of u into one of the original equations. Choosing the first equation,
3(1)
+ 5v= —2
Se OV eee 2
DV it
yo =]

Substitution of the values of u = 1 and v = —1 into each of the original equa-

tions confirms this answer to be correct.
that the coordinates of the
solution in the x, y plane are
We must now recall that our original system involved x and y rather than u
identical to. the coordinates
and v, so this cannot be the solution to our original problem. We must now recall
our substitution expressions and use them to determine the values of x and y.
found in the u, v plane. This is
Since u = 1/x, we have 1 = 1/x, or (by taking the reciprocal of each side of
not typically the case, as we
the equation) 1 = x. Similarly, since v = 1/y, we have —1 = 1/y, or —1 = y, and
will see in coming examples!
hence the solution in the x, y plane is the point (1, —1).

BXAMPLE'5.11 Solve the following system:

7
=45=5
y
8
“4-33
ny
SOLUTION
Again we make the substitutions u = 1/x and v = 1/y, which transforms the
given system into
4u ob VSS
ZUG IOV eo
In this case, there again is no clear choice for a variable to isolate, so substitution
appears to be unwise. Choosing to opt for the method of elimination and con-
sidering the coefficients of u and v, it appears that it would be easier to eliminate
the variable u from the system. We multiply the second equation by —2, which
makes the coefficients of u equal but opposite in sign:
4u+ 7v=5
—4u — l6v = —6

Adding the equations together produces an equation involving v only:

oo Viet
which we solve to find

Although this is not a terribly attractive value of y, it is (as fractions go) not
horrible, so we will substitute into one of the system’s equations and solve for u.
Arbitrarily choosing the second equation with which to work, we find
 Solving Systems of Linear Equations Algebraically and with Matrices 179

1
2u+8{<=3
°)
8
2u+—=3
-)
LBttraG a= 27
1S 319
19
u=—
18
Substitution of these values into the equations of the system indicates that
our work is correct, and now it merely remains to convert these solutions into
x, y terms.
Note At first, we may be hesitant to tackle the computations because the values
“Tysmaagi an ~~ of wand v are fractions, and this may create the concern that difficult fractions
When thesubstitutions
could arise in our computations, but this is actually never the case. Since our
involved are u = 1/x and
substitutions were u = 1/x and v = 1/y, what is actually the case is that u is the
v = l/y, the value of x is the
reciprocal of x and v the reciprocal of y. Therefore, to find the solutions in terms
reciprocal of the value of
of x and y, it is merely necessary to take the reciprocals of our u and v solu-
u, and the value of y is the 18
reciprocal of the value of v! lions hence, x = z andy = Z and our solution in the x lane is [. 9),
19 y ik XP 19
Substitution of these values into the original equations of our system confirms
the correctness of the solution.

Note that this is not the only substitution that could lead to a linear system
of equations in u and v. Consider this case:
3 8
f=!
De eae lai
7 4
=O
ame eY

In this case, the substitutions involved would be u = 1/(2 — x) andv= 1/(1 + y).
This would lead us to a system of linear equations in u and v, but when we
translate those answers into x, y coordinates, the process will not be as simple as
merely taking the reciprocal of u and v! In such a case, when the substitution is
more involved, we must insert the values ofu and v into the substitution expres-
sions and compute the values of x and y directly.
In this case, the system in terms of u and v would be
3u + 5v= 11
7u — 4v = 10

There is no obvious choice for a variable to eliminate, so we will choose to elimi-

nate v (since the coefficients of the v terms already have opposite signs). We
multiply the first equation by 4 and the second by 5, yielding
12u + 20v = 44
29H — 20 = 50
180 Chapter5

When we add the equations together, we find

47u = 94
ee
Substitution into the first equation of the system then shows us that v = 1, and
a check in both of the equations confirms the accuracy of this result.
To find the values of x and y, it is necessary to actually insert the values of u
and v into the substitution expressions and compute rather than simply take the
reciprocal of what we have already found:
cael
2-x
2(2-x)=1
4-—2x=1

—2x =-3
3
x=-

and

1
] =——_
l+y

1(1+y)=1
1+ y=1

y=0
Substitution of these values into the equations of the original system verifies
their correctness, and we conclude the solution to the original system was the
3
oint |—, 0 |.
: [

Exercises
Solve the following systems of linear equations using an appropriate substitution.

ag4 3
08 epee~ ef
x y LOX a2
Pe | 210
oS ao ica’) == ol)
x yy x y

ae
a 4. Sees,
x her. Xp ¥
oh le -
 Solving Systems of Linear Equations Algebraically and with Matrices 181

5, 4 y =12 i eater
LX. OY

ae) aay aa

8 : 4
68 Se LZ a ene
a Lee ie Gee ee)

(ee Sridiens pd eee

Ca lay Ka v6

Pie 2 13 Eee
Mu CO wd
8 2
alo Bae i
45 We cee
=5 7
8. ee |) 14. a
SSX ey Hed ey
8 5) Zt
+—- = =3 af,
ey te omen,

Ue ee
9. shire LS : Badinng 95
banda eda ve) DEE TTY.
RL 8 1 2 aye
I) aeVtaE Dae x tay.

8
LOLA UG: : faite
y Ke See Vcr | eel 0)
1
S04 == 9 apccloneerely
y = Se vel 20)

5 A INTRODUCTION TO MATRICES
The methods shown thus far as tools for the solution of systems of linear equa-
tions work well in the case where two equations in two unknowns are present.
When more variables and more equations are involved, it becomes more effi-
cient to use matrices to solve systems.
The word “matrices” is the plural of matrix, and applications of matri-
ces are both far-flung and exotic. They are used extensively in computer pro-
gramming, particularly in the rendering of images numerically (that is, in the
digital representation of images rather than preservation on film) and in the
creation of virtual reality three-dimensional simulations. It is prudent to in-
troduce matrices now, when work with them will remain relatively elementary,
so that we will have familiarity with them when the need arises to use them in
more elaborate situations.
182 Chapter 5

We should mention that we are, in this text, going to only introduce the
topic of matrices and not present an exhaustive development. A full examina-
tion of matrices and their properties is better suited to a course in linear algebra
or matrix theory, and we will leave quite a bit of the theory and examination of
matrix properties to those courses.

The Fundamentals of Matrices

The technical definition of a matrix is that it is an array of numbers. This is
not particularly helpful unless we are familiar with the notion of an array, so
we should include the definition of that term in our introduction to matri-
ces. An array is an arrangement of objects into horizontal rows and vertical
columns.
Always keep in mind that The array of numbers that make up the matrix is typically enclosed within
rows go across, and columns oversized brackets or parentheses to indicate the limitations of the array and is
go up:and down! Also, try to named with some capital letter, such as A, B, or C. You can use any letter you
accustom yourself to referring wish to name matrices as long as you remember that the common practice is to
to rows first and columns use capital letters.
second! An example of a matrix would be

3 0
A=|
1
4 5
Observe that this matrix has three rows and two columns. The numbers within
the array are referred to as the entries of the matrix and are referred to using
their row and column location (note that, once again, the rows are referenced
first and the columns second). The rows are numbered top to bottom and the
Note columns left to right. Thus, in the example, the entry 5 is said to be in row 3,
ln a matrix, the entry located column 2, and is described as the 3, 2 entry.
in row sand column, is When a matrix has been named with a capital letter, it is customary to re-
referred [Link] the 1, j entry. fer to the entries of the matrix using subscripted notation, incorporating the
Always mention the row first lowercase form of the letter used to name the matrix. That is, when we said
and the column second! the 5 was the 3, 2 entry, the typical practice is to express this as a, » = 5, and
one often sees the notation A = [a; ;] to emphasize that individual entries will
be identified in that manner.
The size, or order, of a matrix is a statement of the number of rows and col-
umns existing in the array. The notation m X n is used to specify the order of a
matrix, where m is the number of horizontal rows and n the number of vertical
columns. The example
Note
eye)
A matrix having m horizontal
A=| 1 -2
rows and n vertical columns
is Said to be of order m Xn, 4 5
or to have size Mm X n.
is then a 3 X 2 matrix, read as “three by two.”
 Solving Systems of Linear Equations Algebraically and with Matrices 183

BPAAMPLE S12 Specify the order of the matrices shown below and identify the particular entry
indicated, if it exists:
a.
(hie BDes Sie |
, entry Cc), and c;3,
4 lie O

b. 4 2
D=|" 7 =I |, entry
d,, and d,,
1/4 10

SOLUTION
a. The matrix Chas two rows and three columns, and hence its order is 2 X 3.
The 2, 2 entry is that value in the second row, second column, and here
that would be %. The 3, 1 entry does not exist, since that would be in the
third row, but matrix C does not have a third row.
b. The matrix D has three rows and two columns, and hence its order is
3 X 2. The 1, 2 entry of matrix D is the entry lying in the first row, second
column, and that would be the entry —2. The 2, 1 entry, on the other
hand, is in the second row, first column, and that would be the entry 7.

In the particular case where a matrix happens to have exactly as many rows
as it has columns, the matrix is referred to as a square matrix. The source of
this term is the square shape taken on by the array when the number of rows
and columns are equal.
Two matrices are said to be equal if they have precisely the same order
and entries in corresponding locations are equal. In technical form, we can say
A = Bifa;; = 6;; for all possible i,j entries.

EXAMPLE 5,13 Matrix A is known to be equal to matrix B. Give the values of the missing en-
tries, identifying them using proper notation:

SOLUTION
The missing entry in matrix A is a, = 4, since the 1, 2 entries of the two matrices
must agree. Similarly, the missing entries in matrix B are b, ,= 2 and b,, = 5.

There are some special matrices we should mention at this point. Any matrix
(of whatever size) having all entries being 0 is referred to as a zero matrix.
When you are interested in specifying a particular zero matrix, you might identify
it (for instance) as the “3 X 4 zero matrix,” possibly using the notation 0; ,, which
would tell the reader that you were talking about a zero matrix of that specific size.
A square matrix having all entries being 0 except for 1 in the entries in the 1, 1
position, the 2, 2 position, the 3, 3 position, and so on is referred to as an identity
 matrix of order n, where n is the number of rows in the matrix. Since the matrix is
square, this also happens to be the number of columns in the matrix. An identity
matrix is always identified as I,,,m where n is the number of rows in the matrix.

EXAMPLE 5.14 Write the matrix we would refer to as J.

SOLUTION
The matrix is the 3 X 3 identity matrix, so all its entries are 0 except for the 1, 1
entry, the 2, 2 entry, and the 3, 3 entry, all of which are 1. The matrix would be

=
oooe
©
_ ©=

The entries of a square matrix that are in the 1, 1 position, the 2, 2 position,
the 3, 3 position, and so on form what we call the main diagonal entries of the
square matrix. We will examine the main diagonal of a square matrix in greater
detail shortly. Any matrix having all entries not lying on the main diagonal
being 0, with main diagonal entries being either 0 or nonzero (of whatever
value), is called a diagonal matrix.

EXAMPLE 5.15 Give an illustration of a diagonal matrix having three rows.

SOLUTION
There are infinitely many such matrices, one of which is

GS)
any
ey SS
Tee(soyiany
f=)

Observe that the main diagonal entries need not all be 0, as the 3, 3 entry in
this matrix illustrates. An identity matrix is a particular instance of a diagonal
matrix, where all the main diagonal entries happen to be 1.

Addition and Subtraction of Matrices

When two matrices are added, which we denote by A + B, the entries in corre-
sponding locations are added to find the entry in the matrix sum. Note that this
requires that the two matrices A and B must have the same order.
Formally, we define A+ B= |a, | + [b,,,| = |4,,,+b, , |, which is the way
to express the content of the preceding paragraph in slightly more formal notation.

BXAMPEE 5:16 Add the two matrices.

oS | % = a) ee
UM BE YEE oe Pe:
 Solving Systems of Linear Equations Algebraically and with Matrices 185

SOLUTION
Both the matrices are 2 X 3, and therefore we can add them together to form
a new 2 X 3 matrix sum. All we need to do is to add the corresponding entries
and simplify the results:

3+(-6) 5+3 -1+411

LeFZ
0+4 =+—= 743
Zou

ipa)
Te
Ae ae 10
6

If we were to take a matrix A and add to it a zero matrix (having the same
order as matrix A), then the sum would simply be the original matrix A. For this
reason, a zero matrix is referred to as an additive identity matrix.
Because the addition of matrices amounts to addition of the individual
entries of the array, the familiar properties of addition of real numbers carries
over to addition of matrices. Matrix addition is commutative and associative.
That is, A + B = B + A, which means that the order of addition can be re-
versed, and also (A + B) + C= A + (B + C), which means that matrices can
be added in whatever grouping we wish.
Matrix subtraction acts, as was the case with matrix addition, entry by en-
try. In order to perform A — B, we subtract entries in corresponding locations,
noting once more that the matrices must have precisely the same order. It may
be excessive to do so, but we will admonish you to be careful about your signs
when performing subtraction!

PRAMPLE Sli Perform the matrix subtraction.

4 5 || © 2
ere) 2 i

SOLUTION
Both the matrices are 2 X 2, so subtraction is defined. We subtract entry by
entry as follows:
A=(-6) —5=2
]
rica) —-4
Y

Note that we are being very careful about our signs in the shown work:

Lee,
wf
wal argark
Z
186 Chapter 5

As was the case with subtraction of real numbers, subtraction of matrices is

neither commutative nor associative. It is always the case that we must be more
While matrix addition
careful when subtracting than when adding, and particular caution should be
possesses the commutative
applied when working with signed numbers.
and associative properties,

matrix subtraction does not! = Multiplication of a Matrix by a Scalar

When we talk about numbers occurring within the terms of polynomials, we
refer to those numbers as “coefficients.” Numbers occurring within the array of
a matrix are referred to as the “entries” of the matrix. We say that such num-
bers appear in a “context.” Numbers not appearing in any particular context are
referred to as scalars.
Given a matrix A, we may want to consider multiplication of that matrix by
a number such as 3, which we would describe as “three times matrix A,’ or 3A.
The 3, in this case, is what we would consider a scalar.
To multiply a matrix by a scalar, each entry of the matrix is multiplied by the
scalar. In a sense, the scalar “distributes” through all the entries of the matrix.

EXAMPLE 5.18 For the matrix A = i Ae i pina

SOLUTION
We multiply each of the entries of the matrix A by the scalar 6 and obtain
12, 2427-18
C0 S22 42

One can reverse the process of scalar multiplication of a matrix and factor
a common scalar factor from the entries of a matrix. There are occasions when
this proves useful in computations, as we'll see later.

Sl” 27
EXAMPLE 5.19 — Factor the largest possible scalar from matrix B=| 99 -36 | while main-
45 108

taining integer-valued entries in the matrix.

SOLUTION
The entries of the matrix have greatest common factor 9, so we will factor the 9
DSS
from each of the entries. This produces 9} 11 -4 |. Noting that the entries
oe 12
of the matrix have no other common factors, we see that we have factored the
largest possible scalar from the entries.
eee

[ ?
1X
BXAMPLE 45S220 ForA=_ Se: ir
|and B=_| —2
aoe 4 ;
, find and completely simplify 5A — 3B.
GS
 Solving Systems of Linear Equations Algebraically and with Matrices 187

SOLUTION
In this case, the matrix A will be multiplied by the scalar 5 and the matrix B by
the scalar 3, and then we will subtract. Observe that, even now, we are following
the well-known order of operations as we simplify our expressions:
Note
When performing matrix
Se aU So delpiw
operations, continue to
55 ae 36 9
follow the order of operations _| 15-(-6) -10-12 |_| 21 -22
introduced in algebra!
35-36 25-9 -1 16

Multiplication of Matrices
To this point, the operations with matrices have been somewhat intuitive in
the sense that the operations are performed in precisely the manner you would
expect them to be done. Matrix multiplication is not quite so straightforward,
though it is not particularly difficult to master.
Before introducing the method, we need to specify when the operation of
matrix multiplication is defined and the notation we will use. In general, matrix
multiplication is denoted by juxtaposition of the matrices, or their names. That
is, if one matrix is written directly alongside another, the implication is that the
matrices are to be multiplied. Alternatively, if our matrices are named A and B,
then if we write AB, we mean that we want to multiply matrix A by matrix B.
Matrix multiplication is defined in the particular case only when the num-
ber of columns of the first matrix matches precisely the number of rows in the
second matrix. Note that this is not necessarily going to happen for matrices of
precisely the same order, except in the case where the matrices are square and
of the same order!
This concept can be a bit slippery to grasp when first encountered, so let’s
illustrate. Suppose that matrix A is of order m X n and matrix B of order p X q.
How do we determine if multiplication is even possible? It all depends on the
orders of the matrices involved and their “inner dimensions.” The term “inner
dimensions’ is not universally used, but it offers an important visual reference.
When we write the matrix names alongside one another and their respective
sizes below them, as in
AB
Note (m X n)(p X q)
The matrix product AB, where the values of n and p are what we are referring to as the inner dimensions be-
Aisanm Xn matrix and B cause of their relative positioning when writing the sizes below the matrix
isap X q matrix, is possible names. The multiplication is possibly only if the inner dimensions are equal.
only ifn = p, and in that case That is, the matrix multiplication shown for AB is possible only if n = p.
the product matrix will have
If those values match precisely, the outer dimensions (the number of rows
size m X q!
in A and the number of columns in B) tell us the size of the product matrix. That
is, if it turns out that we have n = p, then the product AB will be of order m X q!
This definition presents a surprising result, unique in our experience to this
point: noncommutativity of multiplication. Consider the following example as
an illustration.
188 Chapter 5

es ee EE — ee

EXAMPLE 5.21 Suppose that A is of order 3 X 5 and B of order 5 X 3. Determine if AB is de-

fined and if BA is defined. If either or both multiplications are defined, state the
order of the product matrix.
SOLUTION
For AB, we note that (written in order) the sizes are (3 X 5)(5 X 3). The inner
dimensions match, and thus the multiplication is defined, and the outer dimen-
sions tell us that the product matrix will be of order 3 X 3.
For BA, we again note that the orders are (5 X 3)(3 X 5), and, as before, the
inner dimensions match. This time, however, the outer dimensions tell us that
the size of the product matrix is 5 x 5!
This is truly unusual for us. It is possible to calculate (by a method not yet
discussed) each of the products AB and BA, but they will not have the same sizes
as one another and therefore will not be equal! In some cases, it may even be
possible to compute the product AB, while the product BA will not be defined.
Even when both orders of
So, how does matrix multiplication proceed? When computing the i, jentry
multiplication are defined,
of the product matrix AB, we use row i of matrix A and column j of matrix B.
it is not necessarily the
The entries of row i are multiplied with the entries of column j, one at a time,
case that AB = BA! That is,
and the products added together. The sum obtained will be entry i, j of the
matrix multiplication is not
guaranteed to possess the
product matrix. That sounds mysterious, perhaps, and we could formalize it
quality of commutativity!
with what is called “double-subscript notation,” but let’s see if an example will
make the procedure clear.

a a2
pees
EXAMPLE 5.22 Suppose A=| 3 4 |and B= a |
. Find the product AB and the product
5 6

BA, if possible. If either or both is not possible, explain.

SOLUTION
Since A is a 3 X 2 matrix and B is a2 X 2 matrix, the product AB is defined
and will be a3 X 2 matrix. Let’s calculate that matrix, moving slowly and taking
care to explain the procedure cautiously. We'll call the product matrix P, so its
entries can be referred to as p;;:

iL 2 Be Pir Pap

3. 4 : Pr =Pry

mere Pa welds
Entry p),, is found by multiplying the entries of row 1 of matrix A against the entries
of column 1 of matrix B and adding those products. That would be (1)(7) + (2)(9).
Note that we move across the row of matrix A and down the column of matrix B, in
order. This simplifies to 25, and that is the 1, 1 entry of the product matrix.
Entry p, is found by multiplying the entries of row 1 of matrix A against
the column 2 entries of matrix B and adding those products. That would be
(1)(8) + (2)(10) = 28, which is the 1, 2 entry of the product matrix.
 Solving Systems of Linear Equations Algebraically and with Matrices 189

Continuing in the same manner for the other entries of the product, we
obtain the following result:

Sg)
5)(7)+(6)(9)_ (5)(8)+(6)(10)
29.) 26
= 57. 64
89 100

For the product BA, recalling that B was a 2 X 2 matrix and that A was a3 X 2
matrix, we see that the product is not defined, since the inner dimensions for
BA are not equal.

EXAMPLE 5:23 Perform the matrix multiplication shown, if possible:

eA
isk Nag 349 | 5
ae
SOLUTION
Since the first matrix is 1 < 3 and the second is 3 X 2, the multiplication is defined,
and we will obtain a product that is a 1 X 2 matrix. For each i,j entry of the prod-
uct matrix, we will use row i of the first matrix and column j of the second matrix:
HOC Nae ee Liar Lyi) (2)(2) (3 )(5) cL) eo) aloe 0)

The Multiplicative Inverse of a Square Matrix

Recalling that a square matrix is a matrix having as many rows as it has columns,
we define the multiplicative inverse of a square matrix A (often simply referred
to as the “inverse of A”) to be the square matrix A ', having the same size as
matrix A, such that AA’! = A_'A = I, where n is the number of rows/columns
of matrix A.

lie 4S
EXAMPLE 5.24 Show that the matrices A = and A‘= 2 are inverses of
2 —-5 el

one another.

SOLUTION
To show the matrices are inverses, we must compute the products (remember,
commutativity is not assured, so we must compute both AA! and A 'A) and
show that both are J.

ft ee3 ST || RS ee pel Oar

ees oe | Osi
(=0, 65-(—5) Oma! :
190 Chapter 5

Bee3 218 clio 6alone |S) -|eat) Jer

Fai 2 -5 —2+2 6+(-5) Oo ‘

Since both products are I,, we conclude that the matrices are inverses of one
another.

It turns out, as we will see in the exercises, that if a matrix has an inverse,
then that inverse is unique. Thus, it makes sense to talk about “the” inverse of
matrix A. When a matrix has an inverse (not all square matrices do!), we say
that it is invertible or that it is nonsingular. If a square matrix has no inverse,
then we say that it is noninvertible, or singular. That terminology is a bit dif-
ficult to master, since “invertible” correlates to “nonsingular, while “noninvert-
ible” correlates to “singular.” [Link] unusual, but the terminology involving the
word “singular” is inherited from a feature of matrices we will explore shortly.
The inverse of a matrix can be challenging to find. There is a shortcut technique
one can apply to 2 X 2 matrices that we will ask you to uncover in the exercises,
but in general the procedure is to follow the Gauss-Jordan reduction method. The
procedure can be tedious but is generally not difficult. It is highly susceptible to
computational error, however, and thus great care should be taken with all your
computations. Even a slight mistake will lead to a result that is tragically far from
the truth and is likely to spawn numerical values that are, to be blunt, horrific.
Conceptually, what we will do is take our matrix and “adjoin” to it an identity
matrix. In this context, by “adjoin” we mean that we will double the width of the
matrix, setting up the right half of the matrix so that it is an identity matrix of
appropriate size. Then we will perform elementary row operations on this wid-
ened matrix until the left half of the expanded matrix is the identity matrix, at
which point the right half of the matrix will be the inverse of our original matrix.
The elementary row operations one may perform on matrices are as follows:
e Interchange any two rows.
e Multiply each element of a row by a nonzero number.
e Add any constant multiple of every element of a row to the corresponding
element of another row.
There are obvious problems here that we are stalling for the moment, prime
among which is how we would determine if a given matrix had an inverse at
all! We have said that finding an inverse can be a challenging procedure, so it
is natural to ask if we should discuss methods for determining existence before
embarking on the path to locating the inverse. Perhaps so, but we'll address that
question after introducing the method.

Dig Olea!
Let's start with the matrix A=| 2 2 1 | andcompute its inverse. The
=3 inl; gel
traditional method for computation of an inverse matrix can be tricky business.
In some cases, when matrices are set up very nicely, the steps can work out with
delightful ease. Most of the time, however, a good deal of ingenuity is involved,
Solving Systems of Linear Equations Algebraically and with Matrices 191

and we do not want to explore cases that are limited to the situation where things
work out in an ideal manner (this would leave you with an unrealistic notion of
the complexity of the problem!). Consequently, we are going to take a slightly
different approach than is customary, one that will give you a systematic method
that will work no matter how hideous the inverse matrix may become.
The first step is to “adjoin” the identity matrix having the same size as A to
the right of matrix A:

<p 0 a MO ee)
Shee NRW ah ald)
aele WG alk GUE RUE

The objective now is to use the elementary row operations to transform this matrix
into a form where the left half of the array is the identity matrix. Assuming that this
can be done, the right half of the resulting matrix will be the inverse matrix we seek.
Our process, which we will not pretend is the most elegant or traditional
possible, will be a methodical approach that will not rely on clever observations
or trickery. The procedure may generate somewhat large numbers along the way,
but things will never be so bad that we cannot handle them with relative ease.

Step One: Multiply each row by a constant such that all the nonzero first
column entries are the same, with the top nonzero entry positive and the
others negative.
Observe that the first column entries are 5, 2, and —3. The least common mul-
tiple of these numbers is 30, so we will multiply the top row by 6, the middle row
by —15, and the bottom row by 10. ‘This produces

30 Win abe onal) OF 0

alle Oi =atew saber pil)
sil} Ne cake gt 0 10

Step Two: Add the first row to each of the rows having nonzero entries in
column 1; only the rows being added to are changing in this process, and the
first row is fixed:

30 WP MP 2's) Fal)
WES Seer Tey ahed Ae)
Oe 1) eG 0 10

Observe that the entries in the matrix are growing in size. If we wish, we can
multiply each row by a suitable entry at any stage in order to remove common
factors and thereby make the entries as small as possible as we proceed. This is
not necessary, but you may wish to do so. In this case, we can multiply the top
row by 1/6, the middle row by 1/3, and the third row by 1/2. This produces

Soe Omens eae een ()) ()

OL) eee eee)
0 oho) Ae Aes
192 Chapter 5

Step Three: Now we focus on the second column. We want the nonzero en-
tries to be the equal, with the middle entry this time being positive and the
other nonzero entries negative.
Since the 1, 2 entry is already 0, we ignore that entry and consider the other
rows. The least common multiple of the existing entries is 10, so we multiply the
second row by —1 and the bottom row by —2. This produces

5 DE ro 0
OFT Oma eS tO
ON 1082-08 Ome 10

Step Four: Add the second row to each of the rows having a nonzero entry
in column 2:

Sat) eee letbona) 0

Opel Ofor are= 255 0
OR Ore See 0

Note that the matrix is gradually assuming the form we desire. None of the rows
have a common factor for their entries, so we are forced to leave the entries as
they are for the moment.
Step Five: We now wish the nonzero third column entries to be equal, with
the third row entry positive and the other entries negative.
Here, the least common multiple of 2, 1, and 1 is 2, so we multiply the first row
by —1, the second by —2, and the third by —2 to yield

5 0-2-1 0 0
0 -20 -2 4 -10 0
One getagl Oee Ose

Step Six: Add the third row to the rows having nonzero entries in the third column:
2) OB OS ROMO
OR 20 Oe OR 0m)
0 Oe e ol Ogee)

Step Seven: We are nearly to the end. All that remains is to multiply each
row by the reciprocal of the first nonzero entry in that row, and this will
complete the process.
Note that this could create fractions, but we have stalled off that possibility as
long as possible. The first row must be multiplied by = , the second by a , and
a) 20
1
the third by (5 This yields
 Solving Systems of Linear Equations Algebraically and with Matrices 193

ID = tam eA
We conclude the inverse is A’ =| -1 1 —-1 |, which we can confirm by
oe EB iE)

computing the product (in both orders) of A with its alleged inverse. We will
suppress that work here and ask you to trust that we have checked its correctness!

PXAMPLG 5:25 Find the inverse of the matrix C =| i ( }given that the inverse truly does

exist.

SOLUTION
As our first step, we adjoin the identity matrix to the matrix C:

Aud: say, la
oO a0 MET

Next, we arrange things so that the first column entries are equal, with the
top entry in the column being positive and the lower entry negative. The
existing entries are 4 and —2, and they possess common multiple 4, so we
merely need to multiply the entire second row by 2 in order to achieve our
objective:

Aree gla)
Ae O peer?

The top row is now added to the bottom row, producing

As eet. e0
Ones ov 2

We now turn our attention to the second column. The existing entries are 7 and
13, which have common multiple 91. We wish the second row to have a posi-
tive 2, 2 entry and the top row to have a negative 1, 2 entry, so the top row is
multiplied by —13 and the bottom by 7. This does, as we forewarned, produce
rather large numbers, but they are integer valued and are not beyond our ability
to handle:

=D.26 9 Lo eelSs al)

Oa 1 7 14

We add the second row to the first row, which gives us a diagonal matrix in the
left half of our expanded array:

2 0 Oe nL
Ov-ol) 7 14

In order to complete the computation of the inverse matrix, we now multiply

each row by the reciprocal of its first nonzero entry. The first row is multiplied
194 Chapter 5

by — fa and the second row by ay This produces rather frightening fractions,

52
but, as we never have to perform calculations with them, it is of little concern:

LP ORG/ S214 52
Oia Ole a4 91

Reducing our entries in the right half of the array, we find

SOMES 26.27) 26
DP ples ey Ale,

Hence, our inverse matrix is

ci |-3126 7126
Yi SoM)

A straightforward calculation reveals that both CC ' and C”'C simplify to L,,
confirming the correctness of our work.

Testing for Invertibility

As we have seen, finding the inverse of a matrix can be a lengthy procedure. We
had indicated prior to presenting the method for computation of the inverse
that it would be desirable to have a means for determining the existence of an
inverse prior to undertaking the process of calculating the matrix, and it is to
that issue we now turn.
Of course, there is already a means at hand for us to determine that an in-
verse would fail to exist, though we have not spoken of it. It could well be that
the process we were attempting in the examples would be impossible, and that
would demonstrate the nonexistence of an inverse.

3 4
EXAMPLE 5.26 Show that the matrix A = |does not have an inverse.
WA WE

SOLUTION
Proceeding as before, we attempt to calculate the inverse:

3 Ay e0
12a tomsOre

The least common multiple of the first column entries is 12, so we multiply the
first row by 4 and the second by —1, yielding

pO ee
Salto NAP Sil
 Solving Systems of Linear Equations Algebraically and with Matrices 195

Now we add the first row to the second, which gives the following rather un-
usual result:

12 1G speatiae 0
WE Ui ae |

Observe that the first half of the second row has all 0 entries. It will be impossible
to generate an entry of 1 in the 2, 2 entry, which is needed for us to transform
the left half of the matrix into the identity matrix. Consequently, computation
of the inverse is, in fact, impossible.

The previous example illustrates that it is possible to determine directly the

failure for an inverse to exist. We can, if we wish, always proceed as in that ex-
ample, knowing that (if our matrix is not invertible) we will ultimately reach a
stage where the left half of one of the rows in our augmented matrix is all 0s and
that this establishes noninvertibility.
There is an alternative to this approach that uses the determinant of
the matrix. We will introduce computation of determinants using cofactor
expansion across the first row of the matrix. We point out here that this is, in
practice, not the most efficient manner of calculating a determinant, but the
other methods rely on further exploration of matrices, and this would take us
beyond the scope of our interest here. Parenthetically, we point out that in ap-
plications it is not uncommon to use matrices whose size exceeds 25 X 25 and
that cofactor expansion of a square matrix with n rows typically requires at least
n! multiplications. For a 25 X 25 matrix, this would call for 25! computations,
or approximately 1.5 X 10°°. If you had a computer capable of performing 1 bil-
lion multiplications per second, it would require more than 500,000,000 years
to complete the determinant calculation! In our work, we will not see matrices
this large, however.
For a2 X 2 matrix, A = [a;,], the determinant, which we symbolize as det
A, or |A|, is defined by the formula a,,;4), — a, a. This can be thought of as
the difference of the products of the entries lying on the crossing diagonals of
the matrix.

PNP LE >. 27 Find the determinant of the matrix A = z '}

SOLUTION
We compute the determinant by taking the product of the main diagonal en-
tries 3 and 2 and subtracting the product of the entries on the other diagonal,
—7 and 5. That is, det A = (3)(2) — (—7)(5) = 6 — (—35) = 6 + 35 = 41.

Before proceeding, we should mention that the value of the determinant,

among other things, provides us with the answer to our question: does A
have an inverse? The answer is yes if the determinant ofA is nonzero and
no if the determinant ofA is 0. Consequently, the matrix A in the example
is invertible.
196 Chapter 5

(RULE: The matrix A has an inverse if and only if its determinant is nonzero.

We will not prove this result in this text, but it is a fundamental result estab-
lished in a course on linear algebra.
For matrices larger than 2 X 2, we use a recursive definition, where each _
larger square matrix determinant is defined on the basis of the determinant of
the next-smaller square matrix.
For n = 2, the determinant of an n X n matrix A = [a; |] is computed ac-
cording to the following rule:

det A= Rely iy [det A, |

ja

where A, is the (n — 1) X (n — 1) matrix obtained from matrix A by deleting

row | and column j. In the formula, note that a,, is the entry in row 1, column
j, of the original matrix A. The formula looks intimidating to use but in practice
is not particularly difficult, as is shown in the next example.

eyees — 0
EXAMPLE 5:28 Compute the determinant of matrix A=| 3 -4 1
Pitges AVN 78

SOLUTION 0
We will be working across the first row and will walk you through the compu-
tations cautiously. You will note that the factor in the summation (—1)'* will
cause an alternation in sign from term to term, beginning with a positive factor:

ia A= Cet |aoe Jr
05
We pause here to remark on where that matrix “came from.’ The value of j in the
first term of the summation is 1, and this accounts for the form of the exponent
on (—1). Also, the 2 is the 1, 1 entry of matrix A. The 2 X 2 matrix can be found
if we examine matrix A and think of what would remain if the first row and first
column of A were deleted. That is precisely the 2 X 2 matrix shown.
Continuing, we find

det 4= ("ie si : [ecu eael ; :|

+ (-1) 143 (0) 3ses-4

Notice that each of these computations relies on the ability to compute the
determinant of a2 X 2 matrix, but that has already been established. This
is what is meant by a “recursive” definition: computation of determinants
of a particular size relies on the determinants of matrices of the “next-
smaller size.”
 Solving Systems of Linear Equations Algebraically and with Matrices 197

Consequently, we simplify to obtain

2 4 1 mee om
det
A = (-1) (el nae Je 1) (5) 5 Z|

4 oa 4
+ (—1) (0) Sanh |

=(D2)e 2050 ts Gay(S 15 = 2] 210) 10 (8)

SU Oe U
= —105
The determinant of A has been shown to be nonzero, and hence we can con-
clude (though this was not part of the question at hand!) that A is invertible.
Notice that the computation of the determinant was made simpler by the
appearance of Os within the matrix. This fact underlies the rationale for other,
faster techniques for computation of determinants that are presented in texts
on linear algebra.

Exercises
For the following questions, answer in complete sentences using proper grammar and spelling.
1. What is meant by the order of a matrix, and how is it determined?
Explain the method for determining when two matrices can be multiplied.
How does one determine if two matrices are equal?
What is meant by the term “diagonal matrix”?
When two matrices have the same order, under what conditions can the
oe
matrices be added, subtracted, and multiplied?
Perform the following matrix operations or explain why the operations are impossible.

ell oie 5 ho 11 3
ji KE Spee & el 2 |+ 1
Bail ee lg: _4
paler ‘| me fo a epee |

an |) eee S) +3; 4 -5 =
198 Chapter 5

In the following exercises, let the matrices A through Fhe given as follows. Perform the
indicated operations or explain why the operations cannot be performed.

es 5 y) 4 1 —2 0

a-| Gas: bea io pea 5 iec I3 igneieeann

59) cee Cs

p One 65

15. B—C IERIE

16. B+E Vee
17kED 155°C 7B
18. DE 26. B-C
19. FE i Lite
20- AD 3 ds
Sy MTe 28 AD
22. A? 2G ON

For the following matrices, find their determinants, if possible. If it is not possible to find their
determinants, explain why.

30. 3 4 34. a lo
af 2: Or 0s a4
ap al al

op 2 ako) 35 De A
oF | oa!
Vj aS

32) e—2, | 36. on 6

Ae ely eS ee Fa
Rp he 5)

ee) 4 aif james

=2 22S
5 > lag

For the following pairs of matrices A and B, compute the products AB and BA and use the
results to determine if the matrices are inverses of one another.

Be TS (oO ae i
—5 4 ia)
 Solving Systems of Linear Equations Algebraically and with Matrices 199

Hechler ysleperayile
=| 4.1 (Nie ane

iw ves Se WARE —lfA =/ 116

4 4 —l/4 =3/16

oe eel iy oh «ah
42, pA |= 2. ah B= PE Noe
on Sih qal stb hdl

Seo wee ey Sp SIN)

AD Ai Ayo ely19 Oy LD eee Db— | eel A)
J eee AOS AL Lo 1, SE aL

For the following matrices, find the inverse matrix (if it exists).

44. oS) 49. Leo

Salle) 2 Sy
oe Al

45. 2 ae 50. i) =2/ 53

m4 One
Awl gar

46. It v8 ol Da eee
Salk fe) ey MNO»:
§~- ned

47. 23 Sy45 os) ee ee L

eyo) 7 256
Bh SN) aaa!

48. Aol) 0 Dot IN ale oh

sal pee 5 |
ie wee, a

Usinc MatTrRICES TO SOLVE SYSTEMS

fo) ae MIN TV; 4 ele) WEL@) IS
With the basics of matrices in our hands and well under control, it is time to see
how they can be used to solve systems of linear equations. Because the known
methods of substitution and elimination work well and efficiently for systems of
two linear equations, we typically apply the matrix method only to systems hav-
ing at least three variables, but the method works for two-equation systems as
200 Chapter 5

well. Equations involving three or more variables are considered linear in the case
where each of the variables is raised to the power one only and no term contains
more than a single variable.
An example of such a system would be
x+y +z=10
2x—-3y+z=4
3x-—y+4z=0

Systems with more than two variables also have “points” as their solution,
but they are no longer ordered pairs. The points will have one coordinate
for each variable occurring within the system, and thus the shown system
would have an ordered triple as its solution point. When three coordinates
are involved, the standard ordering of coordinates is alphabetical, as (x, y, Z),
but if more than three variables are present in a system, we typically dis-
pense with distinct letters and use subscripted variables such as x), x3, x3, X4,
and so on. This indicates to us the order of occurrence of the variables as
coordinates within the point.
It is not uncommon to find systems of extremely large size occurring within
applications. When aircraft construction companies create computer models to
examine airflow around experimental aircraft, the systems can have as many as
2 million equations and variables! Such systems are, of course, impractical to
solve by hand and require the use of some of the world’s fastest supercomputers
for their solution.
When we attempt to use matrices to solve systems, we express the left side
of the equation as a matrix product and the right side as a one-column matrix.
The result will be a matrix equation of the form AX = B.
If we view the equation in this way, it is natural to suppose that the equation
can be solved by “dividing” each side of the equation by the matrix A, but we
have never seen an exhibition of matrix division (for very good reason—such
a process is undefined!). However, if we were to multiply each side of the equa-
tion on the left by A’, note that we can solve the equation for X:
A 'AX=A™'B
X=A™'B

Naturally, this process relies on the existence of A ', but if the inverse does not
exist, then the system of linear equations does not have a unique solution. Let’s
see how the method would work by examining the system once more and pro-
ducing the matrices involved.
Recall that we were considering

x+y+z=10
2x-3y+z=4
3x—y+4z=0

To express this as a matrix equation, we will use two matrices on the left: a
coefficient matrix (whose entries are the coefficients of the equations on the left
side of the equations) and a variable matrix (a single-column matrix whose entries
 Solving Systems of Linear Equations Algebraically and with Matrices 201

are the variables of the system). The right side of the matrix equation will be a sin-
gle-column matrix whose entries are the constants to the right of the equals signs.
il it mall x 10
THUS ein ethisecase, Are | Ques Wale oll yeulseand |B med
joes . 0
With these matrices, the system of equations is equivalent to the matrix equa-
tion AX = B, which is easily seen if one multiplies the matrices A and X, and we
recall the definition of equality of matrices.
The determinant of matrix A is —9 (see Exercise 37 in Section 5.4), and
hence A is invertible, and the system has a solution. The inverse of matrix A is
11/9 5/9 =4/9
A'=| 5/9 -1/9 -1/9 | (see Exercise 53 in Section 5.4), and hence we
-7/9 -4/9 5/9
can find the solution to the system as

11 eS 4) 10
X=| 5/9 -1/9 -1/9 t
7)|) eet) eee 0

130/9
X=| 46/9
—86/9
130 46 86
The solution to the system is the ordered, triple S oi = :

EXAMPLE 5.29 Solve the system of linear equations x +3y+4z=-3, using the inverse matrix
KEL 322
x+4y+3z=-6
method.
SOLUTION
The system of equations can be expressed as the matrix product AX = B, where
ly oe x =3
wehavethe matrices A=} 1 2 3 |,X=] y |, andB=/ —2 |.Astraight-
by 4a z 6

1
forward computation reveals that the determinant of the matrix A is (5), and

therefore the inverse matrix exists, and the system has a unique solution.
Using our technique of Section 5.4, we find the inverse matrix to be
Shy SA BiB
A'=| 0 -1/2 1/2 |,and thus the solution to the system can be found
ee Ld
 by computation of A 'B. The computation yields the product matrix | —2 |,

and hence the solution to the system of equations is the ordered triple (—1, —2, 1).
We can confirm this result by substitution into the equations of the origi-
nal system, demonstrating the coordinates satisfy the original system equations
simultaneously: |
(—1) #°3(—2) + 40) = -1-6+4=-3V
(—1) + 2(-2) + 30) = -1-4+3=-2V
(—1) + 4(—2) + 30) = -1-8+3=-6V

EXAMPLE 5.30 Solve the system of linear equations 2x+4y+10z=3 using the inverse matrix
4x+6y+16z=10
—2x+2y+4z=5
method.
SOLUTION
2. A. 1Q
The coefficient matrix of the systemis A=} 4 6 16 |, whichhas determi-
eet 4

nant —8, and hence is invertible. The inverse matrix is found using our method
1 =i 2, he
of Section 5.4 and is equal to the matrix A‘ = 6. 7/2 = 1
oY EE IN
The solution to the system is found by computation of the matrix product

1 —-1/2 -1/2 3 —9/2

NASR oy OG". je2Foo) eet 10 |=| -22 |. As in the previous
—5/2 3/2 iby 62 SD 10
example, this indicates that the solution to the system of equations is the
ordered triple (=. — 22, 10

Exercises
Solve each of the following systems of linear equations, using the matrix method.
"4x = Sy=1 3. —8x+6y=6
3x+y=9 4x-6y=—5

2. 8x—-l4y=—46 4. 6x-5Sy=1
2x+5y=-3 8x+3y=11
 Solving Systems of Linear Equations Algebraically and with Matrices 203

Se 2K Oy 5 10. 5x=6y—42 =—47

3x+2y=—-4 7xX+8y—2z=27
3x+y+4z=14

65 ez X ey.
= 2 Live oe 2 Zi 1
6x+4y=22 5x+3y=13
NEV ae AD

T= oly = Pape Oaae eo)

Ska Oy tog
pee ae 2x-y+3z=3

I
On i aaVa LSS Xer7 VEZ 3)
1 1 3X +2y—4Z =25
Sulit 2x+4x-z=18
9 2e>oy
2a lh 14. x-2y+z=-8
SRY PS = 12 BYBe eae Pn)
Vi OE 2x—4y+2z=—16

Summary
In this chapter, you learned about:
© Solving systems of linear equations by the substitution method and the
method of elimination.
“ Conversions into linear equations by substitution.
“> Matrices, their attributes, and their operations; equality, addition, sub-
traction, multiplication, and multiplicative inverse; and solving systems
of linear equations.

Glossary
additive identity matrix A matrix all of whose entries commutative An operation whose order can be
are Os. Also called a zero matrix. reversed without changing its outcome.
associative A reorganization of mathematical determinant of a matrix A numerical value
operations using symbols of grouping, which does corresponding to a particular square matrix, used to
not change its outcome. determine invertibility.
coefficient matrix A matrix created using the coefh- diagonal matrix A square matrix whose entries not on
cients of the variables of a linear system of equations. the main diagonal are all 0.
cofactor expansion A technique for calculating the elementary row operations Adding a multiple of any
determinant of a matrix. row of a matrix to another, multiplication of any row of
column of a matrix Any of the vertical alignments of a matrix by a nonzero scalar, and interchanging of rows
entries of a matrix. of a matrix.
204 Chapter 5

elimination One of the techniques of solving systems nonsingular matrix An invertible matrix.
of equations.
order The size of a matrix, given as m X n, where m
entries The numbers within a matrix. is the number of rows and n the number of columns
within the matrix.
equal matrices Two matrices having the same order
and identical entries in corresponding locations. ordered triple A three-coordinate expression identify-
ing a location in three-dimensional real space.
identity A true statement of equality.
row of a matrix Any of the horizontal alignments of
identity matrix A square matrix having 1s in all main
entries within a matrix.
diagonal entries and 0s elsewhere.
scalar A number not appearing within any particular
inconsistent system A system of equations having no
mathematical context.
solution.
singular matrix A square matrix having no inverse.
invertible matrix A square matrix for which a
multiplicative inverse exists. square matrix A matrix having as many rows as
columns.
main diagonal In a square matrix, those entries
occurring in row k, column k. substitution One of the techniques of solving systems
of equations.
matrix An array of real numbers.
variable matrix A column matrix whose entries are the
multiplicative inverse of a square matrix A A second
variables of a linear system of equations.
square matrix, named A‘, having the same size as
matrix A, such that AA ' = AA =I. zero matrix A matrix all of whose entries are 0s.
Also called an additive identity matrix.
noninvertible matrix A matrix for which no
multiplicative inverse exists.

End-of-Chapter Problems
Solve the following systems using the method of substitution.
Ll. 4=y=3 Set = 9
8x+y=7 2x+5y=14

2: ox tye 4. 2x-3y=7
7X+7y=7 3x+4y=11

The following problems can be solved by constructing and solving a suitable system of linear
equations. Construct the appropriate system, solve using the method of substitution, and then
interpret the solution in the context of the problem.
5. Toys of type A cost $0.20, and toys of type B cost $0.30. You want to
spend $7.00 and buy 40 toys. How many toys of type A do you need to
buy and how many of type B? Does your answer make sense?
6. The demand for computers on a college campus is given by the equation
2342 — 0.01p, where p is the price in dollars. The supply is given by the
equation 253 + 0.01p. At what price are the supply and demand equal?
 Solving Systems of Linear Equations Algebraically and with Matrices 205

re One train leaves a point 37 miles east of Chicago, heading east at

60 miles per hour. At the same time, another train leaves a point 3 miles
east of Chicago, heading east at 90 miles per hour. When will the second
train overtake the first, and how far east of Chicago will they be?
George has $1.05 in pennies and nickels. He has 15 more pennies than
nickels. How many of each kind of coin does he have?
You want to eat 250 grams of food, and consume a total of 1500 calories.
Carbohydrates run 4 calories/gram, while fat runs 9 calories/gram.
How many grams of carbohydrates should you eat and how many grams
of fat?

Solve the following systems of linear equations using the method of elimination.
3
12? Pre ee
10. 3x-—4y=7
9x —3y =37 4x—-—5y=13

x—y=5 13 : 264x722
Xe =
e's
14 be,
ok Ly — 7 3x+3y=2

The following application problems can be solved through the use of a system of linear equa-
tions. Construct a system appropriate to each application example, solve the system using the
method of elimination, and interpret the solution in the context of the problem.
14. The County Fair admission fee is $8 for adults and $6 for seniors; 5432
people were admitted, and $38,024 was collected. How many adults and
how many seniors were admitted?
15. Suppose you mix x liters of one fluid with y liters of another and end up
with 30 liters of fluid weighing 40 kg. Furthermore, suppose that the first
fluid weighs 1 kg/liter and that the second fluid weighs 1.5 kg/liter. What
are the values of x and y?
16. You can buy shingles at store A for $175 per square, but the store charges
a flat rate of $25 for delivery. At store B, shingles are only $160 per square,
but the charge for delivery is $85. What is the break-even point, where
the total cost is the same from the two stores?

14-karat gold is we gold, while 24-karat gold is pure gold. How many
24
grams of 14-karat gold must be mixed with 24-karat gold to obtain 20
grams of 18-karat (= |gold?

. A cell phone company has two different plans. In the first plan, you pay
$10 per month plus $0.25/minute of calling time. In the second plan,
you pay $20 per month but only $0.10/minute of calling time. How
many minutes do you need to talk per month to make the two plans cost
the same?
206 Chapter 5

Solve the following systems of linear equations using an appropriate substitution.

19. -s4=7 te te
potas 12 Seana te

DO aces—1 22. sy+5h=7

er ———=

Perform the following matrix operations or explain why the operations are impossible.

dntents©)
alate care 7

25 See al Om OnE AO,

eye ee iam OL” NOL)
ore) S ae uN nn \O

26. ] co ww
Zee
Zee
oe Oe
a 3 7

In the following exercises, let the matrices A through D be given as follows. Perform the
indicated operations or explain why the operations cannot be performed.

LeS SunG i x4i 75 cee

A= the 6:
, B= 5.9
,C= 2 —3 4
,D=| 2 0
Sear

27. A+B BL De
28. A+C 3 2maeA-
29. A-B 33. 2A —3D
30. AC 34. 3B—A
For the following matrices, find their determinants, if possible. If it is not possible to find their
determinants, explain why.

aD: 3) 168 37% ot ae

Ae) Gans

36. Leer 38. 258

ee ae! Sak
ao an, 5
 Solving Systems of Linear Equations Algebraically and with Matrices 207

For the following pairs of matrices A and B, compute the products AB and BA and use the re-
sults to determine if the matrices are inverses of one another.

39, a-| mae p-| / |

oA kl phe Oil

eA aA i 24
Ale A=|0 35 Ie SB =e ee ee al
Seeley, c/n lane

eal) Lie
LO NN EE Sh ye ey,
a eI Soe

For the following matrices, find the inverse matrix (if it exists).

43. | dae) 45. 2d

Zee a iy

44, A eeDae 46. eer es.

Duele eZ a6 2
Sa I) oBey 5

Solve each of the following systems of linear equations using the matrix method.
47. 2x+3y=7 49, 5x+3y+4z=12
x 3x+7y—2z=14
selma
i ace)
2 7x—4y+3z=35
48. 5x+7y=35 50. —2x+3y—4z=17
xty=2 SX 2y 432 19
x-4y+2z=21