HULL CO) LEGE ,

vEDEsshtl ¥

ee poe Se

[1S.a% FTISOBINS 664 >

ga | Pi ate mo

Bar . os} ity | .

CODE a a 1 ¢ |

‘ a
“™ > - ¢ vi a : fr r

yo %
en‘ owe Wick 3m= Ly
he 7 a a : oe , 7 - oe
7 © ; ie a a
 V0 owe nn

Linear
Algebra
K. A. Stroud
Formerly Principal Lecturer
Department of Mathematics
Coventry University

Dexter J. Booth
Formerly Principal Lecturer
School of Computing and Engineering
University of Huddersfield

INDUSTRIAL PRESS, INC.

a VITIIUUMIN
x (WWNRAAAOA
 Library of Congress Cataloging-in-Publication Data

Booth, Dexter J. r
Linear algebra / Dexter Booth, K.A. Stroud.
p. cm.
Includes index.
ISBN 978-0-8311-3188-3 (softcover)
1. Algebras, Linear. 2. Matrices. I. Stroud, K. A. II. Title.
QA184.2.B66 2008
512'.5--de22 2008022573

Published in North America under license from Palgrave Publishers Ltd,

Houndmills, Basingstoke, Hants RG21 6XS, United Kingdom

. Industrial Press, Inc.

989 Avenue of the Americas, 19th Floor
New York, NY 10018

Sponsoring Editor: John Carleo

Copyeditor: Robert Weinstein
Cover Design: Janet Romano

Copyright © 2008 by Industrial Press Inc., New York.

Printed in the United States of America. All rights reserved.
This book, or any parts thereof, may not be reproduced,
stored in a retrieval system, or transmitted in any form without
the permission of the publisher.

10" 9) 28" Ae ORs een!

Contents

Hints on using this book Vii

Useful background information ix
Preface xiii

1
Linear equations 2
Solution of simple equations 2
Simultaneous linear equations with two unknowns 5
Solution by substitution 5
Solution by equating coefficients 6
Simultaneous linear equations with three unknowns 8
Pre-simplification 10
Existence of solutions 14
Non-existence of solutions 17
Simultaneous equations in three variables 19
Summary 21
Review exercise 21
Can You? Checklist 1 24
Test exercise 1 24
Further problems 1 25

Learning outcomes 28
Determinants 29
Determinants of the third order 36
Evaluation of a third-order determinant 37
Simultaneous equations in three unknowns 40
Review exercise 43
Consistency of a set of equations 49
Properties of determinants 53
Can You? Checklist 2 58
Test exercise 2 58
Further problems 2 59

Learning outcomes 62
Matrices - definitions 63
 Contents

Matrix notation 65
Equal matrices 65
Addition and subtraction of matrices 66
Multiplication of matrices 66
Scalar multiplication 66
Multiplication of two matrices 67
Transpose of a matrix 70
Special matrices del
Determinant of a square matrix 1)
Cofactors 74
Adjoint of a square matrix 75
Inverse of a square matrix 76
Product of a square matrix and its inverse joc)
Solution of a set of linear equations /9
Gaussian elimination method for solving a set of linear equations 82
Summary 85
Can You? Checklist 3 87
Test exercise 3 88
Further problems 3 89

Program 4 Matrix algebra

Learning outcomes 91
Singular and non-singular matrices 92
Rank of a matrix 93
Elementary operations and equivalent matrices 94
Consistency of a set of equations 98
Uniqueness of solutions 99
Solution of sets of equations 103
Inverse method 103
Row transformation method 107
Gaussian elimination method digg!
Triangular decomposition method 114
Comparison of methods 120
Eigenvalues and eigenvectors 120
Cayley-Hamilton theorem 127
Systems of first-order ordinary differential equations 128
Diagonalization of a matrix 133
Modal matrix teetes
Spectral matrix i
oe
Systems of second-order differential equations 138
Matrix transformation 145
Rotation of axes 147
Summary 149
Can You? Checklist 4 152
Test exercise 4 153
Further problems 4 154
@
Contents

Cie were 157

Learning outcomes 157
Introduction: scalar and vector quantities 158
Vector representation 159
Two equal vectors 159
Types of vectors 160
Addition of vectors 160
Sum of a number of vectors 161
Components of a given vector 163
Components of a vector in terms of unit vectors 167
Vectors in space 169
Direction cosines 171
Summary 172
Can You? Checklist 5 172
Test exercise 5 173
Further problems 5 173

Learning outcomes 174

Introduction 175
Addition of two-dimensional geometric vectors 175
Commutativity 176
Associativity 176
Identity element for addition 177
Scalar multiplication of two-dimensional geometric vectors 179
Properties 179
a(a+b) = aa+ab 180
Proof that a(a+b) = aa+ab 181
From the specific to the general 181
Identity element for scalar multiplication 183
Mathematical structures 183
Vector space 184
Matrices 185
Scalar multiplication 185
Notation 186
Axioms for the addition of matrices 187
The identity element for addition 187
Scalar multiplication 188
Polynomials 191
Axioms governing addition 192
The identity element 193
Scalar multiplication 194
Subspaces 196
Ordered n-tuples of real numbers 196
Review exercises 197
Linear combinations and spanning sets 200
Linear dependence and independence 204
 Contents

Test for linear independence 208

Basis vectors and dimension 210
Minimal spanning sets 213
Solution spaces 218
Summary 221
Can You? Checklist 6 223
Test exercise 6 224
Further problems 6 224

Learning outcomes et
Functions 228
One-to-one correspondence ~ 228
Operators and transformations 230
Linear transformations 231
Some specific linear transformations 232
Bijective linear transformations 233
Invertible linear transformations 236
Linear transformations and bases 238
Invertible linear transformations and inverse matrices 242
Some specific linear transformations 244
The identity transformation 244
The zero transformation 245
Reflection in the x-axis 245
Reflection in the y-axis 246
Counterclockwise rotation about the origin through 6 247
Summary 248
Can You? Checklist 7 249
Test exercise 7 250
Further problems 7 250

Answers Zon
Index 257
Hints on using
this book

This book contains lessons called Programs. Each Program has been written in
such a way as to make learning more effective and more interesting. It is like
having a personal tutor because you proceed at your own rate of learning and
any difficulties you may have are cleared before you have the chance to
practise incorrect ideas or techniques.
You will find that each Program is divided into numbered sections called
frames. When you start a Program, begin at Frame 1. Read each frame carefully
and carry out any instructions or exercise that you are asked to do. In almost
every frame, you are required to make a response of some kind, testing your
understanding of the information in the frame, and you can immediately
compare your answer with the correct answer given in the next frame. To
obtain the greatest benefit, you are strongly advised to cover up the following
frame until you have made your response. When a series of dots occurs, you
are expected to supply the missing word, phrase, number or mathematical
expression. At every stage you will be guided along the right path. There is no
need to hurry: read the frames carefully and follow the directions exactly. In
this way, you must learn.
Each Program opens with a list of Learning outcomes which specify exactly
what you will learn by studying the contents of the Program. The Program
ends with a matching checklist of Can You? questions that enables you to rate
your success in having achieved the Learning outcomes. If you feel
sufficiently confident then tackle the short Test exercise which follows. This
is set directly on what you have learned in the Program: the questions are
straightforward and contain no tricks. To provide you with the necessary
practice, a set of Further problems is also included: do as many of these
problems as you can. Remember, that in mathematics, as in many other
situations, practice makes perfect — or more nearly so.

vii
 —

6H haseail iy. 4/4 ce, esha - 4 7v ey _ ™ od ae

=f Perr? ¥ ipa} 1 hi tn
ew, paves b¢ =) Our oon) bebivii, 2: meigorl-coss fec8 On
; If DIEM | JNGTDS ‘7d 4 Nei dov ten ‘a
uhingees onto!isin Saal af SEGAL
E Pee, Segoe MOY Wek set yran To Sut UrArd Yi aa? THB? |
eet YII9 ogg
t we," 3 Dee | ‘2gadcitt OF Lenln psi ote Morr
ane ote TG aide dea beware 242
144
Prey i? : oii) LDA m6 WY SOD
a7
24 ote sag Lay silonsd teaidotg a2t7 MiB 9 yy
edd patton qweoy sbem.sved wor Hii orn) sa¢
; vineo yn Sc¥ yigqueot IsiaaQey sik», «
x ‘ MW ay inSeria wdHw way spss LO 3 nalae21g A 5
eam GR AE NE LD bask deutad'oh boinc
OF ARE ede
rim e n o a m e r g u r ’ at toed
le 39
rastyo
ee
gm
ee
an inte
io
ee ee oad: nese ope vont
eeaaue” peered
cpanel, OR:—

<7 6a
Useful background
information

Syn OIS used in the text

is equal to —- tends to
ra) is approximately equal to fF is not equal to
> is greater than = is identical to
> is greater than or equal to < is less than
n! factOtial ise 1X 2K S XR < is less than or equal to
\k| modulus of k, i.e. size of k x infinity
irrespective of sign Lim limiting value as n > x
De summation os

Useful mathematical information

1 Algebraic identities
(a+b)*=a2+2ab+b? (a+b) =a*+3a"b+3ab?+b°
(a—b)* =a?—2ab+b? (a—b)* =a* —3a*b + 3ab* — b°
(a+b)* = a* + 4a*b + 6a*b? + 4ab* + b*
(a — b)* = a* — 4a°b + 6ab* — 4ab? + b*
a—b?=(a—b)(at+b) a—b> =(a—b)(a* +ab+b’)
a? +b? = (a+ b)(a* — ab+b’)
2 Trigonometrical identities
(a) sin? 6+ cos? =1; sec? @ = 1+ tan? 6; cosec?@ = 1 + cot” 6
(b) sin(A + B) = sinA cosB + cosA sinB
sin(A — B) = sinA cosB — cosA sinB
cos(A + B) = cosA cosB — sinA sinB

( —B) = cosAcosB + sinA sinB

cos(A
ene +tanB
tan(A + B)= —tanAtanB
por —tanB
tan(A— B)=
1+tanAtanB
(c) LetA=B=80.. sin 20 = 2sin@cosé@

cos 26 = cos? 6 — sin? @ = 1 — 2sin” @ = 2cos*6—1

2tané@
ri
2 earns
1 —tan?6

ix
 Useful background information

(d) Let = $ a sing = 2sin$cos$

cos = cos?S — sin? =1 —2sin?£ = 2cos?S —

2tan$
ie es ae)

(e) sinC +sinD = 2sin“ 4 cos “=7

sin C ~sinD = cos“ + sin°— 7

cos C+ cosD = 2cos +? cos =? ~

cosD — cos C = 2sin< +” sin S =”

(f) 2sinA cosB = sin(A + B) + sin(A — B)

2cosA sinB = sin(A + B) — sin(A — B)
2cosA cos B = cos(A + B) + cos(A — B)
2 sinA sin B= cos(A — B) — cos(A + B)
(g) Negative angles: sin(—é@) =—siné
cos(—@) = cosé
tan(—0@) = —tané
(h) Angles having the same trigonometrical ratios:
(i) Same'sine: 6 and (180° — @)
(ii) Same cosine: 6 and (360° — 6), i.e. (—6)
(iii) Same tangent: 4 and (180° + 6)
(i) asin@+ bcosé = Asin(6 +a)
asin #6— bcos@ = Asin(6 ~ a)
acos@
+ bsin@ = Acos(6 — a a)
acos@— bsin@ = Acos(6+ a)
A=+/a*+
where
b
a=tan — (0° < age 90°)

3 Standard curves
(a) Straight line

Sooe
SUE eedy. eee
ee ¥2 es
=i

Angle between two lines, tan @ = 72

—
1+ mm 2
For parallel lines, mz = m,
For perpendicular lines, mym2 = —1

2
Useful background information

Equation of a straight line (slope = m)

(i) Intercept c on real y-axis: y=mx+c
(ii) Passing through (x,, y,): y—y) = m(x— x1)

(iii) Joining (x1, y1) and (xp, y2); YY = *= 1

Vea Vina ao aT
(b) Circle
Center at origin, radiusr: 9x + y? =
Center (h, k), radius r: (x —h)*? + (y—k?* =
General equation: x7 +y" + 29x + 2fy+c=0
with center (—g, —f): radius = ,/g? + f2 —c
Parametric equations: x = rcosé, y =rsin@
(c) Parabola
Vertex at origin, focus (a,0): —_y* = 4ax
Parametric equations : x=at?, y =2at
(d) Ellipse

Center at origin, foci (+Vv a + b?, 0): -5

mn Ai |
where a = semi-major axis, b = semi-minor axis
Parametric equations: x = acos@, y = bsin@
(e) Hyperbola
x?
Center at origin, foci (+V a2 + b?, 0): Pm
Parametric equations: x = asec@, y = btané
Rectangular hyperbola:
a
Center at origin, vertex (4 =2
ee) ie)
; a
ie. xy=C where c = —=
v2
Parametric equations: x = ct, y=c/t

4 Laws of mathematics
(a) Associative laws — for addition and multiplication
a+(b+c)=(a+b)+c
a(bc)= (ab)c
(b) Commutative laws — for addition and multiplication
a+b=b+a
ab = ba
(c) Distributive laws — for multiplication and division
a(b+c) =ab+ac

A (provided a ¥ 0)
a aa
 Mia hitii te

wool nfetw- ie votwsv

4A whee CMe etd 3 nGieupe sions

Ae aA eI) — nA seal (b)
eel A Fr, ; .
= ey Sy t
deaca ta) POR NR BIN?
: Sree
‘ss = 4 Samba GASH
— te pe 210) he ys} fo? MiaoJa toln
319 tiw ;
, : iN Pe ti — ‘ » ie ca
® : —_ | DES aN beg = analinups Gstuainsi®
an 7

Coes ms — “si ~ plexhequll (4)

Preface
It is now nearly 40 years since Ken Stroud first developed his approach to
personalized learning with his classic text Engineering Mathematics, now in its
sixth edition. That unique and hugely successful programmed learning style is
exemplified in this text and I am delighted to have been asked to contribute to
it. I have endeavored to retain the very essence of his style that has
contributed to so many students’ mathematical abilities over the years,
particularly the time-tested Stroud format with its close attention to
technique development throughout.
The first Program introduces methods of solving simultaneous linear
equations in two or three unknowns and concludes with the problems
encountered when such solutions are either not unique or do not exist. In
Program 2, determinants are introduced and their use in solving simultaneous
linear equations is exemplified. Again the Program concludes with the
conditions for consistency. Matrices are introduced in Program 3 and their use
in solving simultaneous equations via the inverse matrix and Gaussian
elimination is described. The early work on Matrices is extended in Program 4
covering eigenvalues and eigenvectors, the Cayley-Hamilton theorem,
systems of first and second order linear differential equations, matrix
diagonalization and matrix transformations. Having covered an extensive
amount of manipulative Linear Algebra in the first four programs the fifth
program provides a quick review of geometric vectors. This is to enable an
introduction to the final two programs introducing the foundations of vector
spaces and linear transformations so as to place the earlier work within the
context of the encompassing theory of Linear Analysis; this in the hope of
inspiring the student to take the subject further.
To give the student as much assistance as possible in organizing their study
there are specific Learning outcomes at the beginning and Can You?
checklists at the end of each Program. In this way, the learning experience is
made more explicit and the student is given greater confidence in what has
been learned. Test exercises and Further problems follow, in which the
student can consolidate their newly-found knowledge.
This is a further opportunity that I have had to work on the Stroud books,
having made additions to both the Engineering Mathematics and Advanced
Engineering Mathematics texts. It is as ever a challenge and an honor to be able
to work with Ken Stroud’s material. Ken had an understanding of his students
and their learning and thinking processes which was second to none, and this
is reflected in every page of this book. As always, my thanks go to the Stroud
family for their continuing support for and encouragement of new projects
and ideas which are allowing Ken’s work an ever wider public.
Huddersfield Dexter J Booth
January 2008

xiii
 onl wucinetlvasiz petvh
walt; acd) clivt godui
nl J2ix9 “an ob te cupin«
Sponhatiuer yoiviGe rio: tise th DO MID ,
Sry. tite aibuioies tmergoud siti mise. beitiiqnress A enoilsups {
mao ir bes © itetgort ni bsouhouei « <r 15M yoielilanies pol enoiibinn
seeicd ine Zitiem serevni oc? uty Guitisuy euoersiiwiit Enivtos ai
2 meen ii helaaiase e29k i NO Mow yis89 5a bodtingsd «i cobsnimils
begs aaneti-reigco 9) wiomevosgiy bran prmleviegia gaievo
f ouje istisstilh ascall wine breve bas tet io anisizye
nb
despa goivel 2rohextvalzte: vite bras. notsxileciogsth
ee x nt ee yeonil svitelugingat lo tayons
Paves i a2 ot 4 an na! iniocn to Wafver Ssiop & eabivoig roergory
Ss aLOME i qowt lant oti atnolrauborial :

J -
ys
pita ot sg 03 tm
:
69 rs

ae"
TA *
S(Ftae
mt PTOPSL

mots
aft Sa : ‘- e'er a
=— & > 6 a

= a —
Linear equations and
simultaneous linear
equations
Learning outcomes
When you have completed this Program you will be able to:
Solve any linear equation
Solve simultaneous linear equations in two unknowns
Solve simultaneous linear equations in three unknowns
Determine whether or not a solution exists to a pair of simultaneous
equations in two unknowns
2 Linear Algebra

Linear equations
A linear equation in a single variable (unknown) involves powers of the
variable no higher than the first. A linear equation is also referred to as a
simple equation.

Solution of simple equations

The solution of simple equations consists essentially of simplifying the
expressions on each side of the equation to obtain an equation of the form
ax+b=cx+d giving ax —cx =d—b and
—b
HeHCen a as provided a ¥#c.

Example
If OX ok — 24 4X = 1 36-7 x
then 15x+4 = 7x+ 36
SO. 8X = 32 therefore x= 4
Similarly:
if 5(x — 1) + 3(2x + 9) — 2*= 4(3x — 1) + 2(4x+ 3)
then be eee

Because
5(x — 1) + 3(2x+9) — 2 = 4(3x— 1) + 2(4x
4+3)
so SXx-5+6x+27 —2=12x-4+8x+6
11x+ 20 = 20x42
SO, 18 == 9x therefore. — Z,
Equations which appear not to be simple equations sometimes develop into
simple equations during simplification. For example:
(4x + 1)(x + 3) — (x4 5)(x — 3) = (8x + 1)(x —- 4)
(4x7
+ 13x +3) — (x?+ 2x — 15) = 3x2 -11x-4
$0 3x7 11x 18 = 3x" —11x7— 4
3x? can now be subtracted from each side, giving:
Tbea ke} 5 — iiss — 2h
S022 therefore x = —1
Linear equations and simultaneous linear equations

It is always wise to check the result by substituting this value for x in the
original equation:
LHS = (4x + 1)(x + 3) — (x + 5)(x — 3)
= (—3)(2) — (4)(—4) = -6+ 16 = 10
RHS = (3x + 1)(x — 4) = (-2)(-5)=10 .°. LHS =RHS
So, in the same way, solving the equation:
(4x + 3)(3x — 1) — (5x — 3)(x + 2) = (7x + 9)(x — 3)
we have x =

x=-3

Simplification gives 7x* —2x+3 = 7x? —12x-—27

Hence 10x = —30....... x =-—3
Where simple equations involve algebraic fractions, the first step is to
eliminate the denominators by multiplying throughout by the LCM (Lowest
Common Multiple) of the denominators. For example, to solve:
Ae pict Son Oe Se hs
Z Pape 6
The LCM of 2, 3, 4 and 6 is 12:
12(x+2) 12(x%+5)_ 12(2x—S) n 12(x + 3)
2 =ae 4 6
6(x + 2) —4(x+5) = 3(2x — S) + 2(x + 3)

That was easy enough. Now let us look at this one.

To solve ee
X=3)eXGex-=5
Here, the LCM of the denominators is x(x —3)(x—5S). So, multiplying
throughout by the LCM, we have
4x(x — 3)(x—S) ‘ 2x(x — 3)(x — 5) > 6x(x — 3)(x — 5)
x-3 x x-5
After cancelling where possible:
4x(x — 5) + 2(x — 3)(x — 5) = 6x(x — 3)
$6, HnisHing If Off, XS. en scenes
Rs eo Linear Algebra

9) 72
ae

Because 4x”— 20x + 2(x? — 8x +15) = 6x” — 18x

4x” — 20x + 2x” — 16x + 30 = 6x” — 18x
6x” — 36x + 30 = 6x” — 18x
Subtracting the 6x term from both sides:

= 36%+ 30 =—l16x ~.. 30=18x—.4, x =2

ncn 3 5 8 ™ =
g the same way
Workinin oo ie aa 0

PiVesmileeSOLMMON 4 esacs
one nse

The LCM of the denominators is (x — 2)(x — 3)(x + 3), so

3(x — 2)(x — 3)(x + 3) . 5(x —2)(x-—3)(x+3) 8(x-—2)(x -—3)(x+3) ei
x-2 x-3 x+3 0

3(4= 3)(x+3) 45 2D 3S 8r— 2a 3y=0

3(% — 9) +:5(x? x +6) — (x? — 5x +6) = 0
3x? — 27 + 5x” + 5x — 30 — 8x? + 40x — 48 =0
45x
— 105 =0

OT
Sil aie.
ee:
eed
Now on to Frame 7

é
Linear equations and simultaneous linear equations
5

Simultaneous linear equations with two

unknowns
A linear equation in two variables has an infinite number of solutions. For
example, the two-variable linear equation y — x = 3 can be transposed to read:
y=x+3
Any one of an infinite number of x-values can be substituted into this
equation and each one has a corresponding y-value. However, for two
such equations there may be just one pair of x- and y-values that satisfy both
equations simultaneously.

1 Solution by substitution
To solve the pair of equations
ox + 2Zy = 14 (1)
3x — 4y = 24 (2)
5x

From (1): 5x+2y=14 -. 2y=14—5x .. y=7->

If we substitute this for y in (2), we get:
3p,¢
3x—4(7-5) =

3x — 28 + 10x = 24
13 Se eee Se
If we now substitute this value for x in the other original equation, i.e. (1), we
Seu
5(4) + 2y = 14
20+2y=14 ©. 2y=-6 .. y=-3
Wehave
x = 4, y=—-—3
As a check, we can substitute both these values in (1) and (2):
(1) Sx+2y=5(4)+2(-3) =20-6 =14
(2) 3x—4y = 3(4) —4(-3) =124+12=24
x = 4, y = —3 is the required solution.

Another example:
To solve
3x+4y =9 (1)
2x+ 3y =8 (2)
Proceeding as before, determine the values of x and y and check the results by
substitution in the given equations.
TIE ese ibteyaieo> eS eee
i —__—___
 Linear Algebra

Because
9) +230
(1) 3x+4y=9 ~ 4y=9-—32 So y=3->

Substituting this for y in (2) yields:

Dhar ‘ 271 xOR ; xu 5 vith «
2x+3(Z-)=8 that is 2x+7-{=8 ee ok tk L== 9

Substituting this value for x in (1) yields:

—15+4y=9 .. 4y=24... y=6 iy

2 Solution by equating coefficients

To solve 3x+2y=16 (1)
4x —3y=10 (2)
If we multiply both sides of (1) by 3 (the coefficient of y in (2)) and we
multiply both sides of (2) by 2 (the coefficient of y in (1)) then we have
9x + 6y = 48
8x — 6y = 20
If we now add these two lines together, the y-term disappears:
17x=68 4. x=4
Substituting this result, x = 4, in either of the original equations, provides the
value of y.
In(1) 3(4)+2y=16 ©. 12+2y=16 ©. y=2
x=4,y=2
Check in (2): 4(4) —3(2) =16-—6=10
Had the y-terms been of the same sign, we should, of course, have
subtracted one line from the other to eliminate one of the variables.

Another example:
To solve 3x+y=18 (1)
4x+2y=21 . (2)
Working as before: 2'= <.<2.:.::... FOE deer eee
Linear equations and simultaneous linear equations —a

x=75,y=-45 9
§ hes "ey 6x + 2y = 36
(2) 4x-+2y = 21
Subtract: 2x =15 afer 75
Substitute x = 7.5 in (1):
3(7.5)+y =18 22.5 +y =18
y=18-22.5 =-45 y= 45
Check in (2): 4x + 2y = 4(7.5) + 2(-4.5) = 30-9 =21
x=7.5,y=-45
And one more — on your own. 10
Solve 7x-4y=23 (1)
4x —-3y=11 (2)
Working as beforenx 2195 0... Alea A Vat Be See

» hy y= 3 11

(1) x3 21x-12y=69
(2) x4 16x-12y=44
Subtract: 5x =25 ., x=53
Substitute in (2); 20-3y=11 .. 3y=9 ©. y=3
x=3,y=3
Check in (1): 7x — 4y = 35 — 12 = 23
8 Linear Algebra

Simultaneous linear equations with three

unknowns
12 With three unknowns and three equations the method of solution is just an
extension of the work with two unknowns.

Example 1
To solve 3x+2y— z=19 (1)
4x— y+2z= 4 (2)
2X ay —5z = 32 (3)
We take a pair of equations and eliminate one of the variables using the
method in Frame 7:
wa 2y > 2H 19 (1)
4x— y+2z= 4 (2)
(1) 32 6x + 4y — 2z = 38
(2) AX—y+2Z= 4
Add: 10x + 3y —42" (4)

Now take another pair, e.g. (1) and (3):

(1) #5 15% 4- 10y =Sz =95
(3) 2x+ 4y—5$z=32
Subtract: 13x4} 6y = 63 (5)

We can now solve equations (4) and (5) for values of x and y in the usual way.

@
Linear equations and simultaneous linear equations

x=3,y=4 13

Because 10x + 3y = 42 (4)

13x + 6y = 63 (S)
(4) x2 20x + 6y = 84
(5S) 13x + 6y = 63

Subtract: 7X = 21 po A me
Substitute in (4): 30+ 3y=42 .. 3y=12 .. y=4
Then we substitute these values in one of the original equations to obtain the
value of z:
eg. (2) 4x-y+2z=12-4+2z=4
2z=-4 ., z=-2
x=3,y=4,z=-2
Finally, substitute all three values in the other two original equations as a
check procedure:
(1) 3x+2y- z=94+8+4+2=19
(3) 2x+4y-—5z=6+16+10=32 -—soall is well.

The working is clearly longer than with only two unknowns, but the method
is no more difficult.
Here is another.

Example 2
Solve 5x -3y-—2z=31 (1)
2x+6y+3z= 4 (2)
4x+2y— z=30 (3)
Work through it in just the same way and, as usual, check the results.
NO ee: Linear Algebra

14
CTS 15x -—9y-—6z= 93
(ay e2 4x+12y+6z= 8
19x + 3y == hGy! (4)
(1) 5x—3y-—2z= 31
(S)ex 2 8x+4y—2z= 60

3x + 7y = 29 (S)
Solving (4) and (S) and substituting back gives the results above:
X=5, y=2, zZ=-6

|15 Sometimes, the given equations need to be simplified before the method of
solution can be carried out.

Pre-simplification

Example 1

Solve the pair of equations:

2(x + 2y) + 3(3x —y) = 38 (1)
4(3x + 2y) — 3(x+ Sy) = -8 (2
&
2x + 4y + 9x —* 3y= 38 .. lix-+y=38 (3)
12x + 8y — 3x - 15y = -8 “. 9x-7y=-8 (4)
The pair of equations can now be solved in the usual manner.
Linear equations and simultaneous linear equations on ee ois
ater se Pas ‘

x=3,y=5 16

Because (3) x 7 77X + 7y = 266

(4) 9x-7y= -8

86x = 200 yoee 3S Ea

Substitute in (3): 33+y=38 - y=5
Check in (4): 9(3) — 7(5) = 27 —35 = -8

x=3,y=5

Example2
ee 1 X= 2p x41
COs (aan
Bye Pax — Sy _ Sx
| Z 4
20(2x — 1) F.20(x—2y) 20(x +1)
For (1) LCM = 20
5 10 4
4(2x — 1) + 2(x — 2y) = S(x + 1)
8x —4+2x-—4y =5x+5
5x-4y=9 (3)
Similarly for (2), the simplified equation is

9x —6y =4 17
Because
eye, sy _ ke
LCM = 12
a Z 4
12(3y + 2) Ss12(4x — 3y) >. 12(5x + 4)
a 2 4
4(3y + 2) + 6(4x — 3y) = 3(Sx + 4)
12y + 8 + 24x — 18y= 15x+ 12
24x — 6v +8 = 15x4+ 12
9x —6y =4 (4)
So we have 5x -—4y=9 (3)
9x -6y =4 (4)

Finishing off, we get

ae ate 2, crac pd eae cota
oe a Se Te ee a ee Se en
ee Se. pe Linear Algebra

x= —19/3, y=—61/6

Because
9 x (5x — 4y = 9) gives 45x — 36y = 81
5 x (9x — 6y = 4) gives 45x — 30y = 20
: : ‘ 61 ge }
Subtracting gives —6y = 61 so that y= ore Substitution then gives:

244 19 190 19
Sx = 9+ 4) a eS ee SO x= =

That was easy enough.

Example 3
S(x + 2y) — 4(3x + 4z) — 2(x+ 3y — Sz) = 16
2(3x —y) + 3(x — 2z) + 4(2x — 3y +z) =—-16
4(y — 2z) + 2(2x — 4y — 3) — 3(x+ 4y — 2z) = —62
Simplifying these three equations, we get

—9x+ 4y-—6z=16
17x — 14y —2z = -16

Solving this set of three by equating coefficients, we obtain

<n
Linear equations and simultaneous linear equations ats bea: ||

20|
Here is the working as a check:
—9x+4y-—6z= 16 (1)
17x — 14y —2z = -16 (2)
x — 16y — 2z = —56 (3)
(1) —9x+ 4y-6z= 16
(2) x3 51x — 42y — 6z = —48
Subtracting: 60x — 46y =-64 .. 30x-—23y =-32 (4)
(2) 17x — 14y -2z=—-16
(3) x — 16y — 2z = —56

Subtracting: 16x+ 2y = 40 .. 8x+y=20 (S)

(4) 30x — 23y = —32
(S\ x23 184x + 23y = 460
Add: 214x = 428 X=2
Substitute in (5): 16+y = 20 y=4
Substitute for x and y in (3):
2 — 64 —2z = —56
-62-2z=-56 .. z+31=28
z=-3
Check by substituting all values in (1):
—9(2) + 4(4) — 6(—-3) = -18 + 16+ 18 = 16
x=2,y=4,z=-3
Ag Linear Algebra

Existence of solutions
o1 Consider the pair of simultaneous equations from Frame 7:
5x + 2y = 14
which have the solution x = 4, y = —3
3x — 4y = 24
There is a graphical interpretation of this solution. Each of the two
simultaneous equations can be re-written:
5x + 2y = 14 can be written as 2y = —5x+ 14
So that

ya-2x47

Similarly 3x — 4y = 24 can De written as y= ......s..0ss

The answer is in the next frame

22 A.
3 eels

3x — 4y = 24 can be written as 4y = 3x — 24

Hence y = Sy —6 |
ar 4
The two equations, written as:

y= ~2x se

v= a —6

are clearly seen to be the equations of two straight lines, the first with slope
—S/2 and y-intercept 7 and the second with slope 3/4 and y-intercept —6
Vs

As can be seen from the figure the two lines intersect at the point with
coordinates x = 4, y=3 which is the solution of the two simultaneous
equations.

a
Linear equations and simultaneous linear equations

Because each of the two equations

SX -+ 2y = 14
3x — 4y = 24
represents a straight line, the existence of a solution means there is a point
that is common to both lines, in other words the lines intersect.
The next question is, what if the lines are parallel and never intersect?

Move to the next frame

Consider the pair of simultaneous equations: 23

2X + Sy = 10
4x-+10y=5
Trying to solve them by multiplying the first equation by —2 so as to equate
coefficients produces:
—4x — 10y = —20
4x+10y=5
which gives, on adding both sides, 0 = —15 which is............

The answer is in the next frame

24
Trying to solve them by substitution produces from the first equation
2x + Sy = 10

Next frame

y=—Sx42
25

Because 2x + Sy = 10:
a
Sy=-2x+10 ©. y= Tgxtte

Substituting this into the second equation gives ............

Next frame
 Linear Algebra

Therefore
?
4x + 10v = 5 becomes 4x + 10(- =x - 2)= 5

that is:
4x — 4x + 20 = S giving 1S’‘= 0 which is ............
The answer is in the next frame

27
But why does this happen?
The first equation 2x + Sy = 10 can be written as y= ............
The answer is in the next frame

28 2
ys——xt2

2 ,
where y= —sxt2 is the equation of a straight line of slope —2/5 and
y-intercept 2. *

ae

The second equation 4x + 10y = 5 can be written as vy= ............

Next frame
Linear equations and simultaneous linear equations
17

Because 4x + 10y = 5S:

2 1
ie y 5 xX+ 5)

2 Le.
Now, y= agt+5 is the equation of a straight line of slope —2/5 and
y-intercept 1/2.

So the two lines are parallel and have no point in common. This means that
there are no values of x and y that simultaneously satisfy both equations — we
cannot simultaneously solve the two equations
2x+Sy = 10
4x+ 10y=5
and that is why we get nonsense when we try.
Next frame

Non-existence of solutions
Whether or not a solution exists to a pair of simultaneous linear equations in
two variables can be determined by inspecting the slopes of the lines that they
represent. For example:
2 1
ay Sok Le ey fo aes
meine ene can be re-written as > ae i that is - 4
4x —7y = -2 X+2=/Y yasxts
18 Linear Algebra

The first equation is the equation of a straight line of slope 2/3 whereas the
second equation is the equation of a straight line of slope 4/7. The two lines
are therefore not parallel and so intersect at a single point. The equations,
therefore, will have a common solution. On the other hand, the equations:
5x + 6y = 1 5x—1=-—6y
can be re-written as
—15x — 18y
= 20 —15x
— 20 = 18y

Wi 5
that is 6 6
eee we
6 9
In this case the two lines are parallel, each having a slope —125/6 so they
never intersect. The equations, therefore, have no common solution.
Check the existence of a solution to each of the following pairs of
equations:
(a) x—3y=6
—4x+12y=10
(bl) —x+4y=8
2x — Sy = 16
(c) 7x-y=14
3y = 21x -2
(d) —9x+2y=-18
36x — 8y = 72
‘ The answer is in the next frame

31 (a) non-existent
(b) existent
(c) non-existent
(d) identical equations

Because:
x
(a) x—3y ='6 Yogre
can be written as
4x + 12y =10 iS
eG
Same slopes so a solution does not exist

(b) -x+4y=8 erie

= x 2

can be written as
2x — Sy =6 at ae: 16
5 5
Different slopes so a solution does exist

@
Linear equations and simultaneous linear equations 19

(c) 7x-y=14 Ye 7x— 14

: can be written as 2
Sy = 21x -—2 Y id =

Same slopes so a solution does not exist

9
(d) —9x +2y
=-18 Vizagee
can be written as
36x — 8y = 72
y= 7% —9

Both equations the same and so the coordinates of any point on the line
satisfy the equation there is an infinity of solutions.
Move to the next frame

Simultaneous equations in three variables

The set of three simultaneous equations in three unknowns from Frame 12: 32 |
$x+ 2y—z=19
4x-—y+2z=4 have the solution
x = 3, y= 4 andz= -2
2X + 4 —$Z = 32
But what does this mean graphically? The equation 3x + 2y — z = 10 when
plotted against a set of three Cartesian coordinate axes produces a............
The answer is in the next frame

33
So when each of the three equations are plotted against the same set of three
Cartesian axes, this will produce three planes.
When any two, non-parallel planes intersect they do soina............
The answer is in the next frame

34 |
A single plane intersects a non-parallel straight line ina............
Next frame

35 |
So three non-parallel planes can intersect in a single ............
The answer is in the next frame
 Linear Algebra

Because any two of the non-parallel planes intersect in a straight line and the
third plane can intersect that straight line in a point, provided the third plane
is not parallel to the line of intersection of the other two planes.
The three planes represented by the equations
3x+2y—z=19
4x-—y+2z=4
2x + 4y — 5z = 32

AN Cs7 7 ee =
Next frame

Kamit pai A Ze eZ.

Because that is the solution to the simultaneous equations found algebrai-

cally.
Recall that if two equations with two unknowns have a solution, then the
straight lines that each represent are not parallel but intersect. If they do not
have a solution, then the straight lines that each represent are parallel and do
not intersect.
A similar situation occurs with three equations in three unknowns:
e If they have a solution, then the planes that they represent are not mutually
parallel and intersect in a single point.
e If any pair of planes are distinct but parallel, then there will be no single
point whose coordinates simultaneously satisfies each equation and the set
of equations will not have a solution.
e If any one plane is parallel to the line of intersection of the other two
planes, then there will be no point in common with all three planes and the
set of equations will not have a solution.

@
Linear equations and simultaneous linear equations PERSE TE

|B) Summary
A linear equation involves powers of the variables no higher than the first.
A linear equation in a single variable is also referred to as a simple equation.
2 Simple equations are solved by isolating the variable on one side of the
equation.
3 A linear equation in two variables has an infinite number of solutions. For
two such equations there may be only one solution that satisfies both of
them simultaneously.
4 Simultaneous equations can be solved:
(a) by substitution
(b) by elimination.
5 For some sets of simultaneous equations a unique solution may not exist.

F Review exercise
1 Solve the following linear equations: 39
(a) 2(x-— 1) —4(x4+2) =3(x+
5) + (x—-1)
x—-1°> x+1 X+2
TESS Ea a Pega’
> 5 2
(<) Oy
i Se aa
2 Solve the following pairs of simultaneous equations:
(a) x-y=2 by elimination
2x+ 3y=9
(b) 4x+2y = 10 by substitution
3x —Sy=1
3 Solve the following set of three equations in three unknowns:
x+y+z=6
2x -y+3z=9
x+2y-—3z=-4

4 Simplify and solve the following set of simultaneous equations:

2(x + 2y) — 3(x+ 2z) — 4(y —2z) =-1
3(2x — y) + 2(3y — 4z) — S(3x — 2z) = —7
—(x—y)+(x-—z)+(y+z)=-2
 Linear Algebra

5 Determine which of the following pairs of equations have a solution:

(a) —3x+7y=21
6x — 14y = 42

(b) x+y=1
x-y=3
(c) x+3y=5
y=2
(d) Sty
—15x+6y=11

40 IO )o2@ 1) = 4(x 2) = 30x 15) 4 (x = 1). ’

Eliminate the brackets to give 2x — 2 —- 4x -8 = 3x+15+x-1.
Simplify each side to give —2x —- 10 = 4x+ 14.
Add 2x — 14 to both sides of the equation to give —24 = 6x so that
x= —4.
x-1 x+1 x+2
(b) 2 oe Wer
Multiply throughout by 4 to give 2(x—1)—2(x+1)=20-—x-2.
Simplify to give —4 = 18 — x, therefore x = 22.
3 mi pipes
5 2 , V(x — ;
(c) males ey Multiply throughout by x(x+2)(x—1) to give

3x(x — 1) — S(x+ 2)(x — 1) = —2x(x+2). Eliminate the brackets to

give 3x2 «3x — 5x? —5x+.10 = —2x?— 4x.
Simplify to give —2x” — 8x + 10 = —2x? — 4x. Add 2x? to both sides to
give —8x+ 10 = —4x. Add 8x to both sides of the equation to give

10:= 4x so.thatx =>.

2 tay — Ves {1]

2x + 3y =9 [2] Multiply [1] by —2
—2x+2y = —4 Add [2] to give Sy = 5 so that y = 1. Substitution
in [1] gives x = 3
(b) 4x+2y=10 {1]
3x Sy=1 [2] Subtract 4x from both sides of [1] and then
divide by 2 to give: y=5-— 2x. Substitute this
expression for y in [2] to give 3x — 5(5 — 2x) = 1.
Simplify to give 13x-—25=1 so that x =2 and
v=) =a

?
Linear equations and simultaneous linear equations Pee Taal in Th

3 x+y+z=6 {1]
2x—-y+3z=9 [2]
x+2y—3z=-4 [3] Add [1] and [3] and subtract [2] to give
(2x + 3y — 2z) — (2x -y + 3z) = (6—4)
— 9.
That is
4y —Sz=-7 [4] Subtract [1] from [3] to give
y—4z=-10 [S] Multiply this by 4 to give
4y — 16z = —40 [6] Subtract [6] from [4] to give 11z = 33 so
that z = 3. Substitute this value in [5] to give
y =-10+ 12 =2 so from [1],
x=6-3-2=1.
4 2(x+2y) —3(x+2z) —-4(y—2z) =-1 [1]
3(2x — y) + 2(3y — 4z) — 5(3x —2z) =-7 [2]
—(x-y)+(x-—z)+(y+z)=-2 [3] Eliminate the brackets to give
—X%+2z=-1 [4]
—9x + 3y+2z=—-7 [S]
2y=-2 [6] Thus y=-—1. Substituting this
value in [5] gives
—9x+2z7=-4 Subtracting [4] gives —8x = —3 so that
3 Wet e
x= 3° Substitution in [4] gives z = — Té

—3x+7y=21 7y = 3x+21
5 (a) ae can be written as y ve
6x — 14y = 42 14y = 6x — 42
|
y= :sa +3
That is . They are parallel (both with slope 3/7) so a
y = 5X -—3

solution does not exist.

=1 =-x+1
(b) nits can be written as f :
x-y=3 y=x-3
They are not parallel (slopes 1 and —1 respectively) so a solution
exists.

x+3y=5 ae a 5
(c) Y=» can be written as’ 3.3
Wwe y=2
They are not parallel (slopes —1/3 and 0 respectively) so a solution
exists.
5x -—2y=17
(d) —15x+ 6y=11
They are parallel (both with slope $/2) so a solution does not exist.
24 Linear Algebra

41 You have now come to the end of this Program. A list of Can You? questions
follows for you to gauge your understanding of the material in the Program.
You will notice that these questions match the Learning outcomes listed at
the beginning. Now try the Test exercise. Work through them at your own pace,
there is no need to hurry. A set of Further problems provides additional valuable
practice.

4 Can You?
|42 Checklist 1 :
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames
e Solve any linear equation? to (6)
Yes | ere No
e Solve simultaneous linear equations in two unknowns? to
Yes [a] [J LJ EJ No
e Solve simultaneous linear equations in three unknowns? to
Yer = lye) ey ee SING
e Determine whether or not a solution exists to a pair of
simultaneous equations in two unknowns? G1) wo
Yes L LJ LJ L] No

Test exercise 1
|43 1 Solve the following linear equations:
(a) 4(x+5S)—— 6(2x+3)
6( =3(x+ 14)eu2(S
—2(5 2x)a+ 9
2X+1 2x+5 x-1
Rescue caaes
ReFa Oeks
eyed
ie
(d) (4x — 3)(3x — 1) — (7x + 2)(x + 1) = (5x + 1)(x — 2) — 10

2 Solve the following pairs of simultaneous equations:

(a) 2x+3y=7
5x —2y=8 by substitution
(b) 4x+2y=5
3x+y=9 by elimination

2
Linear equations and simultaneous linear equations 25

3 Solve the following set of three equations in three unknowns:

2x+ 3y-z=-5
xX —4y+2z=21
Sx + 2y —3z = -4
4 Simplify and solve the following set of simultaneous equations:
4(x + 3y) — 2(4x + 3z) — 3(x —2y —4z) = 17
2(4x — 3y) + 5(x — 4z) + 4(x — 3y + 2z) = 23
3(y + 4Z) + 4(2x —y —z)+2(x+ 3y —2z) =5
5 Without solving them, show that a solution exists to the pair of simultaneous
equations:
4x —2y =6
—2x+4y=6

Further problems 1
Solve the following equations, numbered 1 to 15: 44 |

1 (4x+5)(4x — 3) + (5 — 4x)(5 + 4x) = (3x — 2)(3x + 2) — (9x — 5)(x + 1)

2 (3x—1)(x — 3) — (4x — 5)(3x — 4) = 6(x — 7) — (3x -—5)”
AXA 7 X= 7 X46
3
a ee)
k= SPAT 2 2X43 XFS
tee 3
§ (3x+1)(x — 2) — (5x —2)(2x — 1) = 16 — (7x + 3)(x +2)
Ad chan ns SU
3 iar: 4
7 8x—Sy=10
6x —4y=11
5 7
8 2 +-=
X43. x x-2
9 (3 —4x)(3 + 4x) = 3(3x — 2)(x + 1) — (Sx — 3)(Sx + 3)
10 3x-2y+z=2
x-3y+2z=1
2x+y—3z=-5S
Al 2x-y+3z=20
x-6y-z=-4l
3x + 6y + 2z = 70
 Linear Algebra

12 3X + 2y+5Z7=2
SX + 3y —-22=4
2X — Sy —3z=14
13 SX — 6y-+3z=—-9
2X — 3¥ +22 5-5
34 —/) +57 =—16
3x+2 x+2y x-3
14
4 2 12
2) oe eee tL
5 4 10
ban Ne Sa ae
15
seLeT
16 Simplify each side separately and hence determine the solution:
x-2 ,tee A, Sa | x-—3
Cy ee ee ee ak wee ee
17 Solve the simultaneous equations:

2 7 2
1 ee y
7 —- 6 Gan Wr

18 Writing u for wana BF and v for at solve the following equations for
x + 8y
u and v, and hence determine the values of x and y:
eed ag 1
x+8y 8x-y ,
x+8y | 8x—y

19 Solve the pair of equations:

oe
x—2y x+y
eee
x—2y x+y
20 Solve:
4 7
xl 4p
2(x+1)
De ee,
2-y
Linear equations and simultaneous linear

21 Without solving them, determine whether or not a solution exists to each of

the following pairs of equations:
(a) 3x-2y=4
—15x+10y =8
a eaive 2ir 52 1
(b) 2x—4y 4x-—2y
(2x —y)(x—2y)
ae
4x—2y 2x-—4y (2x—y)(x—2y)
Pitan TAY 5
oS
2Xx+y 2 oye*
hag 43 = 3x
Determinants

Learning outcomes in
When you have completed this Program you will be able to:
e Expand a 2 x 2 determinant
e Solve pairs of simultaneous linear equations in two variables using
2 x 2 determinants
e Expand a 3 x 3 determinant
e Solve three simultaneous linear equations in three variables using
3 x 3 determinants
e Determine the consistency of sets of simultaneous linear equations
e Use the properties of determinants to solve equations written in
determinant form
&
Determinants TAR 7 29

Determinants
You are quite familiar with the method of solving a pair of simultaneous
equations by elimination.
e.g. To solve 2x+3y+2=0 (a)
3x+4y+6=0 (b)
we could first find the value of x by eliminating y. To do this, of course, we
should multiply (a) by 4 and (b) by 3 to make the coefficient of y the same in
each equation.
So 8x+ 12y+8=0
9x+12y+18=0

Then by subtraction, we get x + 10 = 0, i.e. x = —10. By substituting back in

either equation, we then obtain y = 6.
So, finally, x =-10, y=6
That was trivial. You have done similar ones many times before. In just the
same way, if:

ax+byy+d,;=0 (a)
a2X+boy+dz=0 (b)
then to eliminate y we make the coefficients of y in the two equations
identical by multiplying (a) by............ ANGHD) DY soovenn- ones

(a) by bz and (b) by b; 2 |

Correct, of course. So the equations:

a,x+biy+d,=0
a2Xx + boy + dz A)
become ayb2x + by boy + bed, = 0
d2b,x + b, boy + bid2 =0

Subtracting, we get:
(a, b2 = azb,)x + body _- bid> = 0)

so that (a,
bz — azb,)x = b,d2 — b2d,
Then bi a Bae ee
ee
30 Linear Algebra

3 ; ve bide = bod
~ ay bz — azb,

In practice, this result can give a finite value for x only if the denominator is
not zero. That is, the equations:
ax+by+da, = ()

a2x + boy + do —()

give a finite value for x provided that (a,b2 — a2b,) £0.

Consider these equations:
3x%-+2y=—5=0
4x+3y—-7=0
In} thiscase, a; — 3) Dn—Z2) as — 45 bo 25

a,b2 — arb, FORO tae,

=—9—38.—.1

will
This is not zero, so there { }be a finite value of x.
will not

4
The expression dbz — dazb, is therefore an important one in the solution of
simultaneous equations. We have a shorthand notation for this:
b
abo = a2b, = is

For a to represent a,b2—a2b,; then we must multiply the terms

diagonally to form the product terms in the expansion: we multiply |
ot 2
2
and then subtract the product s a ie. + \,and—- /
2

eg. F HP 2|-|5 | =3x2-5x7=6-35=-29

So sr KS n Dime
ne 9 1 Sn ree
Determinants 31

indie
ae 1

ay b,
is called a determinant of the second order (since it has two rows and
a2 bo

two columns) and represents a;b2 — a2b,. You can easily remember this as
+\- 7.
Just for practice, evaluate the following determinants:

(a)
42 » (Db)
7 4 21
5 3 Gear “MORoe
Finish all three: then move on to Frame 6

A 52,3|=4x3-5x2=12-10= [2]
@ |§

Blo643\—
i=? xo-6x4=21124—(-3]
ie Feet faeces5| =
Zoe)1,|=2-3)-4«1=-6-4=[Eig
(c) 5
ayx+byy+d, =U
Now, in solving the equations ee + boy +
dz =0
b, dz — bod
we found that x = 2"! and the numerator and the denominator can
ayb2 — d2b,
each be written as a determinant.
bi d2 — bod, COT
IO aMa ; ayb2 = a2b, eeees) or

by dy|, |a, by 7
b> do f a2 bo

If we eliminate x from the original equations and find an expression for y, we

a\d2 mm a
obtain y= —, ————_|_
‘ ee — azb,
So, for any pair of simultaneous equations:
ax+biy+d, —\)

a2x + boy + dz = 0

we have wee b,d2 oa bod i. ad = aod,

a,b2 a ab, a, b> 4 ab,

Each of these numerators and denominators can be expressed as a

determinant.
So, PME niesys re er
Cv Wak ene
 Linear Algebra

and =
[b, d| Tag
ay By |
bz do a2 bz

We can combine these results, thus:

x =y. 1
lb, dl
dy, a, dy J a, by
bz do a2 dz a2 bp
Make a note of these results and then move on to the next frame

: ax+byy+d,=0
mae lo =—\()
x nat as 1
ae (hyd,dy | ay “ay | la,
ay by!

b2 d> a2 do a2 bo

Each variable is divided by a determinant. Let us see how we can get them
from the original equations.

(a) Consider oat Let us denote the determinant in the denominator

1 ay
bz dz
: _ |b, dy
by Ai, 1.5 Aj se by do

To form A, from the given equations, omit the x-terms and write down
the coefficients and constant terms in the order in which they stand:
elias eet rn
a2x + boy + dz =0 8 b2 d2

(b) Similarly for rere let Az = “s i

a2 d»

To form A2 from the given equations, omit the y-terms and write down
the coefficients and constant terms in the order in which they stand:
aqx+biy+d,=
#7 gives A, = a,
Gasdy
A2X + boy + do

2
Determinants

; 1
(c) For the expression Pe My denote the determinant by
Ao.
1 Dy
a2 b2

To form Ao from the given equations, omit the constant terms and write
down the coefficients in the order in which they stand:
ax+by+d,=0 a, by
da2x + boy +d, =0
gives Ss

Note finally that ae Bee

Ai 42 Ao
Now let us do some examples, so on to Frame 10

Example 1 10

5x+2y+19=0
To solve the equations {
3x+4y+17=0
The key to the method is:

tt dy ed
Le TRS OLS
To find Ao, omit the constant terms:
a 2
: do=|3 4 =5x4-3x2=20-6=14

‘, oid {a)

Now, to find A;, omit the x-terms:

5 SANE HAP Sere

eee) 11

Because A = i a =34-76=-22 tb)

Similarly, to find Az, omit the y-terms:
IOP oe ine
da =|3 17 = 85-—S7=28 (c)

Substituting the values of A;, Az, Ao in the key, we get

et RS
—42 28 14
from which x =............ Pi en 2 eee re
34 Linear Algebra

Now for another example.

Example 2

Solve by determinants {3x —2y+5=0

2x+3y—-14=0

First of all, write down the key:

SOR ot ee
A; Az Ao -
(Note that the terms are alternately positive and negative.)
2te 3
Then Ao 3 3 |=-4-9=-13 (a)

Now you find A; and Az in the same way.

13 Ay =e18a 92

Because we have {ax + Sy —14=0

3x—2y+5=0

Ate g| | al
3 ,-14 a —14
—2 5 5 —2
= 128 = 13. Ay 13

eee i mete
2 —14 2 -
3 = 5 3
= 10 — (—42) = 52 Oy es
+ Sillssteele
So that — =
ses Mh ast = Ao
and Ay
= =13; ~“A2
= 523" Ag
= =—13
A; —-13
Ae =13 1 a= Al

NY
Be SIC i ge
Do not forget the key:

per Laser
Ai Az Ao
with alternate plus and minus signs.

Make a note of this in your record book

Determinants

Here is another one: do it on your own. 14

Example 3

Solve by determinants:
4x — 3y + 20 =0
3x+2y—2=0
First of all, write down the key.
Then off you go: find Ap, A; and A> and hence determine the values of x
and y.
When you have finished, move on to Frame 15

15
Here is the working in detail:
Pe eae x -y 1
3x+2y—2=0 Ai Az Ao
4 -3
Ao =
v3 >|= —9)=8+9=17
ee?
-3 20
= = 6-40= -34
ie 4
4 20
= —8 — 60 = —68
i i 4
Bie el 2 LX Se,
An S1Le
A2 —68 ,

Now, by way of revision, complete the following:

) Ae
(a) se a Se

in le
(b) a gee

() a fla
c et ae a

d P q a ae ee
Cre ee
 Linear Algebra

16 Here are the results. You must have got them correct.

(a) 20-422
(b) —20—6=-—26
(c) ac — bd
(d) ps —rq

For the next section of the work, move on to Frame 17

Determinants of the third order -

17 A determinant of the third order will contain 3 rows and 3 columns, thus:

Each element in the determinant is associated with its minor, which is found
by omitting the row and column containing the element concerned.
ay by Cy
: : C2 P
e.g. the minor of a, is obtained a2 b2 C2
cm
a3 b3 C3
i a ay by Cy
: ; 2 :
the minor of b, is obtained a2 b2 C2
a3 C3
a3 b3 C3
ay Cj
a2 b
the minor of c, is 2 Pa obtained a2 C2
a3 b |
C3
So, in the same way, the minor of az is............

|18
Minor of a is bic
bz C3

Because, to find the minor of a2, we simply ignore the row and column
containing dz, i.e.

a bb
a2 b2 C3

Ge Ds) Ce
Similarly, the minor of b3 is ............
rr
Determinants reay an «
t

> Tew

.
Minor of b3 is |
c' 19
|}42 C2

i.e. omit the row and column containing b3.

Now on to Frame 20

Evaluation of a third-order determinant 20

To expand a determinant of the third order, we can write down each element
along the top row, multiply it by its minor and give the terms a plus or minus
sign alternately:
ay b, Cy

az D3 G3
Then, of course, we already know how to expand a determinant of the second
order by multiplying diagonally, + \. — /

Example1

lee 3|4 | 2\° =

re ae 3| 2 8 2 4
2 4 8
=1(5x8—-4x
7)—3(4x 8-2x 7)+2(4x 4-2x5)
= 1(40 — 28) — 3(32 — 14) + 2(16 — 10)
= 1(12) — 3(18) + 2(6)
= 12 —544+12=
—-30

D1
Here is another.

Example2

J : a|° vf As rs S|. A
oa eae 2 2|*>|2 9
2 £9 2,
— 3(12 — 63) — 2(8 — 14) + 5(36 — 12)
= 3(—51) — 2(-6) + 5(24)
1593-12-00 21
Now here is one for you to do.
38 Linear Algebra

Example 3
PET EES
Evaluate 4A*6 3
Seon
Expand along the top row, multiply each element by its minor, and assign
alternate + and — signs to the products.
When you are ready, move on to Frame 22 for the result

22
Pa RE)
Because {4 6 3]= |
oe Sot|
= 2(6 — 27) — 7(4 — 24) + 5(36 — 48)
= 2(—21) — 7(—20) + 5(-—12)
= —42+ 140 — 60 = 38
We obtained the result above by expanding along the top row of the given
determinant. If we expand down the first column in the same way, still
assigning alternate + and — sighs to the products, we get:
22 L295

lo alle6 33
Cas Leas
4.4.6 3 = 2
he! Oat
Day et
= 2(6 — 27) — 4(7 — 45) + 8(21 — 30)
= 2.21) —4( 381 8(-9)
242 4159 72 38
which is the same result as that which we obtained before.

93 We can, if we wish, expand along any row or column in the same way,
multiplying each element by its minor, so long as we assign to each product
the appropriate + or — sign. The appropriate ‘place signs’ are given by
+ —- + - +

+ - + - +

Ct. <OLC.
The key element (in the top left-hand corner) is always +. The others are then
alternately + or —, as you proceed along any row or down any column.
So in the determinant:
‘te aens:
heey the place sien’ of the element.9 is. ops -ene
4720-3
Determinants 39

—7 Because in a third-order determinant, the ‘place signs’ are:

CO
ioe? Cag Remember that the top left-hand element always has a
ry a ee + place sign. The others follow from it.
+ — +
ry Ath”
Now consider this one: 6 8 4
i 9-5
If we expand down the middle column, we get

oe
3 es
SA 6 4 aa
ee esd eee
3.2

|I
Finish it off. Then move on for the result

~78 20
6) 4: 3) Sey
Because —/7 +8 _ q |
eS HOG 6 4
= —7(30 — 4) + 8(15 — 2) — 9(12 — 12)
= —7(26) + 8(13) — 9(0) = —182 + 104 = —78
So now you do this one:
Rew oa
Evaluate Set nee by expanding along the bottom row.
5 A 4
When you have done it, move to Frame 26 for the answer

119 26

ay Say
We have 6108 and remember
ee +141+ +1+

=5|ae
3 6 3
ee |
+2|6 11 |=8-4) - 76-24) + 202-18)
5(5) — 7(—18) + 2(-16) = 25 + 126 — 32 = 119
ive <a
One more: Evaluate ag TEP| by expanding along the middle row.
4 69
Result in Frame 27
ET ae ee en
40 ; Linear Algebra

27 | 143

1 2578

*3]a olla
Pda #3
Because ame | =-7 49 a te
679
A659
= —7(18 — 48) + 3(9 — 32) —1(6 — 8)
= —7(—30) + 3(—23) — 1(—2)
= 210 -69+2= 143

We have seen how we can use second-order determinants to solve

simultaneous equations in 2 unknowns.
We can now extend the method to solve simultaneous equations in
3 unknowns.
So move on to Frame 28

Simultaneous equations in three unknowns

28 Consider the equations:
ax+byy+coz+d,=0
a2X + bey + coz + do = ()

a3xX + b3y + c3Z + d3 = 0

4
If we find x, y and z by the elimination method, we obtain results that can be
expressed in determinant form thus:
Be 4 ey Z .¥ =!
byte dy|- are dy |} arb | | ay ty
bz C2 de a2 C2 de az bz do a2 bz C2
bz ¢3 43 a3 03 a3 a3 b3 d3 Tue Dani Ce
We can remember this more easily in this form:

hee een eae

By Dg Bia Lip
where A; = the det of the coefficients omitting the x-terms
Az = the det of the coefficients omitting the y-terms
A3 = the det of the coefficients omitting the z-terms
Ao = the det of the coefficients omitting the constant terms.
Notice that the signs are alternately plus and minus.
Let us work through a numerical example.
Determinants 41

Example 1

Find the value of x from the equations:

2x + 3y—z-—-4=0
3x+y+2z-—13=0
x+2y-—5z+11=0

First the key: inate Sonanlaal pete

mr og. Lg Ae
x ; :
To find the value of x, we use —- = —, i.e. we must find A, and Ao.
1
(a) To find Ap, omit the constant terms.

epee 1
ieee desl
2 2 |

igp2es - G
= -184+51-5=28
(b) Now you find Aj, in the same way.

Ay se86 29
3 -1 -4
Because ose Wh 2 -—13| =3(22 — 65) + 1(11 + 26) — 4(—S — 4)
2 -5 11
= 3(—43) + 1(37) — 4(-9)
= —129+
37+ 36
= —129+
73 = —S6
x Aston X HiStd
sete WatRAO wiectighe 1628
56
X= 58 x

Note that by this method we can evaluate any one of the variables, without
necessarily finding the others. Let us do another example.

Example 2
Find y, given that:
2x+y—5z+11=0
x-y+z-—6=0
4x + 2y —3z+8=0

First, the key, which is ..........-.

jnnnnnnnnn ee eat EEE dEUS SESS
 Linear Algebra
e

To find y, we use Fe=

Therefore, we must find A> and Ao.

2x+y—5z+11=0
The equations are x-y+z-6=0
4x + 2y —-3z+8=0
To find Az, omit the y-terms. ig
2 -5 11
EO abs Berar ant +1], _
ieee mas bev kes ees oes oe 3
4 -3 8
= 2(8 — 18) + 5(8 + 24) + 11(-3 — 4)
= —20+ 160 —77 = 63

To find Ap, omit the constant terms

9AN Ro aha me .

31 RE bi
4
Because
2 iL =§
oh ite | Li 1 -1
Agee dese beady beeed oy | zs
—3 4 -3 ay 2
482) ta3
= 2(3 — 2) — 1(-3 — 4) — 5(2 + 4)
=2+7-30=-21
2y owed ; A2_ 63
So we have AS? SL V= Bo 221

. y=-3
The important things to remember are:
: OY Re eee

with alternate + and — signs.

(b) To find A,, which is associated with x in this case, omitt the x-terms and
form a determinant with the remaining coefficients and constant terms.
Similarly for Az, A3, Ao.
Next frame
Determinants ORI

F| Review exercise
Here is a short review exercise on the work so far.
32
Find the following by the use of determinants:
x+2y—3z-—3=0
1 (areata Find
y.
3x+2y+z+5=0
3x —4y+2z7+8=0
2 x+5y—3z2+2=0 Find
x and z.
5x+ 3y—-zZ+6=0
2x —-2y —-z-3=0
3 4x + Sy —-2Z+3=0 } Find x, y and z.
3x + 4y —-3z7+7=0

When you have finished them all, check your answers

with those given in the next frame

Here are the answers: 33 |

1 y=-4
2 eke 7 5
a xe2 Vol *z=3
If you have them all correct, turn straight on to Frame $2.
If you have not got them all correct, it is well worth spending a few minutes
seeing where you may have gone astray, for one of the main applications of
determinants is in the solution of simultaneous equations.
If you made any slips, move to Frame 34

The answer to question 1 in the review exercise was 34

Did you get that one right? If so, move on straight away to Frame 41. If you
did not manage to get it right, let us work through it in detail.
x+2y-—3z-3=0
The equations were 2x-—y—-—z-—11=0
3x+2y+zZ+5=0
Copy them down on your paper so that we can refer to them as we go along.
The first thing, always, is to write down the key to the solutions. In this case:
x —
Ai pe ite) I

To fill in the missing terms, take each variable in turn, divide it by the
associated determinant, and include the appropriate sign.
So what do we get?
On to Frame 35
44 Linear Algebra

The signs go alternately + and —.

In this question, we have to find y, so we use the second and last terms in
the key.
See
pa Rareoge Ye
ee

So we have to find Az and Ao.

LOIN ty. We vane nseees

36 form a determinant of the coefficients omitting those of the y-terms |

L.-3 -3
So Ao = 2-1 --11
3 seek S
Expanding along the top row, this gives:
-1 -11 2 -I1 +(-3)/2 a
raeheatian I| 5 — (-3) cna ea |
We now evaluate each of these second-order determinants by the usual
process of multiplying diagonally, remembering the sign convention that
+ \,and—- / 4

50 We get Dig west:

37 Ay = 120

Because

Ap lt 54 LIVE B(LO 33) = 312 3)

= 6 + 3(43) — 3(5)
= 6 +129 =15 = 135-15
= 120
sy = 120
We also have to find Ao, i.e. the determinant of the coefficients omitting the
constant terms.

So ING lok

2
Determinants 45

Ao = -30 40
Because Ao = 1(—-1+ 2) — 2(2 + 3) — 3(44+ 3)
= 1(1) — 2(5) — 3(7) = 1 — 10 — 21 = —30
So Ao= —30

A2 120
oO we have y Ao ~30 4 4 4

Every one is done in the same way.

Did you get review exercise 2 correct?
If so, turn straight on to Frame SO.
If not, have another go at it, now that we have worked through 1 in detail.
When you have finished, move on to Frame 41

The answers to review exercise 2 were x=peak 2 41

Did you get those correct? If so, turn on right away to Frame SO. If not, follow
through the working. Here it is:
The equations were:
3x —4y+2z+8=0
x+S5y-—3z+2=0
5x+3y-—z+6=0
Copy them down on to your paper.

The key to the solutions is: aan

1
Fill in the missing terms and then move on to Frame 42
 Linear Algebra

We have to find x and z. .. We shall use:

x vo Ay
i ukaer ssinamenmely 7)

and ey hats ie 822

A3 Ao Ao
So we must find Aj, A3 and Ao.
(a) To find A,, form the determinant of coefficients omitting those of the
x-terms. =

43

Finish it off: then on to Frame 44

44
Because A; = —4(—18 + 2) — 2(30 — 6) + 8(-5 + 9)
= —4(—16) — 2(24) + 8(4)
= 64 — 48 + 32 = 96 — 48 = 48
aN = 48

(b) To find A3, form the determinant of coefficients omitting the z-terms.

% ao]
Determinants
47

45

Expanding this along the top row gives

On to Frame 47

47
Because A3 = 3(30 — 6) + 4(6 — 10) + 8(3 — 25)
= 3(24) + 4(—4) + 8(—22)
=72—16-176
= 72 — 192 = —120
.. 23 =-120

(c) Now we want to find Ao.

Now expand this along the top row as we have done before. Then evaluate the
second-order determinants which will appear and so find the value of Ao.
Work it right through: so that
 Linear Algebra

49
eo ieee 1 5
Because Age 3 |
+ i |
+ 2
out eel 5 :
3 (5440) seA( Le 15) 218 — 25)
= 304) 494( 14) 2222}
= 12+ 56 — 44
= 68 —44 = 24
NG = oe

So we have: Ag= 48, #&3=—120, Ag =24

Also we know that:

1 A3
= —— and z= —-—
0 Ao
SOlnatey —— eee BYvatel 4 = Pee

50

Well, there you are. The method is the same every time — but take care not to
make a slip with the signs.
Now what about review exercise 3. Did you get that right? If so, move on
straight away to Frame S52.
If not, have another go at it. Here are the equations again: copy them down
and then find x, y and z.
2x -—2y-—z—3=0
4x + S5y—2z+3=0
3x+ 4y —-3z+7=0
When you have finished this one, move on to the next frame and check your results

)
Determinants , — 3
iy =

Here are the main steps, so that you can check your own working:
Mie Vis 2 —1
Beis ha As
—2 -1 -3 2-1 -3
Bie) Ss —-2°.3 |= 54 As=14 —2 . 3 |=27
4 -3 7 a =a OF

2 -2 -3 2 —2 -1
Apt! 41 eS St 81 Bp =14.05 21-27

3. 4 7 3 4 -3
x 1 A
Simba =—7 mere ay
Eee
A2 Ao
Baie
Ag —27
as aie
Z 1 A3 81
Grn Ke Sy oe tO z=+3

All correct now?

On to Frame 52 then for the next section of the work

Consistency of a set of equations

Let us consider the following three equations in two unknowns. 52
3x-—y-4=0 (a)
2x+ 3y-—-8=0 (b)
x-2y+3=0 (c)

If we solve equations (b) and (c) in the usual way, we find that x = 1 and y = 2.
If we now substitute these values in the left-hand side of (a), we obtain
3x -y—-4=3-—2-—4=-3 (and not 0 as the equation states).
The solutions of (b) and (c) do not satisfy (a) and the three given equations
do not have a common solution. They are thus not consistent. There are no
values of x and y which satisfy all three equations.
If equations-are consistent, they have a <...<.2..080".
5h \eee
50 Linear Algebra

53
Let us now consider the three equations:
3x+y—5S=0 (a)
2x + 3y —8=90 (b)
x-—2y+3=0 (c)

The solutions of (b) and (c) are, as before, x = 1 and y = 2. Substituting these
in (a) gives:
3x+y—-5=342-5=0
i.e. all three equations have the common solution x=1, y=2 and the
equations are said to be............

54
Now we will take the general case:
a,x+by+d,=0 (a)
Agx + boy +d2,=0 (b)
a3X-- b3y + d3=0 (c)

If we solve equations (b) and (c):

pela that BR 8 x -y 1
e. we set —— = = ——
€3X + b3y+ dz =0 Ai A2 Ao

bo d> a2 do a2 bo
where Fa ge », A= , KK =
bz d3 a3 3 a3 bz
ipiAy ro uag
so that arn EOE Earae:

: : Ai —A2
If these results also satisfy equation (a), then a + a a +dy=0
0 0
i.e: ay;.A, = by.A2 ar d;.4o0 ea)

bo d a2 dz| a2 b
i.e. ay g : — b, 3 3 + dy 4 a =)
bz d3 a3:id3 a3 b3

aq by ay
i.e. a2 b2 do =a)

a3 bz d3

which is therefore the condition that the three given equations are consistent.
So three simultaneous equations in two unknowns are consistent if the
determinant of coeilicients:152 a2. <..5..

2
Determinants 51

55
Example 1

Test for consistency:

2 1 -S5
For the equations to be consistent |1 4 1) must be zero
3 -1 -10

oie eee 4 1 ie at
eer
cp) Pst)ee. =| ee 1 ae ;
. aolaie1
eo eAtae ye 11103) 5 (112)
= 2-39)\e-(13) —~5(—13)
== 79134 6h —78 +78 =0
The given equations therefore ............ consistent.
(are/are not)

are 56

Example 2

Find the value of k for which the following equations are consistent:
3x+yv+2=0 Pn bn oe
4x+2y—k=0 For consistency: |4 2 -—k|=0
2x —-y+3k=0 2 -1 3k
2 —k 4 -k 4: 2
Zz =0
if:
ne ind—1 é elt bg Pa
3(6k — k) — 1(12k + 2k) + 2(—4 — 4) =0
* 15k-14k-16=0 .. k-—16=0 .. k=16
Now one for you, done in just the same way.

Example 3

x+(k+1)y+1=0
Given 2kx+ Sy—3=0
3x+7y+1=0

find the values of k for which the equations are consistent.

enn
Nee eeenn —————————E———eeeeeee_
4 7S el ag Meg Linear Algebra

ss |

The condition for consistency is that:

1, ck ee
ok es eee ea)
3 7 41
ees Bat tas
a (Ra4 1 a
P p|-«+n | e |
(5-221) — (k+ 1)Ck+ 9) + (14k — 15) =0
26 —2k* —11k—9+4
14k -—15=0
SO 3k 20
1 2h 23K
2= 0... (2K+1)(k—2)=0
1
2k S77 Onn =

Finally, one more for you to do.

Example 4
x+y-k=0
Find the values of k for consistency when <¢ kx —3y+11=0
2x+4y—-8=0
&

|58 k=lor
1
— 3

Because
1 1 -k
3, eT = 2)
2 4 -8
Ast! 11 1% 11 lk Fe a 6
4 -8 2 -8 2. AR
", (24 — 44) — (-8k — 22) — k(4k+ 6) =0
= 20 + 8k+-22 = 4k — 6k=0
—4k? + 2k+2=0
"2? —-k-1=0. .. (2k+1)(kK>1)=0
1
tee kal ork = 5

2
Determinants
53

Properties of determinants
Expanding a determinant in which the elements are large numbers can be a 59 |
very tedious affair. It is possible, however, by knowing something of the
properties of determinants, to simplify the working. So here are some of the
main properties. Make a note of them in your record book for future reference.
1 The value of a determinant remains unchanged if rows are changed to columns
and columns to rows.
a; a2|_|aq, by
by b * a2 bo

2 If two rows (or two columns) are interchanged, the sign of the determinant is
changed.

az bz} _ ja, by
ay b, x" a2 bo

3 If two rows (or two columns) are identical, the value of the determinant is zero.
ay ay
a2 a2

4 If the elements of any one row (or column) are all multiplied by a common factor,
the determinant is multiplied by that factor.
kay kbo a, by
a2 b» a2 bo
5 If the elements of any row (or column) are increased (or decreased) by equal
multiples of the corresponding elements of any other row (or column), the value of
the determinant is unchanged.
a,+kb, b, a, by ay by a, b,
a2+kbz bz a2 be a2 +ka, b2+kb, a2 bp

Note: The properties stated above are general and apply not only to second-
order determinants but to determinants of any order.
Move on for some examples
 Linear Algebra

60 Example 1

Evaluate
427 429
2A
Of course we could evaluate this by the usual method
(427)(371) — (369)(429)
which is rather deadly! On the other hand, we could apply our knowledge of
the properties of determinants, thus:

427 429| _ 427 429 — 427 | (Rule 5) Subtract column 1 from

369 371 ~ |369 371 — 369 column 2}

427 2| >
369 Z
58 0
3 | (Rule 5) Subtract row 2 from row 1
369
= (58)(2) — (0) = 116
Naturally, the more zero elements we can arrange, the better.

For another example, move to Frame 61

61 Example 2

Evaluate

Column 2 minus column 3 will give us one zero

Column 3 minus twice(column 1) will give another zero

N
| n

Now expand along the top row

SP
PR
PPR
PR

| | w We could take a factor (—1) from the top row

NO
| and another factor (—1) from the bottom row
(rule 4)
2°23
=(-1(-0| Samak
= 1(2—15) =-13
Next frame
Example 3 62

Bae De Te,
Evaluate 2 £02
2 2 24

You do that one, but by way of practice, apply as many of the listed properties
as possible. It is quite fun.
When you have finished it, move on to Frame 63

The answer is [32], but what we are more interested in is the method of
applying the properties, so follow it through. This is one way of doing it; not
the only way by any means.
Bez
aa) We can take out a factor 2 from each row, giving
a factor 2°, i.e. 8 outside the determinant
2a02, 74
RRA a |
Column 2 minus column 3 will give one zero in
= ro i Dene ae|
the top row
1S Lez,
22) —]
Column 1 minus twice(column 3) will give
= §)1 Led ;
another zero in the same row
1 -1 2
0 ea fe
cgay ae 4 Expanding along the top row will now reduce
og - this to a second-order determinant
-3 -1 2

—1 1
= 8 Now row 2 + row 1
-3 -l

-1 1
= 8
—4 0

oe
= = = —8(—4) = 32
4 0
ee
56 Linear Algebra

64 Here is another type of problem.

Example 4
x S >
Solve the equation |S x+1 tel
=3 =—4 x=2
In this type of question, we try to establish common factors wherever possible.
For example, if we add row 2 and row 3 to row 1, we get:
(x+2) (x+2) (x+2)
5 x+1 1 =Q)
—o ay y=)

Taking out the common factor (x + 2) gives:

mic Sol 1
(x+2)['S8o41- “11 S0
oti eae 2
Now if we take column 1 from column 2 and also from column 3, what do
we get?
When you have done it, move on to the next frame

65 1 0 0
We now have (x+2)} 5 x-4 -4 |=0
-3 -1 x+1
Expanding along the top row reduces this to a second-order determinant:
x-4 -4
= 0
(A+2) 507 4d
If we now multiply out the determinant, we get
(x + 2)[(x — 4)(x+ 1) —4] =0
=. (%+2)@7 —3x—8) =0
“stb 2 0.or xX — 3X58 = 0
which finally gives x = —2 or x = a

Finally, here is one for you to do on your own.

Example 5

Solve the equation:

5 se 8
NaS? 222 lO
ee eX.

Check your working with that given in the next frame

@
Determinants

x=-40r1+V6

Here is one way of doing the problem:

2 x3
x+2 2 11=0 Adding row 2 and row 3 to
jee ~~ row 1, gives
= x
X+4 x4+4 x44
x4+2 2 1 |=0 Taking out the common
3 9 » eedactor (x4)
=4 x

a ake column
Tak ] 33 ffrom column
|
4 —
oe ae : ‘ . 1 and from column 2
= x
0 0 1 ie
Ged ruxtil 1 ‘bean is now reduces to second
order
—-x—-3 2-x x

(x +4) x+1 1 | | Subtract column 2 from

Y= 22 — x |) CONT

(x + 4) 1 |. We now finish it off

—§ 2-x|
(Xk -4)(2x—
x" +5) = 0
X+4=0orx?-2x-5=0
which gives x = —4 orx =1+ V6

You have now reached the end of this Program on determinants except for the
Can You? checklist and Test exercise which follow. Before you work through
them, brush up on any parts of the work about which you are at all uncertain.
 Peete Linear Algebra

(4 Can You?
|67 Checklist2 )
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames
e Expand a 2 x 2 determinant? to (6)
Yes ab ieee! seal searilzl No
e Solve pairs of simultaneous linear equations in two variables
using 2 x 2 determinants? to
mS rae AS a [| No 2
e Expand a 3 x 3 determinant? to
Yes Se be EBD gtSie SANS, No
e Solve three simultaneous linear equations in three variables
using 3 x 3 determinants? to (1)
Yes iF} CJ LJ LJ CJ No
e Determine the consistency of sets of simultaneous linear
equations? G2) to
:<2 epg Ba og ekeared ODaaa Pe wR IN
e Use the properties of determinants to solve equations written
in determinant form? to
Yes O aa A C] OC No

Test exercise 2
|68 _ !f you have worked steadily through the Program, you should have no difficulty
with this exercise. Take your time and work carefully. There is no extra credit for
speed. Off you go then. They are all quite straightforward.
1 Evaluate
it gale ees
(ayes sl (bDyerae Le
a 2k te ah owal|

2 By determinants, find the value of x, given:

2x>+ 3y—Z2+3=0
x—4y+2z-—14=0
4x + 2y —3z+6=0

fas
Determinants J = Stes ees

3 Use determinants to solve completely:

x-—3y+4z-S5=0
2x+y+z-3=0
4x + 3y+5z-1=0
4 Find the values of k for which the following equations are consistent:
3x + Sy+k=0
2x+y-5=0
(k+1)x+2y-—10=0
5 Solve the equation:
X+1 -S5 -6
—1 x 2 =0
—3 2 x+1
Now you can continue with the next Program

4 Further problems 2
ae SG 1 428 861 69
1 Evaluate: eNO ta iad * en (b) |2 S35 984
pW yewsWhee (2) 3 642 1107
dy Neat aero) 155 226 81
2 Evaluate: (a) |16 10 —18 (ob) | 77 11Z 39
34 6 38 74 111 37

3 Solve by determinants:
4x —-Sy+7z=-14
9x + 2y +3z=47
x-y-Sz=11
4 Use determinants to solve the equations:
4x —-3y+2z=-7
6x + 2y — 3z = 33
2x -4y-z=-3
5 Solve by determinants:
3x + 2y — 2z = 16
4x+3y+3z=2
2x-y+z=0

6 Find the values of \ for which the following equations are consistent:
5x+(A+1)y—S=0
(A—1)x+7y+5=0
3x+Sy+1=0 >
 Linear Algebra

Determine the values of k for which the following equations have solutions:
4x —(k—2)y—S=0
2x-+y—-10=0
(k+1)x-—4y-9=0
(a) Find the values of k which satisfy the equation:
K> Y=
1k olie0
Os® Leak
(b) Factorize
‘Lee SLR
oh Ya) ie Bey
gait bBvec3

Solve the equation:

x 2 3
De Xx 3 oy | 0
3 4 x+6

10
x 34+x 24+%x
3 -3 —1 |=0
2 -—2 —2

11 Express
1 1% 1
a b? ce
(b+c)? (c+a)> (a+b)
as a product of linear factors.
12 A resistive network gives the following equations:
2(i3 = iz) = 5(i3 = i) == 4

(iz _ iz) + 2i2z + (iz _ i) =,()

S(i1 — 13) + 2(i — in) +i,=6

Simplify the equations and use determinants to find the value of iz correct to
two significant figures.
13 Show that (a+b-+c) is a factor of the determinant
bica @
Cta ip DP
Buby c =
and express the determinant as a product of five factors.
Determinants
61

14 Find values of k for which the following equations are consistent:

x+(1+k)y+1=0
(2+k)x+5y—-10=0
4474-9 =0

1+x* yz 1
1S Express|1+y? zx 1| as a product of four linear factors.
bie ayy 1
peeAe Seay 3
16 Solve the equation 2 ne, pony == ()
x+3 1 x+2

17 If x, y, z satisfy the equations:

1
(sm +Ma))x —-M2y =W
Z
—M2x + 2Moy + (M, — M2)z = 0
1 * Mn)z =0
—Moy + (5M
evaluate x in terms of W, M; and Mp.

18 Three currents, i, iz, iz, in a network are related by the following equations:

2h + 312 + 8i3 = 30
61, —i2 + 213 = 4
31, — 121g + 8i3 = 0
By the use of determinants, find the value of i; and hence solve completely the
three equations.
iL Saat: k(x —a)+2x—z=0
k(y-—a)+2y—z=0
k(z-—a) -x-y+2z=0
ak(k + 3)
show that x =
k2 + 4k+20
20 Find the angles between @ = 0 and 6 = x that satisfy the equation:
1+sin?@ cos?6 4sin 20
sin?@ 1+cos*@ 4sin26 |=0
sin? 0 cos?6 1+4:sin26
Matrices
Learning outcomes .
When you have completed this Program you will be able to:
Define a matrix
Understand what is meant by the equality of two matrices
Add and subtract two matrices
Multiply a matrix by a scalar and multiply two matrices together
Obtain the transpose of a matrix
Recognize special types of matrix
Obtain the determinant, cofactors and adjoint of a square matrix
Obtain the inverse of a non-singular matrix
Use matrices to solve a set of linear equations using inverse matrices
Use the Gaussian elimination method to solve a set of linear equations
Matrices . ic 63

Matrices — definitions
A matrix is a set of real or complex numbers (or elements) arrangedinrowsand 4
columns to form a rectangular array.
A matrix having m rows and n columns is called an m x n (i.e. ‘m by n’)
matrix and is referred to as having order m x n.
A matrix is indicated by writing the array within brackets
ae?
C8: 2 3 ;) is a 2 x 3 matrix, i.e. a ‘2 by 3’ matrix, where

5, 7, 2, 6, 3, 8 are the elements of the matrix.

Note that, in describing the matrix, the number of rows is stated first and the
number of columns second.
5 Ome:
2135 23| , F
7 g 7 | isa matrix of order 4 x 3, i.e. 4 rows and 3 columns.
6 und

6 4
Spotne. matrix |'O.° Wis Of Ofder 7..+. yesh. ss
23

and the matrix ( ; : f Cle) ae) 1 6(3 ee eee gee

3x2; 2x4 2

A matrix is simply an array of numbers: there is no arithmetical connection

between the elements and it therefore differs from a determinant in that the
elements cannot be multiplied together in any way to find a numerical value
of the matrix. A matrix has no numerical value. Also, in general, rows and
columns cannot be interchanged as was the case with determinants.
Row matrix: A row matrix consists of 1 row only.
e.g.(4 3 7 2) is a row matrix of order 1 x 4.
Column matrix: A column matrix consists of 1 column only.
6
e.g. | 3 ] is a column matrix of order 3 x 1.
8
To conserve space in printing, a column matrix is sometimes written on
one line but with ‘curly’ brackets, e.g. { 6 3 8 } is the same column matrix of
order 3 x 1.
Move on to the next frame
 nN parva ;
@ (3) isa FO as et matrix of order... 0.0...
(i) (4.0) 73) WA. Ke matrix of order... .....<..
Rm € BBs... 14 matrix of order... .........

. (a) column, 2x 1 (b>) row, 1 x 4 (2 column, 3 x 1 |

He eA intact, allen Bicone imitate ea ee

We use a Simple row matrix in stating the x and coordinates of 2 point

relative to the x- and »-axes. For example, if P is the point (¢, 9) then the 3 8
the x-coordinate and the 5 the »-coordinate, In matrices generally, however,
no commas are used to separate the elements,
Single element matrix: A single number may be regarded as a 1 x 1 matrix, Le.
having 1 row and 1 column.
Double suffix notation: Each element in a matrix has its Own particular ‘adkdiress”
or location which can be defined by a system of double suffixes, the fimt
indicating the row, the second the column, thus:
Q, 22 & Bu
@2n @m 23 GH
@, @3> B33 Gy
@2; indicates the element in the second row and third column.
Therefore,
in the matrix
6-5 1 -3
2-4 8 3
: 47-6 5
—2 9 7-1

the locationof (a) the element 3 can be stated as 2.2... 5...

(b) the element — 1 can be stated as —. 2.2.02...
(c) the element 9 can be stated as ~.... 20.6...

|5 @ an Oe
Mowe eet

a
a
Matrices

Matrix notation
Where there is no ambiguity, a whole matrix can be denoted by a single
general element enclosed in brackets, or by a single letter printed in bold type.
This is a very neat shorthand and saves much space and writing. For example:

411 412 413 414

421 422 423 424 | can be denoted by (aj) or (a) or by A.
A431 432 433 434

x}
Similarly (:)can be denoted by (x;) or (x) or simply by x.
x3
For an (m x n) matrix, we use a bold capital letter, e.g. A. For a row or column
matrix, we use a lower-case bold letter, e.g. x. (In handwritten work, we can
indicate bold-face type by a wavy line placed under the letter, e.g. A or x.)
So, if B represents a 2 x 3 matrix, write out the elements bj in the matrix,
using the double suffix notation. This gives ............

B= (pi by2 ey 7
bo; b22_— —b23

Next frame

Equal matrices
By definition, two matrices are said to be equal if corresponding elements 8 |
throughout are equal. Thus, the two matrices must also be of the same order.
4 Ai, ay2 a3 ee enters
pout aa 22 as |= (2 3 3)
then.d31,.=4;. d12 = 6; ..413.=5;. a2 =2,;,.. etc.

Therefore, if (aj) = (xij) then aj = xy for all values of i and j.

fe Wale Ge 5 —J as
po9 an? ea fl 26
&§ h k 0 4 8

TROT96 oases
xptte 7 ee eee ae nd ee ee ee
66 Linear Algebra

Addition and subtraction of matrices

|1 8) To be added or subtracted, two matrices must be of the same order. The sum or
difference is then determined by adding or subtracting corresponding
elements.

RAD Amey SA Ar booked

FB blast andes Bors Dacha 7 was Ete 4
vi /(5a 10 ae
Ne 172710

65°. 12 Sl A oa 6-3 5e7 Bite)

and _ =
(; 4 ig (s: 10 “es 6. EN EES
= (Song)
bab Girea

aha: ah Le eae 5
So,(a) ( 37 s)+(¢ iF on he ed hesa. wietenes

oud! G 1-323
(ees ‘|-(s Ofte ee
120 4 7 889

11 a (2 a
2 -7

Multiplication of matrices

|12 1 Scalar multiplication

To multiply a matrix by a single number (i.e. a scalar), each individual element
of the matrix is multiplied by that factor:
Pee Dp UP
8: een ey, 24
4 28
i.e. in general, k(aj;) = (kaj).
It also means that, in reverse, we can take a common factor out of every
element — not just one row or one column as in determinants.
10 25 45
Therefore, (35 15 50 )Can. Deswritten=aeaemestycse:
Matrices
67

sx (7 3 10) es
2 Multiplication of two matrices
Two matrices can be multiplied together only when the number of columns in
the first is equal to the number of rows in the second.

by
a4. a2 a
e.g. if A= (ay) = ( + en a and b = (bj)= | bo
421 422 423
b3
to a See by
then A-b= igs ay2 aa b>
421 422 423 b3

5 (ee + ay2b2 + pa)

d21b, + d22b2 + a23b3
i.e. each element in the top row of A is multiplied by the corresponding
element in the first column of b and the products added. Similarly, the second
row of the product is found by multiplying each element in the second row of
A by the corresponding element in the first column of b.

Example1
8
AY ATG ole! gen aiid D1 uel np one
ve ate WARE SA ER Rae oe a Mes Pees ea

ene Eas etngeta|

Similarly 4 :
Oo ae mee
FW
ON

ea ead ie 14
12+ 24+0+ 63

(ir)17 .
The same process is carried out for each row and column. b
 Linear Algebra

Example 2

|
pal Mgeds
igo Rsos
lf A= (ay) =" 2a 7 and B= (bi) = (
DD eo
3 4

|
he 53
aes.)
then @A= B=) ie7e.7
2o SeSai6
a we

lees Sisco mle 45 bole aoe

Oe Gumle elt
DSCAGE Pau OO axa Xe ©
BROS 424% 2s ABE 4 93 5 4 x8 3x 14<6

|
8+10 4425 3+40 1+30
=| 164+14 84+35 6+56 2+42
244+8 12+20 9432 3424
Reo AS oe: Nee eH
= | 30 43 «G2 44
pence 41e 27
Note that multiplying a (3 x 2) matrix and a (2 x 4) matrix gives a product
matrix of order (3 x 4)
i.e. order (3x2) x order (2 x 4) — order (3 x 4).

(same)
In general then, the product of an (/ x m) matrix and an (m x n) matrix has
order (Ixn). &
aL
2 6
lf A= dB=/]-2 9
3 5
4 3
thenA: B= a5 e.teacesad

16

2V4 6 as
since A-B= ( ) —2 9
54-975
4 3
a eae 2+36+18\ (30 56
T2118 20ers 8 lepdSis I NG in99
Matrices

Example 3

It follows that a matrix can be squared only if it is itself a square matrix, i.e. the

tan($ 2)
number of rows equals the number of columns.

ee 2) (30038 28+14\ (S51 42

PA 2010 cave ACT =)
Remember that multiplication of matrices is defined only when

the number of columns in the first 17

= the number of rows in the second

That is correct. ¥ . Blin “ A has no meaning.

ABDUL Sy 1
If A is an (m x n) matrix
snihl cantar }th en products
d an
A-BandB-A are possible
ible.

Example
7 10
IA=(4 : 3)aman (5 1)
9 12
, de pest fA 10
then A B=(, 5 a) oe a
9 12
7+16+27 10+22+36\ / SO 68
<A ponent a
Se panes
andB-A=|8 11 Gi 5 >)
9 12
7440 14+50 21+60 47 64 81
—~ 18444 16455 24+66|=| 52 71 90
9+48 18460 27+72 57 78 99
Note that, in matrix multiplication, A. B # B.A, i.e. multiplication is not
commutative. The order of the factors is important!
In the product A- B, B is pre-multiplied by A
and_ A is post-multiplied by B.
70 Linear Algebra

shee
‘ 9 2 4
SOn eA Cs et and B = ( )
Se]
thenzA= B= ee andi BSA eee

18 AP 16. 32
A-B= {155 26 SZ]; BA=(Jo )
2a: opte 1S

Transpose of a matrix
19 If the rows and columns of a matrix are interchanged:
i.e. the first row becomes the first column,
the second row becomes the second column,
the third row becomes the third column, etc.,

then the new matrix so formed is called the transpose of the original matrix. If
A is the original matrix, its transpose is denoted by A or A’. We shall use the
latter.
4 6
if Tfiet{ 4a 2
was(7 »)een at(§ 9
bo
in
Ae ee

Therefore, given that

2
Matrices ” 79;

Special matrices
(a) Square matrix is a matrix of order m x m. 21
|
e.g. | 6 8 9] isa3 x3 matrix
7s 4
1
A square matrix (aj) is symmetric if ay = aj, e.g. | 2
~; Ww
©
oOome
OWN

i.e. it is symmetrical about the leading diagonal.

Note that A= A’.
Die eae oo
A square matrix (aj) is skew-symmetric if aj = —aje.g. | -2 O 9
-5 -9 0
In that case, A= —A!.
(b) Diagonal matrix is a square matrix with all elements zero except those on
a)
the leading diagonal, thus | 0 2 O
0 0 7
(c) Unit matrix is a diagonal matrix in which the elements on the leading
On 0
diagonal are all unity, i.e. | 0 1 O
0 0 1
The unit matrix is denoted by I.

Ss 2, 4 1 0 0
IfA={1T"s 8] andi= 10 YF 0] then A-I=............
i Ie «' 28 Oia |

22
72 Linear Algebra

Similarly, if we form the product I- A we obtain:

LoD 0 5. Cnatth
L2A-=-|309 107) = 10 ao
05a) Pee ae:
5+0+0 2+0+0 4+0+0 S
=|041+0 0+3+0 04+8+0]=] 1
0+0+7 04+0+9 0+0+6 7
A-IT=I1-A=A
Therefore, the unit matrix I behaves very much like the unit factor in ordinary
algebra and arithmetic.
(d) Null matrix: A null matrix is one whose elements are all zero.
OO. *O
ie. |O O O | and is denoted
by 0.
On G
If A- B = 0, we cannot say that therefore
A= 0 or B= 0

Zi hicss ad 2
for if A= ~~ \andB=|4 -6
6 3 -9
: 2), 4

Pao
chen en Ayan
i ah unis
--
6.3 -9
2 4
= 2+4-6 Me te? | 0 0
ONG tal | G54
1s = 36). 09.0
That is, A- B= 90, but clearly
AA 0 and B90.

Now a short revision exercise. Do these without looking back.

A Gans 7 OF 8-3 4
t a= (5 19 1) and B= (5 ee a
determine (a) A+ B and (b) A—B.
4 3
9? ifA= 2a, 7)\ andBad? 7ts
4086
determine (a) 5A; (b) A- B; (c) B-A.
DaeG aa |
See ieAG—n eon aeanc)
si— se Ome aa) aChlelie
AG > — ee
4 1 Ze
A822 6
4 Given that A = ({ 8 7 )determine (a) A’ and (b) A- A’.

When you have completed them, check your results with the next frame
Matrices 73

Here are the solutions. Check your results. 23

1 ds ead 3 uta) eres Ae ee ee
wie (5 Syed 10 Ji) A B=( > heats 5)
2.15 oo so Se
a ea Saat O 35 CD) A-B= {38 18 @) ce)
ou 34 54 20

Za0 a a2
3 A-B= [|4 («)is not possible since the number of columns in
4 1 2 3
the first must be equal to the number of rows in the second.
4 1
4 A= & 3 4 Al= 1298
6 7
art
a-at= (4 2 *\. naa A fogabh 4+16+42
ge ays Pa Nt 216 AD Paty,

Now move on to the next frame

Determinant of a square matrix

The determinant of a square matrix is the determinant having the same 24 |
elements as those of the matrix. For example:
Oo. eF Gl sh th 4
the determinant of | 0 6 3] is|0O 6 3) and the value of this
Sio4, 27 S407

determinant is 5(42 — 12) — 2(0 — 24) + 1(0 — 48)

= 5(30) — 2(-24) + 1(-48) = 150 + 48 — 48 = 150
5 0 8
Note that the transpose of the matrix is ( 6 ; and the
1huis maa d

eee
determinant of the transpose is |2 6 4) the value of which is
LS SF
5(42 — 12) — 0(14 — 4) + 8(6 — 6) = S(30) = 150.
That is, the determinant of a square matrix has the same value as that of the
determinant of the transposed matrix.
A matrix whose determinant is zero is called a singular matrix. »
SAX: ue Linear Algebra

3 a2
The determinant of the matrix . mn ) Tras (he Value eae ce
F 1 $26 200
and the determinant value of the diagonal matrix | 0 S 0 | has the
00 4
valué...s.<s6a9e

25
= 3(—30) — 2(15) + 5(25) =5

= 2(20) +0+0 = 40

Cofactors
If A = (aj) is a square matrix, we can form a determinant of its elements:

411 412 @43.—=«w-» SsGin Each element gives rise to a cofactor, which is
Q@21 422 423 .-- Gan ‘ simply the minor of the element in the
431 432 433 «-- G3n determinant together with its ‘place sign’,
: : : which was described in detail in the previous
Gni Q@n2 4n3_--- Ann program.

2h tau
For example, thé determinant of the matrix A = [|1 | is
1 —~4--0
ROSA IS
det A= |A|=/4 1 6} which has a value of 45.
Le 4.0

The minor of the element 2 is 4 : = (0 — 24 = —24.

The place sign is +. Therefore the cofactor of the element 2 is +(—24) i.e. —24.
Similarly, the minor of the element 3 is (es
10 5) =0-6=-6.

The place sign is —. Therefore the cofactor of the element 3 is —(—6) = 6.

In each case the minor is found by striking out the line and column
containing the element in question and forming a determinant of the
remaining elements. The appropriate place signs are given by:
+ - + - 2...
ae ie alternate plus and minus from the top left-hand
ey corner which carries a +
Matrices ae a te

Therefore, in the example above, the minor of the element 6 is : :

i.e. 8—3=5S. The place sign is —. Therefore the cofactor of the element 6
is —5.
oe Wee
So, for the matrix |6 5 4 |, the cofactor of the element 3 is............
FikSs 6S
and that of the element 4 is

Cofactor of 3 is 4 — (—10) = 14
Cofactor of 4 is —(56 — 3) = —53

Adjoint of a square matrix

Papas mast
If we start afresh with A= | 4 1 66|], its determinant
Ley? t
a ha,
det A=|A|=|4 1 6] from which we can form a new matrix C of
Ly 4is0
the cofactors.
A A A
ae rs 4s oe tes Aj, is the cofactor of a);
er aie a Aj is the cofactor of aj etc.
A31 A32 A33

On = — 24 )=- 24
ye Plasto 0thee
\., A2 : =
he Olt (0O—6)=6
=

4 1
ABn=t
13 |; | =+(16-—1)=15
( )

a 7! |
A21 == y | = —(0
(0 — 20)
) = 20 ( )
A22 22 = + F 0 = +(0-5)=-S

yA.
Ao323 = — |: | = —(8
( —3)=-S
)

Anat|) g[et08-9=18
oy eS
Aw=-[,2 ||= -02-20)
An=+|, “ = +(2—12) = -10
—-24 6 15
The matrix of cofactors is C = 20 -S -S
13 8 -10
—-24 20 13

and the transpose of C, i.e. C' = 6-5 8

15 -S -10
This is called the adjoint of the original matrix A and is written adj A.
76 Linear Algebra

Therefore, to find the adjoint of a square matrix A:

(a) we form the matrix C of cofactors,
(b) we write the transpose of C, i.e. C!.
ae aL
Hencethe adioint.ot. | tour Lee eis eee
40 3

27

Inverse of a square matrix

28 The adjoint of a square matrix is important, since it enables us to form the
inverse of the matrix. If each element of the adjoint of A is divided by the
value of the determinant of A, i.e. |A]|, (provided |A| 4 0), the resulting matrix
is called the inverse of A and is denoted by A“.
Zh
For the matrix which we used in the last frame, A= [:1 |
145
g..3 5
det A=|AJ=M 1 6|=2(0-24) -3(0-6)+5(16—1) =45,
lato
—24 6 iS
the matrix of cofactors C = 20 =5. =5
13 8 -—10

—24 20 13
and the adjoint of A, i.e. C' = 6 =5 8
15 -—5 -10
Then the inverse of A is given by

24 20 18
45 pp (49-5 4S 24205) 13
Es 6 5 8 1
A= = = Coe)
45 45 45 45
15 5 10 15 -—5 —-10
45 745. 5

)
Matrices Re aes

Therefore, to form the inverse of a square matrix A:

(a) Evaluate the determinant of A, i.e. |A|
(b) Form a matrix C of the cofactors of the elements of |A)
(c) Write the transpose of C, i.e. C', to obtain the adjoint of A
(d) Divide each element of C! by |A!
(e) The resulting matrix is the inverse A~! of the original matrix A.

Let us work through an example in detail:

12S
To find the inverse of A= |4 1 5
Ga a,

(a) Evaluate the determinant of A, i.e. |Al. |A| =

|Al= 28 | 29
Because

[a2 @
JAJ={4 1 5} =1(2-0)-2(8—30)+3(0-6)
=28
gaa
(b) Now form the matrix of the cofactors. C=............

30

Because
Au =+(2-0)=2; Ai=—-(8-30)=22; Az =+(0-6)
An = —(4-0) =-4; Az2=+(2—18)=-16; Az3 =—(O-—12)=12
A31 =+(10-—3)=7; A32 = —(S — 12) =7; A33 = +(1 — 8)

(c) Next we have to write down the transpose of C to obtain the adjoint of A.

31

(d) Finally, we divide the elements of adj A by the value of |Aj, i.e. 28, to arrive
at A~!, the inverse of A.
78 Linear Algebra

32

Every one is done in the same way. Work the next one right through on your
own.

33

Here are the details:

Di Fone4
det A=|A|=|3 1 6|=2(8) —7(-6)+4(-5)
= 38
S00 rk
Cofactors: &
Ai = +(8 —0) =8; Ai2 = —(24—30)=6; Ay3=+(0—5)=-5
Az = —(56—-0) = —56; Az2 = +(16—20)=-—4; Az3 = —(0— 35) = 35
A31 = +(42—4)=38; A3.=-(12—12)=0; A33=+(2—21)=-19
ee eee EOS et.
Ce | =56. —4 > 35 hy 2ayGl= 6 LA 0
Scan O19 Samo 19
Sie SG oe a8
then A == paeeed 0
=a 1 35. <9
Now let us find some uses for the inverse.
Matrices 79

Product of a square matrix and its inverse

—6 12 —-7 ao FZ
i 2—16+442 4-—4+0 6—20+ 14
=5,| 22-64+42 44-16+0 66-—80+14
—-6+48—42 -124+12+0 -18+60-14
osx O.. 0 1 0 O

Finish it off

1 0 0 34
om Pet
Omit

AcAT = Aa A=]
That is, the product of a square matrix and its inverse, in whatever order the
factors are written, is the unit matrix of the same matrix order.

Solution of a set of linear equations

Consider the set of linear equations:
35
QyiXy + AigX%o + Gigx3 + ..s00.0s- + AinXn = by
91X11 + Go2X2 + G23X3 + ......... + 42nXn = b2

AniX1 + GAn2X2 + An3X3 + oS & Bleue, & 60 + AnnXn = by -

 Linear Algebra

From our knowledge of matrix multiplication, this can be written in matrix

form:
Gy" Qo 'Gi3" >. in x by
A21 @22 423 «.-- Gon X2 bo
: oe ie.A-x=b

Ani Gn2 4n3 --- Ann Xn bn

ay. 412 Ain X1 by
d21 422 d2n X2 bo
where A = ; ; x= | “and b=

Ani G@n2 --- Ann Xn bn

If we multiply both sides of the matrix equation by the inverse of A, we have:

A!-A-x=A!-b
ButA7!-A=I .. I-x=A™-b ie. x=A™?'-b
Therefore, if we form the inverse of the matrix of coefficients and pre-multiply
matrix b by it, we shall determine the matrix of the solutions of x.

Example

To solve the set of equations:

X1 + 2x2 +%x3=4
3X1 — 4X2 — 2x3 =2

5X; + 3x2 + 5x3 = —-1

%
First write the set of equations in matrix form, which gives ............

ie. A-x=b “. x=A!l-b

So the next step is to find the inverse of A where A is the matrix of the
coefficients of x. We have already seen how to determine the inverse of a
matrix/so in this case. A; =... se
Matrices _
Pe Bt;

37

Because

1 Yi 1
IAJ=|3 -4 -2|=-14-504+29=29-64 .. [A] = —35
5 3 5

Cofactors:

A11 = +(-20+ 6) = -14; Ayz = —(15 + 10) = —25; Ay3 = +(9 + 20) = 29
Az = —(10 — 3) = -7; A2z2 = +(5 — 5) =0; A23 = —(3 - 10) =7

A31 = +(—4+
4) =0; A32 = —(—2 — 3) = §; A33 = +(—4-—
6) = -10

~14 -25 29 =14 =7 0

C={\-7 0 7| .«. adj A=C'=|-25 0 5
0 $.5=10 29 7 -10
ane ; 14 ‘7 0
Bow Al
Alm 35 oe A ee
|A| oo PRor
0 2
29 7 -10
; —14 —7 0 4
x=Av-b=—~-|-25
ye! as See 0 5 Ay a Pe ae
29 —10 —1
Multiply it out

38

xX] 2
So finally x = | x2 | = 3 eA = ee ee = oie eae
X3 —4

Once you have found the inverse, the rest is simply x =A™!-b. Here is
another example to solve in the same way:
iff} 24m 0—¥2' +3%3 a= 2
X1 +3xX2 —X3 =H IA
2x; —2x2 +5x3 =3

then Mecseroae- oss -000r Pisa ses orth eo es

 Linear Algebra

xX, =-1] Xo =) Xoi—

The essential intermediate results are as follows:

ye x1 2
1 Sle 23] A iecAcx=b => x=Al-b
2 -—2 5 X3 3

det A= |A| =9
19 27, =8 : beer ss
CH= |e) 4522 adj
A= Cla |--7 <4 5
ES EES Ege ee 7

ated s F
|A| 9
a he, oe eee
ter] eS D a9 =]
x=A) b=3 ey he er 11 |==] 45 | = 5
Sh etecg S07 3 Foie 3
X1 —]
Mise) Xo hs > Sete Lew, Xo = aS
X3 3

Gaussian elimination method for solving a set of linear equations

%

ay G12 @13. ... @in X1 by

Az, 22 23... Aan X2 De ye
: ; : d oH Be s = : ie. A-x=b

Gni Gn2 a4n3 --- Ann Xn bn

All the information for solving the set of equations is provided by the matrix
of coefficients A and the column matrix b. If we write the elements of b
within the matrix A, we obtain the augmented matrix B of the given set of
equations.

B11) \G12 G13 ees yeltin by

: A2i 22 23 ... Gan b2
i.e. B= : / 2 P

Gnt Gn2 4n3.-- Ann bn

(a) We then eliminate the elements other than aj; from the first column by
subtracting d21/a,; times the first row from the second row and 43)/a),
times the first row from the third row, etc.
Matrices 83.

(b) This gives a new matrix of the form:

Qiy “diz dig. Ai by

O: Ce C23 «0c CoH do

0 Cn2 Cn3 te Cnn dn

The process is then repeated to eliminate cj. from the third and subsequent
rows.
A specific example will explain the method, so move on to the next frame

To solve x; + 2x2 — 3x3 =3 40

2X1 — X2 —x3=11

3x1, + 2x2 + x3 = —5

1 2 —3 Xj 3
This can be written |2 -1 —-1 ]|- x2] = 11
3 Zz il X3 —5

12 3 3
The augmented matrix becomes | 2 —1 —1 11
ey tna | —5
Dae.
Now subtract I times the first row from the second row

and :times the first row from the third row.

1 2 —3 3
This gives | 0 —S 5 5
QO —4 10 —14
—4 4
Now subtract =’ i.e. 5! times the second row from the third row.

1 2 -3 3
The matrix becomes |0 —S 5 5
0 0 6 —18
Note that as a result of these steps, the matrix of coefficients of x has been
reduced to a triangular matrix.
Finally, we detach the right-hand column back to its original position:
1 2 -3 Xy 3
QO -S Sy Pe = 5
Ou abe cee 6 X3 —18

Then, by ‘back-substitution’, starting from the bottom row we get:

6x3 =-18 ©. x3= —3

—5x2 +5x3=5 .. — §x2=5+15=20 ©. x2=-4

X1 + 2x2 — 3x3 =3 * 4 —-84+9=3 ~, a

X,
= 2; x2 = —4; x3 = —3
84 Ry Linear Algebra

Note that when dealing with the augmented matrix, we may, if we wish:
(a) interchange two rows
(b) multiply any row by a non-zero factor
(c) add (or subtract) a constant multiple of any one row to (or from) another.

These operations are permissible since we are really dealing with the
coefficients of both sides of the equations.
Now for another example: move on to the next frame

41 Solve the following set of equations:

X1 — 4X2 — 2x3 = 21
2X1 +X2 + 2x3 =3
3X1 + 2x2 — x3 = —2

First write the equations in matrix form, which is............

42

43

DY Hates
se tees C66 Rees Ae

| subtracting 2 times the first row from the second row

-and 3 times the first row from the third row.

1 -4 -2 ZN
So the matrix now becomes |O0 9 = 6 —39
O 14 5 —65 ]
and the next stage is to subtract from the third row ............ times the
second row.
Matrices «Bs

14 45

1 —4 —2 | AA |
If we do this, the matrix becomes | 0 9 6 —39
0 O -4.33 | -—4.33
1 -4 —2 X1 21
Re-forming the matrix equation | 0 9 Galati xs | = —39
0 QO —4.33 X3 —4.33
Now, starting from the bottom row, you can finish it off.

Gp Se Se = Sr oe

As we have seen, a basic knowledge of matrices provides a neat and concise 47 |

way of dealing with sets of linear equations. In practice, the numerical
coefficients are not always simple numbers, neither is the number of
equations in the set limited to three. In more extensive problems, recourse
to computing facilities is a great help, but the underlying methods are still the
same.
All that now remains is to check down the Summary and the Can You?
checklist. Then you can work through the Test exercise.

Summary
1 Matrix — a rectangular array of numbers (elements). 48 |
2 Order — a matrix order of (m x n) denotes m rows and n columns.
3 Row matrix — one row only.
4 Column matrix — one column only.
5 Double suffix notation — a34 denotes element in 3rd row and 4th column.
6 Equal matrices — corresponding elements equal.
7 Addition and subtraction of matrices - add or subtract corresponding
elements. Therefore, for addition or subtraction, matrices must be of the
same order.
: oP ee a ae
/ al iy A aes a Linear Algebra
vi. aS ee

8 Multiplication ofmatrices
(a) Scalar multiplier - every element multiplied by the same constant,
i.e. k[ aj] = |kaj]. ;
(b) Matrix multiplier — product A-B possible only if the number of
columns in A equals the number of rows in B.

ao bh. eye : _ (9g+bh+ci aj + bk + cl)

de f ay Sy dg+eh+fi dj +ek+ fl /
l

Square matrix — of order (m x m)

She Apc |
(a) Symmetric if aj = aj, e.g. |2 6 1
re MG
0 2 4
(b) Skew symmetric if ay = —aji, e.g. | —2 O 1
—4 -1 0
10 Diagonal matrix — all elements zero except those on the leading diagonal.
11 Unit matrix — a diagonal matrix with elements on the leading diagonal all
1 0 0 ‘
unity, ie. | O 1 O } denoted by I.
0 0 1
12 Null matrix — all elements zero.
13 Transpose of a matrix — rows and columns interchanged. Transpose of A
is Al. .
14 Cofactors - minors of the elements of |A| together with the respective
‘place signs’ of the elements.
15 Adjoint of a square matrix A — form matrix C of the cofactors of
the elements of {Aj, then the adjoint of A is C’, i.e. the transpose of C.
adj A=C!.
16 Inverse of a square matrix A

a ajA_c
|A| [Al
17 Product of a square matrix and its inverse
A-A'=A!-A=I
18 Solution of a set of linear equations
A-x=b x=A'!.b
19 Gaussian elimination method — reduce augmented matrix to triangular
form, then use ‘back substitution’.
Now for the Can You? checklist
Matrices

4 Can You?
Checklist 3 49
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames

e Define a matrix?
Yes [ ]
CG)» @)
No
e Understand what is meant by the equality of two matrices?
Yes j
Z°@
| J No
e Add and subtract two matrices?
Yes kel bes LJ
(Go)
» G1)
] No
Multiply a matrix by a scalar and multiply two matrices
together?
C2)
» Ga)
Yes Se fh E) Np
Obtain the transpose of a matrix?
aoe
C2)
» ()
lie td). tg) GW? O/ No
e Recognize special types of matrix? (21)
» @)
Yaa eo Lo Lk dd ida ice
e Obtain the determinant, cofactors and adjoint of a square
matrix? (2)
» @)
Prmerte st AGI a te UND
e Obtain the inverse of a non-singular matrix? Gi)» G4)
Yes CL) C) CO C) CJ No
e Use matrices to solve a set of linear equations using inverse
matrices? (Gs)
» Ge)
Yes s C) OC) C) L) No
Use the Gaussian elimination method to solve a set of linear
equations? Ga)» &)
Foe te a LL) No
88 Linear Algebra

& Test exercise 3

|5Q The questions are all straightforward and based on the work covered. You will have
no trouble.

Daas Ome Ey sy |PGE :

1 A= (4 7 0 ;) and B= (3 16 ;,)determine

(a) A+
B and (b) A—B.

2 Given that A=
Ave and B= 8
a esO | determine
it Sp <4 its v7

(a) 3A, (b) A-B, (c) B-A.

YEA Pee
3 IfA= ( 7 i)form the transpose A' and determine the matrix
S006

product A’ -I.
aoe:
4 Show that the square matrix A = jee) ; is a singular matrix.
-1 8 2

1 4 3
5 TiA= (:ead )determine (a) |A| and (b) adj A.
eZ
Dies lars
6 Find the inverse of the matrix A = [ 5 7
2 OiaG
7 Express the following set of linear equations in matrix form:
2X1 + 4x2 — 5x3 = —7
X, — 3x2 +x3 = 10
3x1 + 5x2 + 3x3 =2

8 Solve the following set of linear equations by the matrix method:

X, + 3x2 + 2x3 =3
2X1 — X2 — 3x3 = —8
5X1 + 2x2 +x3=9

9 For the following set of simultaneous equations:

(a) form the augmented coefficient matrix
(b) solve the equations by Gaussian elimination.
X; + 2x2 +3x3=5
3X1 — X2 + 2x3 = 8
4x, — 6x2 — 4x3 = —2
Matrices meteor ay
i Zi Ere

4 Further problems 3
4
a= (5 2 =(3 g)»determine:

(a) A+B, (b) A—B, (c) A-B, (d) B-A

baw
(5 5) >:\ ne
8=(4 0) e=(5 0) #(0 1):
ee Oudi =( 1
20 Cast Aa 0

where j = V—1, express (a) A- B, (b) B-C, (c) C- Aand (d) A’ in terms of other
matrices.

3 IfA= (33 oa) enh y= & a) detennine

(a) B', (b) A-B, (c) B’-A
4 Determine the value of k for which the following set of homogeneous
equations has non-trivial solutions:

4x; + 3x2
— x3 = 0
7X1 — X2 — 3x3 a)

3x1 — 4x2 + kx3 =0

5 Express the following sets of simultaneous equations in matrix form:

=2
(a) 2x; — 3x2 —x3 (b) x; — 2x2 — x3 + 3x4 = 10
X1 + 4x2 + 2x3 =3 2x; + 3X2 +X4 = 8
X1 —X¥2+X%3=5 xX; — 4x3 — 2x4 =3
—X2 + 3x3 +X%4 = —7

method:
In 6 to 10 solve, where possible, the sets of equations by a matrix
6 2i; +i2 + i3 = 8 7 3x+2y+4z=3

5i, — 3i2 + 213 = 3 x+ty+z=2

Ti; + in + 3i3 = 20 2x -y+3z=-3

8 4i, — Siz + 6i3 = 3 9 3x+2y+5z=1

x-y+z=4
8i; — 7iz — 3i3 = 9

7i; — 8i2 + 913 = 6 6x +4y+10z=7

10 3x, + 2x2 - 2x3.= 16
Ax; + 3x2 + 3x3 =2
—2x1 + X2 —X%3 = 1
 Linear Algebra

In 11 to 13 form the augmented matrix and solve the sets of equations by Gaussian
elimination: ;
11 5 —in +273
=3 12) i, + 2i2 + 3i3 = —4
2i, + 4i2 +i, = 8 2i; + 6i2 — 3i3 = 33
ij + 3i2 — 313 = 2 4i, — 2i2 +i3 = 3

13 7i, — 4i2 = 12
—4i, + 12i2 — 6i3 =O
—6i2 + 14i3 = 0

14 Ina star-connected circuit, currents i, iz, i3 flowing through impedances Z;,

Z2, Z3, are given by:
ij +i2 +i3 =O
Zi = Z212 =@; —@2

Zoi; — 2313 = €2 — 63
If Z; = 10; Zz = 8; Z3 = 3; e; — e2 = 65; e2 — e3 = 160; apply matrix methods
to determine the values of ij, iz, i3.

15 Currents of ij, i2, iz in a network are related by the following equations:

Z11, + Z3i3 = V
Z2i12 = 2313 a0)

ij —i2 —i3 =O

Determine expressions for ij, iz, iz, in terms of Z;, Z2, Z3 and V.

4
Matrix algebra
Learning outcomes
When you have completed this Program you will be able to:
Determine whether a matrix is singular or non-singular
Determine the rank of a matrix
Determine the consistency of a set of linear equations and hence
demonstrate the uniqueness of their solution
Obtain the solution of a set of simultaneous linear equations by using
matrix inversion, by row transformation, by Gaussian elimination
and by triangular decomposition
Obtain the eigenvalues and corresponding eigenvectors of a square
matrix
e Demonstrate the validity of the Cayley-Hamilton theorem
Solve systems of first-order ordinary differential equations using
eigenvalue and eigenvector methods
Construct the modal matrix from the eigenvectors of a matrix and the
spectral matrix from the eigenvalues
Solve systems of second-order ordinary differential equations using
diagonalization
Use matrices to represent transformations between coordinate
systems
92 Linear Algebra

Singular and non-singular matrices

|1 Every square matrix A has associated with it a number called the determinant
of A and denoted by |A). If |A|] 4 0 then A is called a non-singular matrix.
Otherwise if |A| = 0, then A is called a singular matrix.

Example 1

: 2 )
IsA={4 7 6 | singular or non-singular?
Qasoeks
2s
(Ave 447° £6 a
DS" #3

=|. S|-a5 Seas 3

as Cae oS
= (21 — 30) — 2(12 — 54) + 8(20 — 63)
= —9+ 84 — 344
= —269
Because |A| 4 0 then A is non-singular.

Example 2

Ao Pah
Is A= {1 S 6) singular or non-singular?
2 / 4
URE an

|2 singular

Because

|3 O22.
IAl=41 5 36
Wy See Oh
= 3(20 — 42) —9(4-—12) + 2(7 — 10)
= —66+ 72-6
=a)

Because |A| = 0 then |A| is singular.

Matrix algebra 93

Exercise
Determine whether each of the following is singular or non-singular.
4 5 3 -4
1 jAl=| 2 3
2 |B\= -6 8
4 1 -2 s 2° 4
3 |Cj=/17 3 4 Dij=[:5 146
et: ae Se

1 non-singular 2 singular 3
3 singular 4 non-singular

Because
Straightforward evaluation of the relevant determinants gives
1 |Aj=2 2 |B) =0
3 |Cj=0 4 |D|=-5S
Closely related to the notion of the singularity or otherwise of a square matrix
is the notion of rank of a general n x m matrix.

Rank of a matrix
The rank of an n x m matrix A is the order of the largest square, non-singular
sub-matrix. That is, the largest square sub-matrix whose determinant is non-
zero. If n =m, so making A itself square, then this sub-matrix could be the
matrix A itself.

Example
ae: eee
To find the rank of the matrix A= {| 1 2 3 | we note that
a 5vG

aa Memes
1, Es 2 ee ed eee
4-5-0
94 Linear Algebra

|4 | .2|
Because
Sra
Als |e 2s
45526
= 3(12 — 15) —4(6 — 12) + 5(S — 8)
=—9+24—15=0
Therefore we can say that the rank of Ais............

5 Because
7
|A| = O and therefore A is singular.
Now try a sub-matrix of order 2.
|
3 4
|
=6—4= —2+ 0. Therefore the rank of Ais | ...........
ded.

6 ie
Because
The largest square, non-singular sub-matrix of A has order 2 therefore A has
rank 2.
This method of finding the rank of a matrix can be a very hit and miss affair
and a better, more systematic method is to use elementary operations and
the notion of an equivalent matrix.
Next frame

Elementary operations and equivalent matrices

|7 Each of the following row operations on matrix A produces a row equivalent
matrix B, where the order and rank of B is the same as that of A. We write
A~B.
1 Interchanging two rows
2 Multiplying each element of a row by the same non-zero scalar quantity
3 Adding or subtracting corresponding elements from those of another row
are operations called elementary row operations. There is a corresponding set of
three elementary column operations that can be used to form column equivalent
matrices.

es
Matrix algebra

Example 1
604.
GivenA=|{1 2
Ain 5

Seta: 5 —2 -4
je ePALL sd baw 2 3 by subtracting 3 times each
7yaoi a 5 6 element of row 2 from row 1

—2, —4
2 | by subtracting 4 times each
element of row 2 from row 3
iaww eo)
nd
by multiplying each element of
row 1 by 3/2

Ao ae by subtracting corresponding
SIS)
GS
eS)
BPRS)
aby
tre
Le
0 O elements of row 1 from row 3

The row of zeros in matrix B means that its determinant is zero and so its rank
is not 3. The largest sub-matrix with non-zero determinant has order 2 and so
the rank of B is 2. Because matrix B is row equivalent to matrix A we can say
that the rank of A is also 2.

Example 2
Lies
Determine the rank of A= |4 7 6
OF Saas

By taking 4 times the elements of row 1 from row 2 we obtain the equivalent
Matti Xipysapicy®F

By taking 9 times the elements of row 1 from row 3 we obtain the equivalent
WISI A a tere occa

By multiplying the elements of row 2 by —13 we obtain the equivalent matrix

Gees PON er
Ges
Ne
eee
Linear Algebra

By adding corresponding elements of row 2 to row 3 we obtain the equivalent

matrix UANt AVE.

11

Because all the elements below the main diagonal of this matrix are zero we
call the matrix an upper triangular matrix. By inspection we can see that the
determinant of this triangular matrix is non-zero, being the product of its
three diagonal elements 1 x 13 x (—69) = —897. Therefore its rank is 3 and so
the rank of matrix A is also 3.
Try another one for yourself.

Example 3

eee ey
PeAerAn OL A =" Pk O ROut WSs ts cccce
se
2p 44.

12
4

Because

S892 —6 —16
Ae 6 Loss 5 6 Subtracting 3 times row 2
9 7 a 7 4 from row 1

5 6 Subtracting 2 times row 2

Pee we from row 3

nn OV Multiplying row 1 by —1/2

Adding row 1 to row 3

Interchanging rows 1 and 2

Ont
OL
tf
ot
et
te
a
©
ON OAH
CAN
Matrix algebra

ak
and oS $f =0. So the rank of this matrix is not 3. The largest
0 0 0
square sub-matrix of this matrix with non-zero determinant is, by inspection,
of order 2 and so the rank of this matrix, and hence the rank of the equivalent
matrix A is 2.
Finally try a non-square matrix.

Example 4

De ae
ihe rank Or A= 1'0 8 2 4.) isi....J.t-...
ae aes ae?

13
Because
Z'2ZEN3% 4 0 -12 -3 -3
Aerie 29 oA 10 g. pny Subtracting2 times
| a des 2 1 7 3 2 row 3 from row 1

0 -8 -2 -2
AAO” yp Ded Multiplying row1
ee eae A Mh
eet Sasa
ao lidiarR 2.4 Adding row
2 to
NEE ae Page row1
It is possible to find a 3 x 3 sub-matrix of this matrix that has non-zero
determinant, namely
OO 2 eZ
8 2 4 ]| where 8 2 4 = 2(24 — 14) = 20.
SSPE ee? fe tee?
Consequently, this matrix and hence matrix A has rank 3.
 Linear Algebra
EES.

Consistency of a set of equations

|14 In solving sets of simultaneous equations, we can express the equations in
matrix form. For example
€y1X1 + 4y2X2 + 413xX3 = by
A21X1 + A22X2 + d23X3 = b2
431 X1 + 432X2 + d33X3 = b3
can be written in the form
a1 42 443 X1 by
a2, 422 423 x2 | = | be
431 432 433 X3 b3 .
i.e. Ax=b
The set of three equations is said to be consistent if solutions for x1, X2, x3 exist
and inconsistent if no such solutions can be found.
In practice, we can solve the equations by operating on the augmented
coefficient matrix, i.e. we write the constant terms as a fourth column of the
coefficient matrix to form Ap.
a1 a2 43 dD,
Ap = | 421 22 23 D2
a31 432 433 bz
which, of course, is a (3 x 4) matrix.
The general test for consistency is then:
A set of n simyltaneous equations in n unknowns is consistent ifthe rank
of the coefficient matrix A is equal to the rank of the augmented matrix Ap.
If the rank of A is less than the rank of Ap, then the equations are
inconsistent and have no solution.
Make a note of this test. It can save time in working

15 Example

ib # on X41 4
If = th
ie Ae (5) i
1 ure
A= ( ) and d= (5 ee
2 6 246-5
les
Rank of A: ; [= 6 = 0 ., rank
of A=1

5
Rank of Ap: = 0 as before
2b
3 4
but 6 «|=15 -24=-9 |. tank
of Ay = Z
Matrix algebra

no solution exists 16
Remember that, for consistency,
Fak OA ce tcc

17

Uniqueness of solutions
1 With a set of n equations in n unknowns, the equations are consistent if
the coefficient matrix A and the augmented matrix Ay are each of rank n.
There is then a unique solution for the n equations.
Note that if the rank of A = n then A is a non-singular sub-matrix of Ay
and so the rank of Ay = n also. Therefore there is no need to test for the
rank of A» in this case.
2 If the rank of A and that of A, is m, where m <n, then the matrix A is
singular, i.e. |A| = 0, and there will be an infinite number of solutions for
the equations.
3 As we have already seen, if the rank of A < the rank of Aj, then no solution
exists.
Copy these up in your record book; they are important

Writing the results in a slightly different way:

18
With a set of n equations in nm unknowns, checking the rank of the
coefficient matrix A and that of the augmented matrix Ap enables us to see
whether
(a) a unique solution exists
rank A= rank Ay =n

(b) an infinite number of solutions exist

rank A= rankAy =m <n

(c) no solution exists

rank A < rank Ay

Example
—4 5 X1 ay —3
—-8 10/\x2/ \-6
Finding the rank of A and of Ay leads us to the conclusion that
100 Bee Linear Algebra

19 there is an infinite
number of solutions

Because

A aes and A cin ee)

-~8 10 Poa 81 0-6
Rank of A: |3 1o[= ~40-+40 =0 = Rank of Al= I.

_ 5 by 8 4 3
Rank of Ay: = 0; = (): =")
zea > a 1 ra ihe sr
Rank of Ay = 1 =
Rank of A= rank of Ay = 1

But there are two equations in two unknowns, i.e. n = 2

Rank
of A= rank of Ap =1<n
‘, Infinite number of solutions.
4x4 —3
The solutions can be written as x; arbitrary and x2 = arr

You will recall that, for a unique solution of n equations in n unknowns

20 4

Now for some examples for you to try. In each of the following cases, apply
the rank tests to determine the nature of the solutions. Do not solve the sets of
equations.

Example1

ee X41 1
ee a Xo ke | 2
1 4 3 X3 3

eZ. — 1 eZ. ak i
A= 1735 4 FZ and5An== [304 wee
1 AY 3 1 4 3 3

)
Matrix algebra 101°

a unique solution exists 21

Because

n=3; rank of A=3; rank of Ay = 3.

rank of A=rank of A, =3=n_ .. Solution unique
And this one.

Example 2
vane Sema} x] Z
4 Lok et ee
3 il X3 1

This time we find that

no solution is possible 22

Because
2 -1 7 2 £102
Aree 4 eo te Ape |4 22°85
3 1. 3 3 Rae ee |
n=B: rank of A = 2; rank of Ap = 3
*. rank of A < rank of Ay
*. No solution exists

and finally

Example 3
Lees XxX 1
pees 4+ X2 = 2
2 35 1 X3 S

In this case, we find that ............

ee
f02 = Linear Algebra

23 infinite number of solutions possible

Because
1) 2a =5 lA 2 dad: 1
A= e738 aad and. Avieas| 1 G3 AN a Z
24. on A Zine 2S
Rank of A:
12243 1° 2 -3
‘Aca iid oO e414 Oe.1 © *7 Subtracting row 1 from row2
2aS eel PD |
12 -3
Ot Subtracting 2 times row 1
Oar from row2

Whe Don sid

ot (PO ail Subtracting row 2 from row3
O7.080
and so rank of A is 2 by inspection.
Rank of Ay:
| ee Aan ea | 12 3.)
Ape |1- 3 4 22-Paepert oe Subtracting row 1
op see iees 2 15 Wpliieg )'afomtqw2
1 2. 28h
Sik bl a Cts Recwed | Subtracting 2 times
017 1 row 1 from row 2

iin eae Tae ah

el Out aoe et Subtracting row 2
00°60 °6 from row 3

and so rank of Ay is 2 by inspection.

Therefore rank of A = rank of Ap = 2 < n (that is 3), therefore there is an
infinite number of solutions.

Now let us move on to a new section of the work

Matrix algebra 103

Solution of sets of equations

1 Inverse method 24
Let us work through an example by way of explanation.

Example 1
To solve 3x,;+2x2-— x3=4
2X1 — X2+2x3=10
X1 — 3x2 — 4x3 =5.
We first write this in matrix form, which is

25

3 2 -1 Xj 4
Thenif A=|2 -1 2] then | x. ] =A?! 10
1 -—3 -4 X3 5

where A’! is the inverse of A.

To find A“!
(a) Form the determinant of A and evaluate it.
peer
|A|=|2 -1 2| = 3(4 + 6) — 2(-8 — 2) — 1(-6 + 1) = S55
1 -3 -4
(b) Form a new matrix C consisting of the cofactors of the elements in A.
The cofactor of any one element is its minor together with its ‘place
sign’
Ay, Ai2 Ai3
ie. C= | Aoi Age Ads
A31 A32 A33
where Aj, is the cofactor of a); in A.
—1 2 2
=10; Ay= — = 10;
Me sat tlie ae ; ZA
2 -l
Ay3 = =-—§
i ; a
2 -l1 3 -l
-
wy a Ed3/113 ve = ; aa=-11;
3 W)
23 ; 4
 10 10 -S
So C=] 11 -ll il
3 -8 -—7
We now write the transpose of C, i.e. C’ in which we write rows as columns
and columns as rows.

27 1621.31.35
c= (10 =}} <§
-S 11 7

This is called the adjoint (adj) of the original matrix A

i.e. adj A=C'
Then the inverse of A, ice. A~! is given by
l 1 LOA. 3
At=—xcC'=—]| 10 -11 -8

As a check that all the calculations have been done correctly and without
error, the product of matrix A with its adjoint should be equal to the unit
matrix multiplied by the determinant of A. That is
AxadjA=detAxI
For this case
3 2 -1 oy rk eee
AxadjA=|]|2 -1 2 10 -11 -8
1 -3 -4 -§ ll —7
aa. ee
= fey oo: e
o 0S
=detAxlI

Thus all is well. We can now continue to find the solution.

xX, 4

So | x2 | =A™'] 10 |becomes
X3 $

Xy a 31 3 o
Xe i= a 10 -ll =f 10.) seuieeis ie.
X3 S53] 7 $
Matrix algebra
105

Because

x1 10 11 3 4
1
X2 |= 55 10 =117 =s 10

X3 —5 11 —/7 5

40 +110 +15 3
1
=s5 40 -110 -40 |=] -2

—20 +110 —35 1

* A= Ae
= — 2: 2X3
=1

The method is the same every time.

To solve Ax=b x=A™'b
To find A“!
(1) Evaluate |A|
If |A| 4 0 then proceed to (2)
If |A| = 0 then there is no inverse and hence no unique solution. Later we
shall discover how to determine whether there is an infinity of solutions
or none.
(2) Form C, the matrix of cofactors of A
(3) Write C', the transpose of C
(4) Then Av! = ialx Cl.
Now apply the method to Example 2.

Example 2

4x, +5x2+ x3=2

—106— Linear Algebra

Here is the complete working.

4 5 il 4 5 1
KE Boe? 3 eA ae 283 20
3 -1 -—2 8. 1 —2

Ay Ai2 Ai3
C=] Agi Azo Az3
Az1 ~A32~ A33
— aS i) = 3 i =?
Au ie = 1 Aj2 ; 4 13 ; 1
5) 1 4 1 4 5
Anes
21 |1 % ECE ita
22 ; a 211 ofAgg23 = ; 3 =19
3 1 4 1 4 5
A31 = i a =—13 Az=
32 : 3 = 13 A33333 = “A= —13

1 —7 5 il 9 —13
C= O11 19 oS) Creat eee lel seme 3
—13 13, .~—13 S 19 .—13
i ty 1 9 —13
A a4 Riek T
SEER otal
26 peed Pe Pe ea
5
Xi 2 —13 2
1
bby || A= 9.N a= 6M a ae
6 = 141 13

X3 1 —13 il

+63 —-13
1
=-— —77
26 aa 13
10 +133 —-13

2
1
=-5-| -78 |=-| -3
5
x, = —2; Koes x3=

With a set of four equations with four unknowns, the method becomes
somewhat tedious as there are then sixteen cofactors to be evaluated and each
one is a third-order determinant! There are, however, other methods that can
be applied — so let us see method 2.

@
Matrix algebra | ee Yd

2 Row transformation method 30

Elementary row transformations that can be applied are as follows
(a) Interchange any two rows.
(b) Multiply (or divide) every element in a row by a non-zero scalar (constant)
k.
(c) Add to (or subtract from) all the elements of any row k times the
corresponding elements of any other row.
Equivalent matrices
Two matrices, A and B, are said to be equivalent if B can be obtained from A by
a sequence of elementary transformations.
Solutions of equations
The method is best described by working through a typical example.

Example 1

Solve 2x%1+ xX2+ x3= 5

X1+3x2+2x3= 1

Ohne = 2X2 = 4x3 = —4,

2. 1 1 xy 5
This can be written | 1 3 y) ey | ]
3 —2 -4 X3 4

and for convenience we introduce the unit matrix

2 1 1 X1 iE Yee B)
1 Wo a oe xo, (eee AS 7.50 1
3 °-2 -4 X3 mt) ad —4

£ 50-6 5
where [«1 |may be regarded as the coefficient of (1
GO. 50s 4 —4

We then form the combined coefficient matrix

ye £1.) 0
155 oy5. 47,0 2D
3 -2 -4 0 0 1

and work on this matrix from now on.

On then to the next frame
a ne LEE
108 Linear Algebra

31 The rest of the working is mainly concerned with applying row transforma-
tions to convertthe left-hand half of the matrix to a unit matrix and the right-
hand side to the inverse, eventually obtaining

bree
ik AO Oy eFe Jae xe
QE OT ate
OF 0 “Liege Je dy

with a, b, c, ... g, h, i being evaluated in the process.

The following notation will be helpful to denote the transformation used:
(1) ~ (2) denotes ‘interchange rows and:
(3) — 2(1) denotes ‘subtract twice row 1 from row 3’, etc.
So off we go.
flare, (2)
i/als- acedinwn2e Liatlcent)
el Bio On-0
SS ee
(2211) 41 3 Zim: 1.40
3)-30( See pe ee 2 0)
(fet tee 0 WOe se)
A ee 1 aes eee. Q\ 8 150
[°Se SE an Fees
0 &—F —4y.4-2)\ £1 yt
—(2)~=(3) /1 32 0 32 Oo
[:Leni titr2 selancth
Ori 3d 2 40
(3)—S5(2) All3 2 O} m1 ot0
[°1 4 2 a SS
Os Obie lie 7, 25
(1)= 32}? /P90 P2108 46 4 3
(Sy (17) [°1 4.2 Ae =
0 0 TPA 79OD PPT TES17
(1)+10(3) /1 0 O 8/17. -2/17. 1/17
(2) — 4(3) [>10 ie 5/17
0 Oe Vogal 7 It 5/17
We now have

(8 2)(2)-2(a at 3)(4)
TOCA

Onis Lele \eks

Gi Xs 1 Sire

1am 5
iene) 5

[ov

Pate oa caoeohe ‘ XD eee $ Sr te oe is We Oe

@
Matrix algebra 109°

X1 \ AO a — Fn ah 34 2
X2 | ==>] —5O0 +11 -12] = + —51 ]=]| -—3
7 17
X3 35 —7 +20 68 4
26 =F X2 = —3; x3=4

Of course, there is no set pattern of how to carry out the row transformations.
It depends on one’s ingenuity and every case is different. Here is a further
example.

Example 2
2X1 = X2 — 3x3 = 1

X,+2x2+ xX3=3
2X1 — 2X2 — 5x3 = 2.

First write the set of equations in matrix form — with the unit matrix included.
TS AY Orcas

If we start off by interchanging the top two rows, we obtain a 1 at the

beginning of the top row which is a help.
(Sieve eet a
2-1 -3 10 0
22-5001
Now, if we subtract 2x row 1 from row 2
and 2x row 1 from row 3, we get
es
410. 3 Linear Algebra

35 ly 2h la Oo ee
PS ea Pe 2
O° =-6 —-7 0 =2 "1

Continuing with the same line of reasoning, we then have

(2) — (3) ieee Oe Ler
0 1 Die L O -1
G: 2% Yanez CHM 26 o]
eCOVe (ie tele Ory ter
ee O -1l
005 6 -2 -5 me

(1) — 2(2) pO ee 2 Notice the three diagonal

(3)+5 |0 1 2 1 O 1} 45 appearing at the
00 1 § -2 -1/ left-hand end

|36 Add three times row 3 to row 1

and subtract twice row 3 from row 2

Right. That gives

(1)+3(3) (1 gO 304 $ 9-2 5-1
(2) 29/3) |Ot a0 gage 33 1
O (0 SFr
Le OcEO X1
Ome =O Xx2 = |
eR
nl | NPR
| nnn NO
OsiOj81 X3 a
oe
ae NN
Sa
nowreeen

Now you can finish it off.

Because

x1 1 8— 3—10 1 —5 =<
x2 mn —7+12+10 foe 1ST ee 3
X3 6— 6—10 —10 —2
Let us now look at a somewhat similar method with rather fewer steps
involved.
SO move on

ey)
Matrix algebra
111

3 Gaussian elimination method

38
Once again we will demonstrate the method by a typical example.

Example 1
2X1 — 3X2 + 2x3 =9

3x1 +2x2-— x3=4

X1 — 4x2 + 2x3 = 6.

We start off as usual

2 —3 2 xX] 9
3 2 -1 Koei] =! [A
1 -4 2 X3 6
We then form the augmented coefficient matrix by including the constants as an
extra column on the right-hand side of the matrix
ine et 8
oe Cee we
Re a
Now we operate on the rows to convert the first three columns into an upper
triangular matrix
(I)~(3) /1 -4 2 6 (2)~(3) /1 -4 2 6
3 2a | Me 2 -3 ZZ aD
2 -3 Ziad 3 2-1 4
(2)—2(1)—-f-1- =4,-- 2 6 (2)+5 (1-4 2 -6
(3)-3(1) |O0 5 -2 -3 (3)+7 |O 1 -% -2
QO 14 -7 -14 0 2-1 -2
(3) 2/2) 71 34iae2y 6 (3)x(-S) /1 -4 2 6
0 1-2 -3 0 1 -% -3
0.46 -} —3 ages tenis Eee

The first three columns now form an upper triangular matrix which has been
our purpose. If we now detach the fourth column back to its original position
on the right-hand side of the matrix equation, we have
112 Linear Algebra

39

Expanding from the bottom row, working upwards

x3 = 4 2. X=

X2 — 2X3 =-3 a8 x2 =—-24+8=1 a bo ll

X, —4%2+2x3 = 6 .. x, -44+8=6 PeaeX hee,

SES Soe1b Set!

It is a very useful method and entails fewer tedious steps, and can be used to
solve efficiently higher-order sets of equations and non-square systems. It can
also solve a sequence of problems with the same coefficient matrix A by using
the augmented matrix (Ab, b2...b,,).

Example2
X1 + 3x2 —2x34+ X4=-1
2X1 —2X2+ X%3-—2xX%4= 1
Xi+ X2—3x%3+ X4= 6
3X1 — X2+2x3-— xX4= 3.

First we write this in matrix form and compile the augmented matrix which is

40

Next we operate on rows to convert the left-hand side to an upper triangular

matrix. There is no set way of doing this. Use any trickery to save yourself
unnecessary work.
So now you can go ahead and complete the transformations and obtain

4
Matrix algebra 113,

Here is one way. You may well have taken quite a different route.
RE Re Aieste on
eae A et
io U3 SG
eT ROCHE e Ae
SO The YE os ere ae
(Qiajmeito #8 Sriey ft Use
Aaa wto 2 21), 6 bly
0 2°03! Lor? 73
yas fi Bh 1 ay
(3) Re en 74 LG
0 teen eo thy
Deak vol) 23
ONDER 5 PEL ees, ae
Bie 104 2 ote Ow 3
OF O=sie peed
060 serie eset 20
a7 ilk eee
Gm eed ott ua
yes ay Pelana ent
Lite Se Se
Returning the right-hand column to its original position
1 SH BA x1 —1
Ope 2.. 3 = 0 hd le 3
0 O -1 0 X3 x 1
0 O O -4 X4 —16
Expanding from the bottom row, we have
—4x, = —16 eS
—x3=1 eee al
—2x2+3x3=3 ©. —2x2=6 J. X2=-3
X,+3x2—2x34+X4=-1 |. x1 -9424+4=-1 See (pee
eeQ 2s to oN Xa— ta 4
114 Linear Algebra

42 We still have a further method for solving sets of simultaneous equations.

4 Triangular decomposition method

A square matrix A can usually be written as a product of a lower-triangular
matrix L and an upper-triangular matrix U, where A = LU.
bk 2:53
For example, if A=] 3 5S 81], Acan be expressed as
A> 9O= 10
if. Dac Uy, Uy2 U3
Aca) — Io4 Ino 0 0 92 nUo3

Ie Hilo M18 OurecOhiiag a

(L) (U)
Kiva, lite 11143
=] buy loiuy2 + lo222 I2113 + I22U23
I3,Uy1 131412 + 132422 131413 + 13223 + 133U33
Note that, in L and U, elements occur in the major diagonal in each case.
These are related in the product and whatever values we choose to put for u11,
Un2, U33.... then the corresponding values of 1/11, /22, 133... will be determined —
and vice versa. :
For convenience, we put 1 = U22 = 433... = 1
hy lyyti2 11413
Then A=LU= | Jy, Ioiti2 +122 M213 + lo2ti23
\lg1 Igiti2 +132 1g113 + 13223 + 133
be |
In ourexample, A=|3 5S 8
4 9°10
ms iS lation —2 feo eZ; nis = ..8ths =
[Ey| — oF Similarly bya e oe A PRPS ohh
oe haus 5
ly = 4; Ey nee, ae ‘ beatae yee ar

loo = —1; u23 = ils 139 =1; 133 =-3

Because
Ip1U12 + Io2U22 = 5 that is 3 x 2+h2x1=5 and sol. =—-1
Ip1U13 + 12223 = 8 that is 3 x 3+ (—1) x U3 = 8 and so 23 = 1
131412 + 132U22 = 9 that is 4 x 2+1]32 x 1=9 and so 132 = 1
13343 + 132U23 + 133433 = 10 that is 4x 34+1x1+h3x1=10
and so [33 = —3
Now we substitute all these values back into the upper and lower triangular
matrices and obtain
Matrix algebra
115

A=EGS3-=1 0 i ey ae |
4} 1 -3 eet

We have thus expressed the given matrix A as the product of lower and upper
triangular matrices. Let us now see how we use them.

Example 1

X1+2x2+ 3x3=16
oe + 5X2 + 8x3 = 43

4x; + 9x2 + 10x3 = 57.

Lees SS xy 16
test tf 3) ye! Soule G43 i.e. Ax = b.
49 10 X3 57

We have seen above that A can be written as LU where

La Oe | Se
A=LU=/;3 -1 0 Se aa
4 1 -3 0 0 1
To solve Ax = b, we have LUx=b_ i.e. L(Ux) =b
Putting Ux = y, we solve Ly = b to obtain y
and then Ux = y to obtain x.
Lie es. 0 V1 16
(a) Solving Ly = b 3 -1 0O y2 | =] 43
4 1 -3 ¥3 a7
Expanding from the top ¥i=16; 3y,-—y2=43 .. y2=S; and
4y, + y2 — 3y3 35) ae 64+ 5 — 3y3 = ees y3=4

Loz <3 X1 16
(b) Solving Ux = y ae eS | x27 = 1 S
OP wee X3 +4

Expanding from the bottom, we then have

 »

Note:
1 Ifd; = 0, then either decomposition is not possible, or, if A is singular, i.e.
A! = 0, there is an infinite number of possible decompositions.
2 Instead of putting uw); = ud: = uy3...=—1, we could have used the
alternative substitution /;; = 2 = /;3... = 1 and obtained values of 8),
U22, W433... etc. The working is as before.
3 One advantage of employing LU decomposition over Gaussian elimina-
tion is in the solution of a sequence of problems in which the same
coefficient matrix occurs.
Now for another example.

46 Example 2
X, + 3x2 + 2x3 = 19

2x;+ X2+ X3=13

4x, + 2x2 + 3x3 = 31.
:ate Some x, 19
211]/|/x]=]13] ie Ax=b
42 3 X3 31
ln OO Q 1 we 3
A=LU= | hi gla 1) on? 23

ly lo b3/\0 0 1
Ii hatha hhiths
=] hn Inti2+h2 bniths + hates
In Ratha + he Its + bettas + bs
: ENS Pe
Fiat Neg
4°23

Now we have to find the values of the various elements. The usual order of
doing this is shown by the diagram.
Matrix algebra pte ae

a 3

etc.

That is, first we can write down values for 1;;, Io), Iz; from the left-hand
column; then follow this by finding u12, u;3 from the top row; and proceed for
the others.
So, completing the two triangular matrices, we have

47

As we stated before: Ax = b; L(Ux) = b. Put Ux=y

then (a) solve Ly = b to obtain y
and (b) solve Ux = y to obtain x.
V1
Solving Ly=b gives | y2 | =
Ys

Because

1 0 0\/y 19
2 =-5 0} y)= {13
os (ee on 31
Expanding from the top gives
¥7=195 yo=S; yz =5.
(b) Now solve Ux=y from which x,=............ ana) ie cn Sn Beeseun
 Linear Algebra

Because we have
Ux >)

a X} 19
ie |Onde 2] [ae] =p 5
ys X3 5

Expanding from the bottom x3 = 5; x2 + = =§ raa=2

and x; + 3x2 +22x3=19 ©. X1+64+10=19:° ©. xr=3

SSeS aS S
We can of course apply the same method to a set of four equations.

Example 3
X, + 2x2 — X3
+ 3X4 9
2x; — X2+3x3+2xy = 23

3x; +3xX2+ X34+ X= S

4x, + Sx2 — 2x3 + 2x4 = -2.
1 2 -1 3 xX} 9

2 -l 3:22 X2 23
i.e. = ~ Ax=b
sh eho 2 Fie cay
4 § =2 2 X4 =—2
iy’ OY UY 0 1 2 Wy3 Wye | re -ae? Ge}
A-LU= ln IpIle O O OQ 1 Ugg Ue: a 2 = 1 2
I3y I3> l33 0 Os 0 1 Uz4 be 3 ihe |

Ig, Iga lag lag § a 0 1 4 § —2 2

hi fiat 1143 hhiuas
re Io, -Iayty2
+ boo Ia1 13 + 12223 Ip) Uy 4 + InQuo4
Ig) Ig,2 +132 Ig1ti3 + I3223 + 193 M31 Ua4 + ba2tag + 133434
Igy Igyya + Igo Lyitig + Lgoti23 + yg grag + Lyrting + Lygtzq + Lay
Now we have to find the values of the individual elements. It is easy enough if
we follow the order indicated in the diagram earlier. So the two triangular
matrices are
 50

As usual Ax=b; L(Ux)=b. Put Ux=y .. Ly=b

(a) Solving Ly =b
1 O O 0 V1 9
2-5 O 0 E 23
yams a OTT ots
4 -3 -1 -8/] \y, —2
Yi
aed Tae Vokes
sat Miva be

Y4

51

(b) Solving Ux =y
12 -1 3 xy 9
01-1 #¢/] x -1
00 1 -8]| x3] | -25
0 O 0 1 X4 5

which finally gives

By cs svcwre conve § XD Sy dec causes>
120 Linear Algebra

Comparison of methods

Inverse method

This is an elementary method but it is very inefficient when the number of

equations to solve increases beyond three.

Row transformation method

An efficient method but each case is different and relies on ingenuity to see
the way forward. iS

Gaussian elimination method

The most efficient method and should be used in most cases. It must be used
when there is a singular or non-square system.

Triangular decomposition method

An alternative to Gaussian elimination in some cases.

Now let us proceed to something rather different,

so move on to the next frame for a new start

Eigenvalues and eigenvectors

53 Matrices commonly appear in technological problems, for example those
involving coupled oscillations and vibrations, and give rise to equations of the
form
Ax
= \x

where A = (qj) is a square matrix, x is a column matrix (x;) and ) is a scalar

quantity, i.e. a number.
For non-trivial solutions, i.e. for x 40, the values of X are called the
eigenvalues, characteristic values, or latent roots of the matrix A and the
corresponding solutions of the given equations Ax = \x are called the
eigenvectors, or characteristic vectors of A.
Matrix algebra 121

The set of equations

Q@iy a2... ~Ajn xX] xX]

A21 G22 ... Aan X2 X2
; : =.

Ani An2 .«.-- Ann Xn Xn

then simplifies to
(a1 — A) a\2 2 aie Gin xy 0
a2 (d22 — X) ee a2n X2 O

an An2 (Ann = X) Xn 0

That is, Ax = \x becomes Ax — \x = 0

ie. (A—AIx=0

the unit matrix I being introduced since we can subtract only a matrix from
another matrix.
For this set of homogeneous linear equations (right-hand side constant
terms all zero) to have non-trivial solutions
|A — \I| must be zero
This is called the characteristic determinant of A and |A—AI{=0 is the
characteristic equation, the solution of which gives the values of A , i.e. the
eigenvalues of A.

Example 1 54

Find the eigenvalues and corresponding eigenvectors of

Ax
= Ax where
A = ¥ “4?

The characteristic equation is |A — AI| = 0

2A 28
- 0, which, when expanded, gives
4 1-A
Nic nk ea ee ANC en ei a aeletewewt
 Linear Algebra

A, =—-2 and A2=5

Because

(2—A)\(1=—A)-12=0 ~~. 2—3A4+A7=12=0

7 —3A—10=0 (A—5)(A+2)=0.' 7. A=-—2eeS
Now we substitute each value of \ in turn in the equation
(A — AI)x = 0
With \ = —2

(G 1)- 0 3G) =o)

2 ool |

jue
O

20)
1 X2

X1
0

0
me “heed cand SD
A X1 0
é sean es)
Multiplying out the left-hand side, we get

from which we get x2 = — $x; i.e. not specific values for x; and x2, but a
relationship bettveen them. Whatever value we assign to x; we obtain a

o=(3)=(4)
=(8)e(a2)
corresponding value of x2.

The most convenient way to do this is to choose x; = 1 and then scale x; to

obtain integer elements. So here we find for x; = 1 then x2 = —4/3 so x, is of
the form

By

3
This is now scaled up by multiplying by 3 to give

xi a(_4) where a is a constant multiplier.

The simplest result, with a = 1, is the one normally quoted.

*, for y=-2, x1= 2

‘Similarly, for A2 = 5, the corresponding eigenvector is
Matrix algebra 123

=O :
Because, with \2 = 5, (A — \I)x = 0 becomes

(G 1)-8( HC) =)
4 1
Ziad
Leipga f
5 0
X2
x,
0
0
iG wip 5) He) = bo)
—3 3 x1 0
( 4 Oa)
— 3X1 +3x2=0 x2 =x
ie.

1
with A2 = S, the corresponding eigenvector is x2 = a(1 )

1
Again, taking 6 = 1, for \2 =5, x2= (A)

So the required eigenvectors are

3
xi = ( a corresponding to A; = —2

1
» Gi (‘i corresponding to A2 = S.

Example 2

Determine the eigenvalues and corresponding eigenvectors of

“area AE,
Ax
= Ax where A=(5 “fr

The characteristic equation is |A — AI| = 0, which in this case can be written as

ee eee eee
LP) eS Rees Linear Algebra

59 |
Because the equation is (3 — \)(4—A)-20=0 .. \*—7\—-8=0 |
. AFtYA=
8) =00 SAS—Lfors

(a) With A; = —1, we solve (A — AI)x = 0 to obtain an eigenvector, which is

2 Because
E@
( aN {(; 10 ( 0
A= A ={-1)
2. Wa ny 01
0
1

>
-. 4%,+10%2=0 ©. x2 =—=X1 X1= a( 2

: 5
‘with a=1 A,;=-—1 and x1 = ( )
‘ —2
(b) In the same way the corresponding eigenvector x2 for \2 = 8 is

Because

G2 a)-*0 3

2,
‘with, 6:=1,. Az—8.-and x2 = ( )

@
Matrix algebra

The same basic method can similarly be applied to third-order sets of

equations.

Example 3

Determine the eigenvalues and eigenvectors of Ax = \x where

10 4
Am V0°2 .0}.
3 1 -3
As before, we have (A — AI)x = 0 with characteristic equation |A — AI| = 0.
1-A 0 4
i.e. 0 2-A 0 = ()
3 1 —3-r
Expanding this we have

Because
(1 —A){(2 = X\(=3 = A) — 0} 4+-4 (0 = 3(2 — A)} =0
(1—A)(2 ~ A)(=3 — A) — 12(2 =A) =0
* (2=A){((1=A\(-3 — A) — 12} =0
* \=2 or W4+2A-15=0 .. (A—3)(A4+5) =0
tA Shay
3) OL =O

(a) With A; = 2, (A — AI)x = 0 becomes

10 4 Laue x) 0
02 Oj; -2 (°1 0 xo/| = 0
oF EL Paes 0° OF4 X3 0
10 4 20 0 x1 0
2 01 = [°Ad | x2 |=10
3 1 -3 0 0 2 X3 0
-1 0 4 Xj} 0

0 0 O pe eo a8
3 1 -5 X3 0

from which a corresponding eigenvector x; is ............

EE
 ae Linear Algebra
ee eae & <

63 4
xX, = —7
1

Because we have — x; + 4x3 = 0 See ie 4X

3x1 +xX2—5x3=0 ©. 3x, +22-321,=0 3. xp =—-Gu1

- - > 7 1 7 i
- X1, X2, X3 are in the ratio 1: —3:7 ie. 4:—7:1 XX,= | —7

(b) Similarly for \y = 3, (A—ADx=0

(8 3)-a(o 49)(2)-(3)
from which a corresponding eigenvector is

64

Because 5
iu. Ss 3 0
02 Oj;-]0 0
3 1 -3 0 0
—2 0
0 0
3 0
J. —2x,+4x%3=0 -
2
Also —x2=0 .. x%2=0 gaxXo — 120
1

Finish it on your own. Method just the same as before.

Matrix algebra 127

65

Check the working.

1\0.; 4
A=|0 2 OO] and A3=-—5 with (A-AI)x=0.
a se
Tee? a4 130-70 x1 0
Owe2s SO 5}) Ov 1 ) to = 10
3 1 -3 Ua ie X3 0
6 0 4 xy 0
.ra <2 [=| 0
Sorky oe x3 0
6x; + 4x3 = 0 x3 = —3X
2
pox = oH 0. <a— 0
—3
Collecting the results together, we finally have
4 2 2
M1 = 2) x1, = —7 r r2 = SF x2 = O Py A3 = —5, x3 = O
1 1 —3

Cayley—Hamilton theorem
The Cayley—Hamilton theorem states that every square matrix satisfies its 66 |
characteristic equation. For example the matrix

Aaha a) 25 3

of Frame 54 has the characteristic equation

2 —3A-10=0
and so the Cayley—Hamilton theorem tells us that
A? —3A-—101=0
128 Linear Algebra

To verify this we note that

A ee 16 9
A = _ | so that
4-4) 12 13
: 16 9 2 3 1 0
A? 3A 20h = 210
12 13 4 1 0 1
16 9 san8 10 0 0 0
= A121 | ow 12-3 a 110)©\o10
You try one. Verify that the matrix A= & A
the characteristic equation
2 -~7\-8=0
satisfies the Cayley-Hamilton theorem, that is ............

67 A2—-7A—81=0
Because
: S10 eG 29 70
Afi so that
Zee 14 36
29 70 3 10 1 0
A“ -7A-8I = =
14 36 4 ant
29 70 70 Sp neo
ye) 44.0436 28 oe) 1-0 6
Now on to something different

Systems of first-order ordinary differential

equations
68 Matrix methods involving eigenvalues and their associated eigenvectors can
be used to solve systems of coupled differential equations, though we shall
only consider cases where the relevant eigenvalues are sisting’. We proceed by
example.
Example 1

Consider the system of two coupled ordinary differential equations

f(x) = 2fi(x) + 3f2(x) where fi (0) = 2 and f2(0) =

f2(x) = 4fi(%) + fal)
These can be written in matrix form as

7)
Matrix algebra 129

(00_(2.3)(fi) =
That is
F(x) = AF(x)

where (fi)
F(x) _= Gl (A)
R(x)_ ESS and. A=
ie
(4 1) and
.
where

F(O) = & oe = (1) are the boundary

fa(0)) ~ \1 4 conditions in matrix form
)
The matrix differential equation F(x) = AF(x) is similar in form to the
single differential equation f’(x) = af(x) (a constant) which has solution
f(x) = ae” (a constant), so to solve the matrix equation we try a solution of
the form
F(x) = Ce where the number k and the constants c; and © of the
, Ci
matrix C = ( )are to be determined.
C2

Substituting F(x) = Ce into the matrix equation F(x) = AF(x) gives

kCe = ACe* 70

Because
F(x) = Ce so F(x) = kCe™. Since F(x) = AF(x) then kCe* = ACe*
Dividing both sides by e gives
kC = AC that is AC = kC.
So, from Frame 53, k is an eigenvalue of A and C is the corresponding
eigenvector. Therefore, we must first find the eigenvalues of A and for this
matrix these have been found earlier in Frames 54 to 57. They are
3
\ = —2 (and so k = —2) with corresponding eigenvector ( ‘)

1
\ = 5 (and so k = 5) with corresponding eigenvector (;)

To each eigenvalue the matrix F(x) = Ce is a solution. The complete solution

to F = AF is then

F, (x) = (<::Jane: and F,(x) = (<::Jaze:

130 Linear Algebra

nm F\ (x) = {cages: and F(x) = ({)aze®

Because

F(x) = Ce is the solution corresponding to the eigenvalue k with associated

eigenvector C.
The complete solution to the equation F(x) = AF(x) is then a combination of
these two solutions in the form

F(x) =A 2\e™ 4B 7
Applying the boundary conditions gives F(0) =............

7 ro)= (in)=(7)
Because

3 1 3 1
F(x) nately le +B(, Je and so F(0) =4(. 4)+B(1)

—4A+B 1
Therefore

with solution A = 1/7 and B = 11/7,

—-4A+B=1
giving the final solution as F(x) =............

73 F(x) = ( re (117
—4/7 11/7

Summary
To solve an equation of the form
F (x) = AF(x)
1 Find the eigenvalues ;, A2,..., An of A (assuming they are all distinct)
2 Find the associated eigenvectors C;, Cz,..., C,
3 Write the solution of the equation as F(x) = >>", (A,e**)C, and use the
boundary conditions to find the values of a, for r= 1, 2,..., n.
Now you try one.
Next frame
Matrix algebra 131

Example 2 74

The system of two coupled ordinary differential equations

f(x) = 3fi(x) + 10f2(x)

where f1(0) = 0 and f2(0) = 1
fo(X) = 2fi (x) + 4f2(x)
has the solution (refer to Frames 57 to 61)

75

Because

fi(x) = 3fi(x) + 10f2(x)

f3(x) = 2f1 (x) + 4f2(x)
can be written in matrix form as

That is
(fs)= S)(G3) i
F(x) = AF(x)

where F(x) = fas} F(x) = (3) and A= ( *,|and where

F(0) = eat
To solve the matrix equation we first need the eigenvalues and associated
eigenvectors of the matrix A. These have already been found in Frames 57 to
61 and they are
S
\ = —1 with corresponding eigenvector ( |

2
\ = 8 with corresponding eigenvector (:)

The complete solution of F’ = AF is then

F(x) =A( 3 Jetva(tle

 Linear Algebra

77 f(x) = SAe* + 2Be™

fo(x) = —2Ae* + Be*

Because

fi(x) 5 = 2 xX

ibe seade a +a(1)¢

and so
fi(x) = 5Ae™* + 2Be™*
fo(x) = —2Ae* + Be**
Applying the boundary conditions, we find a

aCe colle ery

VY OV TS AT OD

78 _f[0\_ ffi()\ _ (SA+2B

Bie @ 5 ee Pee:
Because
The boundary conditions are f;(0) = O and f2(0) = 1 therefore

0) = (1)=(ico)) = Cana)
0 0 5A + 2B
1 f2(0) —2A +B
This gives the pait of simultaneous equations
SA +2B=0
which have solution
—2A+B=1

79 A= -—2/9. and B=5/9

r= (60) = (MB ero (0

This gives the complete solution as

2
Matrix algebra 133

Diagonalization of a matrix

Modal matrix 80
We have already discussed the eigenvalues and eigenvectors of a matrix A of
order n. In this section we shall assume that all the eigenvalues are distinct. If
the n eigenvectors x; are arranged as columns of a square matrix, the modal
matrix of A, denoted by M, is formed
le. M= (X1, b, Neeae Sr,

For example, we have seen earlier that if

Oe 4
A={0°2 0] then A; =2, 42 =3,\3=-5
SM 13
and the corresponding eigenvectors are
4 2 2
eee os | ON Xe O
1 1 —3

Spectral matrix
Also, we define the spectral matrix of A, i.e. S, as a diagonal matrix with the
eigenvalues only on the main diagonal
Ati Owe Dee
Om tor) 3 0
Leo oes :

A ool 0Tn 6 ee Fe oO
So, in the example above, S=............

81

Note that the eigenvalues of $ and A are the same.

5 -6 1
Noyated, Ve= [Pal 1 O | has eigenvalues \ = 1, 2, 4 and
che ata!
0 1 3
corresponding eigenvectors | 1 |, | 1}, | 1
6 2 a
134 Linear Algebra

poate
QO. 3
M= fl 1 Die Soe 0
Gea 0 0 4

Now how are these connected? Let us investigate.

The eigenvectors x arranged in the modal matrix satisfy the original
equation
Ax = AX
Also Ma(xX, x2™ 47x,
Then AM = A(X; X2 ... Xn)
=(Ax, AxX2 ... Ax») s
=(AjX; AzX2 --. AynX,) . since Ax = Ax

\ «OO 0
0 A2 0
Now S= y , t Pa (Aix \OD.GE moe Anda) = MS

OF OR aA,
“. AM = MS

If we now pre-multiply both sides by M_! we have

M'AM=M'MS_ But M'M=I
“. M'AM=S
Make a note of this result. Then we will consider an example

83 Example 1

From the results of a previous example in Frame 65, if

te 4
Av |eOme Q | then A; =2, Az =3;,A3-=.—5 and
3 1 -3
4 2 2
Xs ee eX 1 (eee gee 0
1 1 3,
42 2,
AlsoM=j{-=7'0 0
Tl 4 =3
We can find M ' by any of the methods we have established previously.
Matrix algebra
135

(Oe =12 84 |
M! = 3/8 1/4 1/4
1/Se 1/28. 1/4

Here is one way of determining the inverse. You may have done it by another.
ole 2 ny NG pa Orel O
tea 0 OF OX Oia |e ee) 04
beiesa! 0 041 i ae a sa LA
1 @.'0 40 —1 pmo 00) a DT u0e aype40
~ [°es Panel a egFas Eee [°Linsdhl Dei /Ee 4
OU Zeman s7 ad 0908 Billo af7e- 2
POP 0s Peetiye ae
~ [°eS tele cla 9 dalla |
eae hanes Wir enBee ae
1° 0 BO dels Clete erecal)
0 ! 3/8 7/28 1/4
0 0 ' 1/8 1/28 -1/4
O- 1/7 * 0
“ M1=1]3/8 1/4 1/4
18° 1/28" 174
1 Os ee a
So now A=|0 2 QO] and M=|] -7 O 0
3 1 =3 1 it 3
le 5 pam oe pe 8 6 -10
AM=|0"°2 O]|=7 0 O|}=/[-14'0° oO
3° 1as3 i bee ye eeu (-:
g* =1/7 "0 86210
Then M'AM=] 3/8 1/4 1/4]] -14 0 0
1/8 1/28 -1/4 7 ave 1S
136 Linear Algebra

85

So we have transformed the original matrix A into a diagonal matrix and

notice that the elements on the main diagonal are, in fact, the eigenvalues of
A
ie. M 'AM=S
Therefore, let us list a few relevant facts

1 M ‘AM transforms the square matrix A into a diagonal matrix S.

2 A square matrix A of order n can be so transformed if the matrix has n
independent eigenvectors.
3 A matrix A always has n linearly independent eigenvectors if it has n
distinct eigenvalues or if it is a symmetric matrix.
4 If the matrix has repeated eigenvalues and is not symmetric, it may or
may not have n linearly independent eigenvectors.
Now here is one straightforward example with which to finish.

Example 2
—-6 5
A= ( ), M =e. fara te Mie = i gto of:
4 2
and hence M-'AMe=]}............
Work through it éntirely on your own:
(1) Determine the eigenvalues and corresponding eigenvectors.
(2) Hence form the matrix M.
(3) Determine M “|, the inverse of M.
(4) Finally form the matrix products AM and M_!(AM).
Matrix algebra
137

: fA. 6 :
( ea
1/6 OY, a
eee
AM = (5 vg)
86

Here is the working. See whether you agree.

a=( -6rei5 ) —6—-\ 5§

=0
Am ON
(-6—A)\(2—A)-20=0_. 074.4. -32=0
(A-—4)(A+8)=0 «. X=40r-8

(a) 4; =4
White =(or)baa) (a)
4 2
=
QO 4

5
X2
-_
0

— 10x, + 5x2 =0 we GI x= (5)

oes 0 a P99al
‘Spe wpe -—6 5 8 0 xX] é:

ts io)le)-(
LD Xj =

=,
EU alkitee Oxo Oo X2 = —2x x m= (>

i eas fa (5
Seedy 4)
: a 1 tae Reng8
To find M € Sp A

Operating on rows, we have

Gacy es all Sams

Spine |e erie
=(5 iti begVE ae

Cr ae eaiaiyed
Sl Oe) Car) a Ciiaety
aon. Ht EGY od
‘. M-'AM = o 3)
138 Linear Algebra

Systems of second-order differential equations

The process of uncoupling a system of differential equations to obtain their
solution can be achieved by diagonalizing the matrix of coefficients. For
simplicity we shall only consider second-order equations and again, we
proceed by example.

Example 1

Consider the system of coupled second-order differential equations

1(x) = 2fi(x) + 3fa(x)

2(X) = 4fi(x) + fa(x)
where f;(0) = 2, f2(0) = 1, f[(0) = 4 and f3(0) =3
These can be written in matrix form as

= That is
(G8 Zea) (Ae)
F’ (x) = AF(x)

where F(x)= (rea E'(z) = te) and A= a i and where

F(0) = Gan. =\7) and F(0)= ct) = é) are the boundary

conditions in matrix form.
The matrix differential equation F’(x) = AF(x) is similar in form to the
single differential equation f”(x) =af(x) (a constant) which has solution
f(x) = ae’ + Be-V™ (a, 6 constants), so to solve the matrix equation we try a
solution of this form. We already know from Frames 54 to 57 that the
eigenvalues and eigenvectors of matrix A are
3
\ = -—2 with corresponding eigenvector ( A)

1
A= 5S with corresponding eigenvector (a

The modal matrix of A is the matrix M and the spectral matrix of A is the
matrix § where

m= (" = io s=(-- me

#
Matrix algebra 139

Because

The modal matrix is formed from the eigenvectors of A. That is

M = > 5 where the two eigenvectors are (by and (

The spectral matrix is formed from the eigenvalues of A. That is

$= Ge 2 where the two eigenvalues are — 2 and 5

If we now define the matrix G(x) by the equation F(x) =MG{(x), then
differentiating gives
F’(x) = [MG(x)|"= MG" (x) where
F’(x) = AF(x) = AMG(x)
and so, from Frame 85, M-'MG"(x) = G(x) = M-'AMG(x) = SG(x). That is
G" (x) = SG(x)
Therefore, in component terms
Mis |Sty Vn — 12 aN eee
e"(x) = (1) )= sew) & 5) \g0(x)
81 (x) = ...81(x) with solution gy (x) =
82 (X) =...82(X) with solution g2(x) =

81(X) = —2g81(x) with solution gy(x) = kyel¥2 + kype V2

§5(X) = 582(x) with solution g9(x) = ko e¥* + koe -V3x
140 Linear Algebra

)a (3ky,e/V2* cia 3kie-IV* sh ke * i koe~ V9

—4k,,e/V2* = Ake WV2* + ka e¥5* + kope7¥*

we = (AG))= M60) = (4 1) (reef ees

BPAY = —( 3 1)\ (ky ei¥2*
+ ky2e5V

and so |

fi (x) = 3kiel¥* + 3ky2eIV™ + ky IV + kype-V*

and

fo(x) = —4kye/¥™ — 4kype

WV 4+ kp eY* 4+ kage V*
This solution can be written in terms of circular and hyperbolic trigonometric
expressions as
ir, PCOS....X
3 O6in. 24x
FX) = (7 epimeebeec ny |

92 F(x) = 3 1\/ Pees V2x

+ Qsin 2x
—4 1/)/\Rcosh V5x
+ Ssinh 5x

Because

3k,e/¥2* te 3k, xe V2

= 3ki (cos V2x +jsin v2x) + 3ki2 (cos V2x —jsin v2x)

= Pcos V2x + Qsin V2x

where P = 3ki; + 3ki2 and Q= (3k — 3ki2)j
and

ko1eY* + ko2e~V5*
= Koy (cosh V5x + sinh v5x) + ko2 (cosh V5x — sinh v5x)

= Rcosh V5x + Ssinh V5x where R = ky; + ko and S$ = ko — ko

Therefore

@
Matrix algebra 141

F(x) = (3P cos V2x + 3Q sin 2x + Rcosh V5x + Ssinh V5x 93

—4P cos /2x — 4Q sin /2x + Rcosh V5x + Ssinh V5x

That is

FX =. reso
FHX) 4s saees

fi(x) = 3P cos V2x + 3Q sin V2x + Rcosh V5x + Ssinh V5x 94

—4P cos V2x — 4Q sin V2x + Rcosh V5x + Ssinh V5x

and so
fi(x) = 3P cos V2x + 3Q sin V2x + Rcosh V5x + Ssinh V5x
fo(x) = —4P cos V2x — 4Q sin V2x + Rcosh V5x + Ssinh V5x
Applying the boundary conditions, we find

FO =(7)= (py) and FO =(3) = ("94's)

(ea islaus F()= (4) =( 3720+ 55.) 95
—4P+R S —4/2Q + V5S

Because

fi(0)= 2, f2(0 , fi(0) = 4 and f3(0) = 3 and so

F(0)=
(2) (8)-(308na
F(0) = @ )- roe Gee
(0)

—4V2Q + V5S
This gives the two sets of simultaneous equations
3P+R=2 3V2Q+ V5S=4
an which have solution
—4P+R=1 —4/2Q
+ V5S=3
Pe cee De Sas atte 3:
142 Linear Algebra

96 P=1/7,R=11/7, Q=1/(7v2) and $= 25/(7V5)

This gives the complete solution as

: *) 3 ab eA
Ave) = 7 COS V2h sin
= 294 — cosh V 5x sinh 52
7/2 7 7/5
25
f2(xx) = —Scos V2x - = pin v2e+ S
cosh V5x+ sinh Va

This method is quite straightforwardly extended to three or more such

coupled differential equations.

Summary
To solve the system of coupled second-order differential equations
F’ (x) = AF(x)
1 Find the eigenvalues and eigenvectors of matrix A and construct the
modal matrix M and the diagonal spectral matrix $
2 Solve the equation G’(x) = SG(x)
(note that even though M | is used there was no need to
calculate it) :
3 Apply F(x) = MG(x) to find F(x).
Try one yourself.
Next frame

97 Example 2

The system of coupled second-order differential equations (refer to Frames 57

to 61)

1 (x) = 3f1(x) + 10f2(x)

2(x) = 2fi(x) + 4f2(x)
where fi(0) = 0, f2(0) = 1, f{(0) = 1 and f{(0) =
has the solution (refer to Frames 57 to 61)
Matrix algebra 143

fi(x) = 10cosx + g sinx + 10 cosh 2V2x + —=sinh 2V2x

9/2
2
f2(x) = —4cosx — =sinx + 5cosh2V2x + s=nh 2V2x
9 9/2

Because

1'(x) = 3fi(x) + 10f2(x)

2(x) = 2fi (x) + 4f2(x)
can be written in matrix form as

feneredDove) a
That is

wn = (0 (maa (5)
F’ (x) = AF(x)

and where F(0) = (ay)and F(0) = (a) = Gy

2
To solve the matrix equation we first need the eigenvalues and associated
eigenvectors of the matrix A. These have already been found in Frames S57 to
61 and they are
5
\ = —1 with corresponding eigenvector ( 3)

2
\ = 8 with corresponding eigenvector € )

The complete solution of F’ = AF is then

5 2
F(x) = (Pcosx + sins) ( a4+ (Reosh 2V2x + Ssinh 2v2x) (1 )

_ [5Pcosx + 5Qsinx + 2Rcosh 22x + 2S sinh 2V2x

~ \ _2Pcosx —2Qsinx + Rcosh 2V2x + Ssinh 2V2x
That is
144 Linear Algebra

1 00 fi(x) = SPcosx + SQsin x + 2Rcosh 2V2x + 2S sinh 2V2x

f2(x) = —2P cosx — 2Qsinx + Rcosh 2V2x + Ssinh 2V2x

Because

F(x) = oe
) (ee
and so
fi(x) = 5Pcosx+5Qsinx + 2R cosh 2V2x + 2S sinh 2V2x
fo(x) = —2P cos x — 2Q sinx + Rcosh2V2x + Ssinh 2V2x
Applying the boundary conditions, we find

FO)= (1) = (prea) mdPO=(4)= ("94vs)

5P+2R
= Es 1
sie and F(0) = (5 5Q4+4V25
) = oars

Because

The boundary conditions are f,(0) = 0, f2(0) = 1, f{(0) =1 and f5(0) =0,

ro=(})e(=f(Sarieeo))#8
therefore

ror=(3)=(=(Sfieagn))
This gives the two sets of simultaneous equations
5P+2R=0 5Q4+4V2S=1
and which have solution
—2P+R=1 ~2Q +2V25 =0

102 Ped/9 R= 519-~O-— 4/9 and $= 1/(9v2)

This gives the complete solution as

10 5. 10 ie *
hg — = COSx + g sinx + = cosh 2v/2x + GS ne 2V2x

4 on 2) oer
fal x)= 9 COSx — 5 Sin x + 5 cosh 2V2x + 9pqsinh 2v2x
Matrix algebra 145

Matrix transformation
y x-y plane v | u-y plane 103 |

If for every point Q (u, v) in the u-v plane there is a corresponding point P
(x, y) in the x-y plane, then there is a relationship between the two sets of
coordinates. In the simple case of scaling the coordinate where
u = ax and v = by
we have a linear transformation and we can combine these in matrix form
¥ _ fa 0\(x
v) \O bj\y

The matrix oe >) then provides the transformation between the

vector (>) in one set of coordinates and the vector 4 in the other

set of coordinates.
Similarly, if we solve the two equations for x and y, we have
1 1
tmeu and saad “i

: 4 arwe a)
the x-y
which allows us to transform back from the u-v plane coordinates to
plane coordinates.
Now for an example.
146 Linear Algebra

Example

If X= (*) = (4) with the transformation T = «9 Qi)determine

Ui by = TX and show the positions on the x-y and w-v planes.

In this case

(=(2 G)=(3)
u-v plane

104 cde (“1 a

There are severak ways of finding the inverse of a matrix. One method is as
follows.

T=
(2
oa1)
eT ey x)
PaallaG 1 0 CO
- — e-
©
oe

Sire (cates ia
So we have U=TX |. X=T'U
of Nara. OF 8
cis) ee ee Ba

Hence a vector as in the w-v plane transforms into 8 in the x-y

plane where (3) SS tse Coe

Matrix algebra
si ; := See.)

6-2) om

transforms into

Rotation of axes
A mote interesting case occurs with a degree of rotation between the two sets
of coordinate axes.

Let P be the point (x,y) in the x-y plane and the point (u,v) in the w-v plane.
Let 6 be the angle of rotation between the two systems. From the diagram we
can see that
x = ucosé —vsin#é
y =usin@
+ vcosé
(1)

‘ , x\ /cos@ —siné@\/u
In matrix form, this becomes (7) = ee:Q Ree ) fa

which enables us to transform from the w-v plane coordinates to the

corresponding x-y plane coordinates.
Make a note of this and then move on
 Tete eae ———

ut
cM

* i.
Linear Algebra
Sete &

106 If we solve equations (1) for u and v, we have

‘ xsin@ = usin@cos @—vsin?é

ycosé = usinécosé+v cos? 6
‘ ycos@ — xsin@ = v(cos? 6 + sin? @) =v
Also x cos 6 = ucos* 6 — vsin @cos 6
ysin @= usin? 6+ vsin@cosé
" xcosé+ysin6 = u(cos? @+ sin* 6) =u
So u=xcosé+ysiné
v = —xsiné+ycos@

and written in matrix form, this is............ ~

107 i\ a cosé@ sin@\/x

v}) \-—sin@ cosé)\y

So we have
(es = Boas sey eel
\y/ \sin@ cosé/\v
a) ( cos @ mon) (e)
and = :
Vv —sin@ cosé/\y
ie.X=TU and U=T'!X
where T is the matrix of transformation and the equations provide a linear
transformation between the two sets of coordinates.

Example

If the u-v plane axes rotate through 30° in an counterclockwise manner from
the x-y plane axes, determine the (u, v) coordinates of a point whose (x, y)
coordinates are x = 2, y = 3 in the x-y plane.
This is a straightforward application of the results above.
Matrix algebra 1490

u 34.3) 108
(") i ae BS a)

Because

Sr eet) eta
pyene 1/2 1@
-1/2 V3/2)\3
= ckon.69 w con
-14+3V3/2 1-60

As usual, the Program ends with the Summary, to be read in conjunction with
the Can You? checklist. Go back to the relevant part of the Program for any
points on which you are unsure. The Test exercise should then be
straightforward and the Further problems give valuable additional practice.

Bp Summary 4
1 Singular square matrix: |A| = O 109
Non-singular square matrix |A| # 0.
2 Rank of a matrix - order of the largest non-zero determinant that can be
formed from the elements of the matrix.
3 Elementary operations and equivalent matrices
Each of the following row operations on matrix A produces a row
equivalent matrix B where the order and rank of B are the same as those of
A. We write A ~ B.
(1) Interchanging two rows
(2) Multiplying each element of a row by the same non-zero scalar
quantity
(3) Adding or subtracting corresponding elements from those of another
row.
These operations are called elementary row operations. There is a
corresponding set of three elementary column operations that can be used
to form column equivalent matrices.
4 Sg gtSe
Linear Algebra
Nee =i oti tt

Consistency of a set of n equations in n unknowns with coefficient matrix

A and augmented matrix Ap.
(a) Consistent if rank of A = rank of Ap
(b) Inconsistent if rank of A < rank of Ap.

Uniqueness of solutions — n equations with n unknowns.

(a) rank of A= rank of Ap =n unique solutions
(b) rank of A= rank of Ay =m <n __ infinite number of solutions
(c) rank of A < rank of Ap no solution
Solution of sets of equations
(a) Inverse matrix method Ax=b; x=A'b
To find A!
(1) evaluate |A|
(2) form C, the matrix of cofactors of A
(3) write C!, the transpose of A
1
(4) Av! = TA ib

(b) Row transformation method Ax=b; Ax=Ib

(1) form the combined coefficient matrix ba!
(2) row transformations to convert to LeAD -
(3) then solve x = A'b.
(c) Gaussian elimination method Ax=b
(1) form augmented matrix [A'b]
(2) operate on rows to convert to [U!b] where U is the upper-
triangular matrix.
(3) expand from bottom row to obtain x.

(d) Triangular decomposition method Ax=b

Write A as the product of upper and lower triangular matrices.
A=LU, L(Ux)=b. PutUx=y ...Lly=b
(1) solve Ly = b to obtain y
(2) solve Ux =y to obtain x.
Eigenvalues and eigenvectors Ax = x
Sets of equations of form Ax = Ax, where A = coefficient matrix, x =
column matrix, \ = scalar quantity.
Equations become (A — \I)x = 0.
For non-trivial solutions, |A— AI| =0 is the characteristic equation and
gives values of \_ i.e. the eigenvalues.
Substitution of each eigenvalue gives a corresponding eigenvector.
Cayley—Hamilton theorem
Every square matrix satisfies its own characteristic equation.

@
Matrix algebra 151

9 Solving systems of first-order ordinary differential equations

To solve the system of coupled first-order differential equations
F(x) = AF(x)

(a) Find the eigenvalues and eigenvectors of matrix A and construct the
modal matrix M and the diagonal spectral matrix $

(b) Solve the equation G’(x) = SG(x)

(c) Apply F(x) = MG(x) to find F(x).
10 Diagonalization of a matrix
Modal matrix of A
IfA has distinct eigenvalues M = (x1, X2,..., X,), where x), Xo; ...) Xs ate
eigenvectors of A, then M~'AM = S$ where § is the spectral matrix of A
Apes: 020
0 AAs eae 0
and $=

Oi opine
A1, A2, -.+, An are the eigenvalues of A.
11 Solving systems of second-order ordinary differential equations
To solve an equation of the form
PF’ (x) = AF(x)

(a) Find the eigenvalues \j, A2,..., An of A

(b) Assuming the eigenvectors are all distinct, find the associated
eigenvectors C;, C2,..., C,
(c) Write the solution of the equation as
n

F(x) = (me 4 bre¥™*) C,

T=1

and use the boundary conditions to find the values of a, and b, for
(Pal erate Ue
12. Matrix transformation
(a) U = TX, where T is a transformation matrix, transforms a vector in
the x-y plane to a corresponding vector in the u-v plane. Similarly,
X = T-'U converts a vector in the u-v plane to a corresponding vector
in the x-y plane.
(b) Rotation of axes
u =( cos 6 aoaley
@ ~ \—siné cosé/\y

(*) a (am ea
y/ \sin@ cos@/\v
152 Linear Algebra

4 Can You?
|110 Checklist 4 | |
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames

Determine whether a matrix is singular or non-singular?

Yes No
»@
Determine the rank of a matrix?
Yes ei No
2»
Determine the consistency of a set of linear equations and ~
hence demonstrate the uniqueness of their solution?
Yes No
>»
Obtain the solution of a set of simultaneous linear equations
by using matrix inversion, by row transformation, by Gaussian
elimination and by triangular decomposition?
Yes L] C L] | | No
»©
Obtain the eigenvalues and corresponding eigenvectors of a
square matrix?
Yes L] |_| a No
v
Demonstrate the validity of the Cayley-Hamilton theorem?
Yes aL | No
Solve systems of first-order ordinary differential equations
using eigenvalue and eigenvector methods?
Yes LJ ene bs No
Do
Construct the modal matrix from the eigenvectors of a matrix
and the spectral matrix from the eigenvalues? (20)
»(28)
Yes iS O O i L] No
Solve systems of second-order ordinary differential equations
using diagonalization?
Vese: A a a | No
(HD
»(i)
Use matrices to represent transformations between coordinate
systems?
Yes | | | No
CB»)(aH)
Matrix algebra cog fT |

&, Test exercise 4

1 Determine the rank of A and of Ay for the following sets of equations and 111
hence determine the nature of the solutions. Do not solve the equations.
(a) X1 + 3x2 — 2x3 = 6 (b) x3 + 2x2 -—4x3= 3
4x; + 5x2 + 2x3 =3 X; + 2x2 + 3x3 = —4
X1 + 3x2 + 4x3 =7 2X, +4x%2+ x3 =-3.
pps Volare 2 4
2 If Ax=b where A=|3 5 -4] and b=|10], determine A’
12 -3 9
and hence solve the set of equations.
3 Given that 3x; + 2x2+ x3=1
Xi — X2+3x3=5

2X1 + 5x2 — 2x3 = 0

apply the method of row transformation to obtain the value of

X1, X2, X3.

4 By the method of Gaussian elimination, solve the equations Ax=b,

1) =2,-—4 —3
where A= | 2 1¥—3 | andb= rd ie
1 3 Zz 5
1 -—2 1 7
nn If Ax =b where A= ( 1 2) and b= (2). express A as the
5 3 3 5
product A = LU where L and U are lower and upper-triangular
matrices and hence determine the values of x;, X2, X3.

6 Determine the eigenvalues and corresponding eigenvectors = Ax

of Ax
bbe be 0
where A = 12, Lele
—2 lL =]

7 If x, and x2 are eigenvectors of Ax = Ax where A = €4 21)determine

(a) M = (X,X2)
(b) M!
(c) M-'AM.
io Solve the system of first-order differential equations
f(x) = Sfi(x) — 2f2(x) where [1 (0) = —3 and f2(0) = 2.
f(x) = —filx) + 4f2(%)
| 154 ee Linear Algebra

9 Solve the system of second-order differential equations

))
+ 6f2(x E. Rs
1(X) = fi(X
3A (x) 26) ee Ot ee
10 (a) Determine the vector in the u-v plane formed by U= TX, where the

transformation matrix is T= im Za and X= cs is a vector

in the x-y plane.
(b) The coordinate axes in the x-y plane and in the u-v plane have the same
origin O, but OU is inclined to OX at an angle of 60° in an

counterclockwise manner. Transform a vector X= ta in the x-y

plane into the corresponding vector in the u-v plane. —

Further problems 4

112 1 If Ax=b where

5
A= [ aoe 2] and
2 3 6
»-( a determine A’!
ms Sl ©
and hence solve the set of equations.
2 Apply the method of row transformation to solve the following sets of
equations.
(a) xX; —3x2-2x4= 8 (b) x, —3x2+ 2x3 =8
2X, +2xX2+ x3= 4 2X1
— X2+ x3=9
3X1 — 4X2 + 2x3 = —3 3X1 + 2x2 + 3x3 = 5.

3 Solve the following sets of equations by Gaussian elimination.

(a) xX, —2x2-— x3+3x%,=4
2X1 + Xo + x3 —4x4=3

3X1 — X2 —2x3+2x4=6
X14+3xX2—-— x3+ x%4=8

(b) 2x; + 3x2 — 2x3 + 2x, =2

4x, + 2x2 — 3x3 — x4 =6

X1— X2+4x3 — 2x4 =7

3X, +2xX2+ X3- X4=5

(Cc) X1+2x24+5x3+ x =4
| 3x1 — 4X2 + 3x3 — 2x4 =7
4x, + 3x2 +2x3-— x4=1
X1 — 2X2
— 4x3 — xX4=2.

2
Matrix algebra Peay aD
i eo: 6 ee RS

4 Using the method of triangular decomposition, solve the following sets of

equations.
1 4 -1 x; —2
(a) (: 2 : X2 = —]
7 -3 +9 x3 —18
1 -2 3 xy —2
(b) [: 1 -5 ks | = iWe
6 -3 y) x3 CIPD
Papert vor x, ae
“Gs eee ene Gan PY
ease eo ila |= | “4
A iene 4) \ x, 10
5 If Ax = Ax, determine the eigenvalues and corresponding eigenvectors in each
of the following cases.
a3 2 5
@ A=(7 s) ) A=(4 Za

—6 5 -5 9
() A=( 4 >) (a) A=( 1 3)
er 5 =o 81
fe, aati 3s I | (f) a-( 1 |
reall 0 Jini 3. O*1
6 4 10 -8
3 mae( 2 2 }
7 E| gk aba.
6 Solve each of the following systems of first-order differential equations.
(a) fi(x) = 2fi(x) — Sfa(x)
fa(x) = fix) — 4fa(x)
where f;(0) = 1 and f2(0) =

(b) f(x) = osx) + 9f2(x)

fa(x) = fi(®) + 3f2(%)
where f1(0) = 0 and f2(0) = —2

(c) Ai(x) = Sfi(x) — 6f2(x) + fa(%)

f2(x) = fi(x) + fax)
F(x) = 3fi(x) + f(x)
where f1(0) = 1, f2(0) = 0 and f3(0) =

(d) fi(x) = 4fi(x) + 10fa(x)—8f(x)

fe fi(x) + 2f2(x) + f(x)
f(x) = —fi(®) + 2f2(x) + 3A(%)
where f(0) = 4, f2(0) = —2 and f3(0) = —
 Linear Algebra

1 cy ie)
lias : 10 3) determine the three eigenvalues \;, A2, A3 of A
0 -3 9
~9 1 gt
and verify that if M=| 3 2 4] then M'‘'AM=S, where § is a
1 3 -3
diagonal matrix with elements 1, 2, A3.

8 10° 7
Invert the matrix A=|5 9 4 | and hence solve the equations
OFT ia

8h +10bh+71,= O
5h + 9+
413 = -9 =
97,+11,+4+88= 1.

T2503 —2 6 -4
If A=|4 6 7] and B={-1 -6 5S], verify that AB=kI
esse 2 2 -2
where I is a unit matrix and k is‘a constant. Hence solve the equations
X1 + 2x2
+ 3x3 =2

4x; + 6x2 + 7x3 =2

5X1 + 8x2 + 9x3 = 3.

10 Solve each of the following systems of second-order differential equations.

(a) fi'(*) = 4%) + 3h)
2(X) = 2fi (x) + Sfa(x)
where fi(0) = 0, f2(0) = 1, f{(0) = 4 and f5(0) =
(b) fi'(x) = —6fi(x) + Sfa(x)
2(X)= 4fi(x) + 2f2(x)
where [1(0) = 0, f2(0) = 1, f{(0) = 1 and f3(0) =
(Cc) fi (X) = 2A(%) + 7falx)
2(x) = fil) + 3fa(x) + f(x)
3 (x) = Sfi(x) + 8f3(x)
where fi(0) = 1, f2(0) = 1, fa(0) =, f{(0) = 0, ff(0) = 0
and f3(0) = 1
(d) f'(%) = -3fi(%) + fa(x)
2(x) = 4fi(x) + Sfa(x) + 3f3(x)
3(x) = fi(x) + 2fa(x) + fa)
where fi (0) = 1, f2(0)= 1, f3(0)=0, f[(0)=0, (0) =0, (0) = 1.
Vectors
Learning outcomes
When you have completed this Program you will be able to:
Define a vector
Represent a vector by a directed straight line
Add vectors
Write a vector in terms of component vectors
Write a vector in terms of component unit vectors
Set up a coordinate system for representing vectors
Obtain the direction cosines of a vector
158 Linear Algebra

Introduction: scalar and vector quantities

Physical quantities can be divided into two main groups, scalar quantities and
vector quantities.
(a) A scalar quantity is one that is defined completely by a single number with
appropriate units, e.g. length, area, volume, mass, time, etc. Once the
units are stated, the quantity is denoted entirely by its size or magnitude.
(b) A vector quantity is defined completely when we know not only its
magnitude (with units) but also the direction in which it operates, e.g.
force, velocity, acceleration. A vector quantity necessarily involves
direction as well as magnitude.
So (a) a speed of 10 km/h is a scalar quantity, but Bs
(b) a velocity of 10 km/h due north’ isa............ quantity.

2
A force F acting at a point P is a vector
quantity, since to define it completely we
must give:
(a) its magnitude, and also
(ER) aS teresaache

3 %

So that:
(a) A temperature of 100°C is a....../..... quantity.
(b) An acceleration of 9.8 m/s” vertically downwards isa............ quantity.
(Cc). Tne weight ofa 7 Ke mass 1580 oro ee quantity.
(Ui The sui Of LOU aSudies. ce eee -G: quantity.
(e) A north-easterly wind of 20 knots isa............ quantity.

4 (a) scalar (b) vector (c) vector (d) scalar (e) vector

Since, in (b), (c) and (e) the complete description of the quantity includes not
only its magnitude, but also its ............

fe
Move on to Frame 6
Vectors 159

Vector representation
A vector quantity can be represented graphically by a line, drawn so that: 6
(a) the length of the line denotes the magnitude of the quantity, according to
some stated vector scale
(b) the direction of the line denotes the direction in which the vector quantity
acts. The sense of the direction is indicated by an arrowhead.
e.g. A horizontal force of 35 N acting to the right, would be indicated by a line
and if the chosen vector scale were 1cm=10N, the line would be
rr eee cm long

7
The vector quantity AB is referred to as B
AB ora
The magnitude of the vector quantity is
written |AB|, or |a|, or simply AB or a.

Note that BA would represent a vector quantity of the same magnitude but
with opposite sense.

On to Frame 8

Two equal vectors 8

If two vectors, a and b, are said to be equal,
they have the same magnitude and the same
direction. a b
If a=b, then
(a) a = b (magnitudes equal)
(b) the direction of a = direction of b, i.e. the two vectors are parallel and in
the same sense.
Similarly, if two vectors a and b are such that b = —a, what can we say about:
(a) their magnitudes,
(b) their directions?
160 Linear Algebra

9 (a) Magnitudes are equal

(b) The vectors are parallel but opposite in sense

i.e. if b = —a, then

|10 Types of vectors

(a) A position vector AB occurs when the point A is fixed.
(b) A line vector is such that it can slide along its line of action, e.g. a
mechanical force acting on a body.
(c) A free vector is not restricted in any way. It is completely defined by its
magnitude and direction and can be drawn as any one of a set of equal-
length parallel lines.
Most of the vectors we shall consider will be free vectors
So on now to Frame 11

|11° Addition of vectors

The sum of two vectors, AB and BC, is defined as the single or equivalent or
resultant vector AC
i.e. AB+BC=AC
or a+b=c

To find the sum of two vectors a and b then, we draw them as a chain, starting
the second where the first ends: the sum c is given by the single vector joining
the start of the first to the end of the second.
e.g. if p =a force of 40 N, acting in the direction due east
q = a force of 30 N, acting in the direction due north
then the magnitude of the vector sum r of these will forces will be

12 Because
72 =p? dlg?

= 1600 + 900 = 2500

fin 25 OOS 0NN

Vectors a 464

The sum of a number of vectors a+b +c+d+...

(a) Draw the vectors as a chain.

(b) Then:

a+b=AC
AC+c=AD
a+b+c=AD
AD+d=AE
a+bic+d=AE
i.e. the sum of all vectors, a,b,c,d, is given by the single vector joining the
start of the first to the end of the last — in this case, AE. This follows directly
from our previous definition of the sum of two vectors.
R Similarly:
Q T PO'4/OR
2 RS STS ee.

13
Now suppose that in another case, we draw the vector diagram to find the sum
of a,b,c,d,e, and discover that the resulting diagram is, in fact, a closed
figure.
D What is the sum of the vectors
d a,b,c, d,e in this case?
E Cc

Cc
{=}
b

A a B

Think carefully and when you have decided, move on to Frame 14

Sum of the vectors = 0 14

Because we said in the previous case, that the vector sum was given by the
single equivalent vector joining the beginning of the first vector to the end of
the last.
But, if the vector diagram is a closed figure, the end of the last vector
coincides with the beginning of the first, so that the resultant sum is a vector
with no magnitude.
162 Linear Algebra

Now for some examples:

Find the vector sum AB+ BC+ CD + DE+ EF.

Without drawing a diagram, we can see that the vectors are arranged in a
chain, each beginning where the previous one left off. The sum is therefore
given by the vector joining the beginning of the first vector to the end of the
last.
-, Sum = AF
In the same way:
AK >) KL LP=P PO 58a oo

15
Right. Now what about this one?
Find the sum of AB—CB+CD-—ED
We must beware of the negative vectors. Remember that —CB = BC, i.e. the
same magnitude and direction but in the opposite sense.
Also —ED = DE
. AB—CB+CD-—ED=AB+BC+CD+DE
= Ar
Now you do this one:
Find the vector sum AB+BC—DC—AD
When you have the result, move on to Frame 16

|16 o
Because

AB+BC —DC= AD = AB+BC + CD+ DA

and the lettering indicates that the end of the last vector coincides with the
beginning of the first. The vector diagram is thus a closed figure and therefore
the sum of the vectors is 0.
Now here are some for you to do:
G)\PO-SOR ERS 51 a eee
(DAG OL = MD cee ee
(C) -GH A Ale KeeGG = ea ee
(d) AB+BC+CD+DB=............

When you have finished all four, check with the results in the next frame

a
Vectors 163

Here are the results: 17

(a) PQ+QR+
RS + ST = PT
(b) AC+ CL—ML=AC+CL+LIM
= AM
(c) GH+HJ+JK+KL+LG=0
[Since the end of the last vector
coincides with the beginning of the
(d) AB+ BC +CD + DB = AB
The last three vectors form a closed D
figure and therefore the sum of
these three vectors is zero, leaving C
only AB to be considered.
A B

Now on to Frame 18

Components of a given vector

Just as AB + BC + CD + DE can be replaced by AE, so any single vector PT can
be replaced by any number of component vectors so long as they form a chain
in the vector diagram, beginning at P and ending at T.
eg A ran
be ak Duc PT=a+bic+d

Example 1
ABCD is a quadrilateral, with G and H the mid-points of DA and BC
respectively. Show that AB + DC = 2GH. oe
We can replace vector AB by any chain of
B vectors so long as they start at A and end at B
e.g. we could say
Q 7
AB = AG + GH + HB
164 Linear Algebra

19 , DC = DG + GH + HC
So we have:

- AB+ DC =AG+ GH+HB+DG+GH+HC

= 2GH + (AG + DG) + (HB + HC)
Now, G is the mid-point of AD. Therefore, vectors AG and DG are equal in
length but opposite in sense.
.. DG = —AG
Similarly HC = —HB
-. KB+DC = 2GH + (AG — AG) + (HB — BB)
= 2GH
Next frame

20 Example 2
Points L, M, N are mid-points of the sides AB, BC, CA of the triangle ABC.
Show that: ‘
(a) AB+BC+CA=0
(b) 2AB + 3BC + CA = 2LC
(c) AM+BN+CL=0
A (a) Wecan dispose of the first part straight away
without any trouble. We can see from the vector
diagram that AB + BC + CA = O since these three
vectOrs Foray allt ee

21 |closed figure |
Now for part (b):

To show that 2AB + 3BC + CA = 2LC

Vectors 4165

From the figure:

AB=2AL; BC=BL+LCc; CA=CL+LA
*, 2AB+ 3BC + CA = 4AL+ 3BL+ 3LC+CL+LA
Now BL=—AL; CL=-LC; LA=-—AL
Substituting these in the previous line, gives
2AB + 3BC + CA =

2LC 22 |
Because

2AB + 3BC + CA = 4AL + 3BL+ 3LC+CL+LA

= 4Al, —3AL 4 3LC — LC — AL
= 4AL — 4AL. + 3LC — LC
= 2LC
Now part (c):

To prove that AM + BN+CL=0

From the figure in Frame 21, we can say:
AM = AB + BM
BN = BC+ CN
Similarly, Clie...) :.daadae

Che GA 4AL 23
So AM + BN + CL = AB+ BM + BC+CN+CA+AL
= (AB + BC + CA) + (BM + CN + AL)
= (AB + BC+CA) +5 (BC+ CA+ AB)

Finish it off

AM +BN+CL=0 24
a Thy arta ee ee ne Aopen gga
Because AM + BN + CL = (AB + BC + CA) +5 (BC + CA + AB)

Now AB+BC+CA is a closed figure .. Vector sum = 0

and BC + CA +AB is a closed figure .. Vector sum = 0
., AM+BN+CL=0
Here is another. >
i. eee Linear Algebra

Example 3

ABCD is a quadrilateral in which P and Q are the mid-points of the diagonals

AC and BD respectively.
Show that AB+AD+CB+CD=4PQ
First, just draw the figure.
Then move on to Frame 25

A D

To prove that AB + AD + CB+ CD = 4PQ

Taking the vectors on the left-hand side, one at a time, we can write:
AB = AP + PQ + QB
AD = AP+PQ+QD

CB = CP+PQ+
QB; CD =CP+PQ+QD

Adding all four lines together, we have:

AB + AD + CB4 CD = 4PQ + 2AP + 2CP + 2QB + 20D
= 4PQ + 2(AP + CP) + 2(QB + QD)
Now what can we say about (AP + CP)?

27 AP+CP=0

Because P is the mid-point of AC .. AP = PC

”. CP = —PC = —AP
“. AP +CP = AP—AP=0.
In the same way, (QB+ QD) =............

28
Since Q is the mid-point of BD... QD = —QB
“. QB+ QD = QB- QB=0
~, AB+AD+CB+CD
= 4PQ +040 = 4PQ

7)
Vectors 167

Here is one more.

29
Example 4
Prove by vectors that the line joining the mid-points of two sides of a triangle
is parallel to the third side and half its length.
Let D and E by the mid-points of AB A
and AC respectively.
We have DE = DA + AE

B Cc
Now express DA and AE in terms of BA and AC respectively and see if you can
get the required results.
Then on to Frame 30

Here is the working. Check through it. 30

DE = DA + AE

Se (BA-+AC
taleg ee 25oN
SORA+<AC
es be
DE = —7 BC

‘. DE is half the magnitude (length) of BC and acts in the same direction.

i.e. DE and BC are parallel.
Now for the next section of the work: move on to Frame 31

Components of a vector in terms of unit vectors 31

y The vector OP is defined by its
P magnitude (r) and its direction (6).
2 It could also be defined by its two
b components in the Ox and Oy
O = - directions.

i.e. OP is equivalent to a vector a in the Ox direction + a vector b in the

Oy direction.
i.e. OP = a (along Ox) + b (along Oy)
If we now define i to be a unit vector in the Ox direction,
then a=ai
Similarly, if we define j to be a unit vector in the Oy direction,
then b=bj
So that the vector OP can be written as:
r=ai-+ bj
where i and j are unit vectors in the Ox and Oy directions.
ence
Ee
168 Linear Algebra

32 Let z; = 2i+4j and zz = Si+2j

To find z; + Z2, draw the two vectors in a chain.

Z1+Z2 = OB= (2+5)i+ (44 2)j = 7i+ 6j

i.e. total up the vector components along Ox,

and total up the vector components along Oy
Of course, we can do this without a diagram:
If z; = 31+2j and z2 = 4i+3j
Z14+2Z2 = 31+ 2j+ 414 3j
= 7i+Sj
And in much the same way, Z2 — Z; =............

33
Because Zz — Z = (444+ 35) — (i+ 2j)
= 41+ 3j — 3i - 2j
=li+lj
=i+j
Similarly, if z; = Si — 2j; Z2 = 31+ 3j; z3 = 41 - 1j
then (a) Zi Loc 23 — eee
and (b) Zi, AZo73 Severs

When you have the results, move on to Frame 34

Vectors
169

(a) 12i (bo) —-28-45 34

Here is the working:
(a) Z1 +22 +23 = Si— 25 + 314 35+ 41-1;
=(5+3+4)i+ (3 -2-1)j=12i
(b) Z; — Z2 — Z3 = (Si — 2j) — (34 + 3§) — (41 — 1j)
= (S—3-4)i+(-2-3+41)j=-2i-4j
Now this one.
If OA = 3i + Sj and OB = 5i — 2j, find AB.
As usual, a diagram will help. Here it is:
First of all, from the diagram, write
down a relationship between the
vectors. Then express them in terms
of the unit vectors.

AB = 2i — 7j 35
Because we have

OA + AB = OB (from diagram)
* AB =OB-—OA
= (Si — 2j) — (3i + Sj) = 2i - 7j
On to Frame 36

Vectors in space
z The axes of reference are defined by 36 |
the ‘right-hand’ rule.
Ox, Oy, Oz form a right-handed set
if rotation from Ox to Oy takes a
right-handed corkscrew action
along the positive direction of Oz.

Similarly, rotation from Oy to Oz gives right-hand corkscrew action along the

positive direction of ............
170 Linear Algebra

37 :
z Vector OP is defined by its comporients
5 eee espo a along Ox
cement ! b along Oy
c along Oz

Livy

Let i = unit vector in Ox direction

j = unit vector in Oy direction
k = unit vector in Oz direction

Then OP = ai+bj+ck
Also OL? = a?.+ b* and OP? = OL? + 7
OP* =a? +b? +c?
So, if r = ai + bj + ck, then r= Va?
+ b? +c?
This gives us an easy way of finding the magnitude of a vector expressed in
terms of the unit vectors.
Now you can do this one:
IP PO = 4143j 4 2k, them (PO|=.........-..

38 [PQ| = V29 = 5.385

Because
PQ = 4i + 3j + 2k

[PQ| = 42 + 32+22
= V16+9+4
— /29 - -
== EE to

Now move on to Frame 39

as
Vectors 171

Direction cosines
The direction of a vector in three dimensions is determined by the angles 39
which the vector makes with the three axes of reference.

Let OP = r= ai+bj+ck
Then
a
7 = COS a: * @=f1Tcosa

b
~ = Cos 8 b=rcos3
Cc
7 = COS c=Trcosy

Also a? + b? + c# =r?
. Pcos?a +7 cos? 3 +r? cos?
y = r*
cos? a + cos” 3 + cos? 7 = 1
If |= cosa
m= cos 3

n=cosy then 2 +m* +n? =1

Note: {I,m,n| written in square brackets are called the direction cosines of the
vector OP and are the values of the cosines of the angles which the vector
makes with the three axes of reference.
So for the vector r = ai+ bj + ck
a b c
l=-; m=-; hess and, of course r= Va? + b? + ¢?
r r
So, with that in mind, find the direction cosines [/,m,n] of the vector
r= 3i-— 2j+ 6k
Then to Frame 40

r = 3i— 2j+6k 40
a=3, b=-2, c=6, Tr=V9+4+36
r= V49 =7

aes eee: ee
sags’ Be ~~ Fao?
Just as easy as that!
On to the next frame
ore ae Linear Algebra

la} Summary
41 1 A scalar quantity has magnitude only; a vector quantity has both
magnitude and direction.
2 The axes of reference, Ox, Oy, Oz, are chosen so that they form a right-
handed set. The symbols i, j, kK denote unit vectors in the directions Ox,
Oy, Oz, respectively.

If OP = ai+ bj+ck, then |OP| =r= Va2 +b? +c?

3 The direction cosines |l,m,n| are the cosines of the angles between the
vector and the axes Ox, Oy, Oz respectively.
a b ‘e J
For any vector: 1=—,m=—,n=—; andl’ +m’ +n’ =1

Now you are ready for the Can You? checklist and Test exercise.
So off you go

&% Can You?

|42 Checklist 5
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames
e Define a vector? o GD
Yes RA del ba Lad No
e Represent a vector by a directed straight line? Ce) to
Yessy oule Fh (pia ll Pn toe ng
e Add vectors? to
Yes ner ear ogee ee ed eNO
e Write a vector in terms of component vectors? to (0)
Ys cL) J) (> Te l No
e Write a vector in terms of component unit vectors? to (35)
Yes Sd ae ee lame 0 |
e Set up a coordinate system for representing vectors? to
SVOSS* potleitoe Alt] oad eer aoe eee NO
e Obtain the direction cosines of a vector? to
Yes LJ L] L] LJ fis] No

es
Vectors 173

& Test exercise 5

Take your time: the problems are all straightforward so avoid careless slips. 43 |
Diagrams often help where appropriate.
1 If OA = 4i + 3j, OB = 6i — 2j, OC = 2i — j, find AB, BC and CA, and deduce the
lengths of the sides of the triangle ABC.
Show that the addition of two vectors a and b is a commutative operation. That
is:
a+b=bia
Show that the addition of two vectors a, b and c is an associative operation.
That is:
(a+b)+c=a+(bic)
Find the direction cosines of the vector joining the two points (4, 2, 2) and
(740,14).

Further problems 5
The centroid of the triangle OAB is denoted by G. If O is the origin and 44
OA = 4i + 3j, OB = 6% — j, find OG in terms of the unit vectors, i and j.
Find the direction cosines of the vectors whose direction ratios are (3, 4, 5) and
(1, 2, —3). Hence find the angle between the two vectors.
Find the modulus and the direction cosines of each of the vectors 3i + 7j — 4k,
i — Sj — 8k and 6i — 2j + 12k. Find also the modulus and the direction cosines
of their sum.
If AB and BC are two adjacent sides of a regular hexagon, what are the other
sides?
Three vectors are directed along the diagonals of three adjacent faces of a cube
from the same vertex. Show that their sum is twice the vector from the same
vertex directed along the diagonal of the cube.
Show that the diagonals of a parallelogram bisect each other.
A point describes a circle, counterclockwise about the origin in the x-y plane
taking 24 seconds to complete one revolution. If the initial position vector of
the point is i, find the position vector of the point after 2, 6, and 10 seconds.

Given that
r= ai+ bj+ck
Show that the magnitude of r is given as:
r=V@+bh?+C?
Show that if vectors and do not lie in the same direction then:
6b = 0 implies that
aa+ a= 8=0
Vector spaces
Learning outcomes
When you have completed this Program you will be able to:
Define a vector space and develop proofs in general terms
Demonstrate how different mathematical objects can form vector
spaces
e Define a subspace of a vector space
Manipulate ordered n-tuples of real numbers within the context of a
vector space ‘
Define a linear combination of vectors
Define a spanning set of a vector space
Distinguish between linear dependence and linear independence
Apply the test for linear independence
Define the criteria for a basis
Deduce the dimension of a vector space
Vector spaces
175

Introduction
The two operations of the addition of a pair of two-dimensional geometric
vectors and the multiplication by a scalar of a single two-dimensional
geometric vector have certain properties that are retained when the operations
are applied to mathematical objects other than two-dimensional geometric
vectors. It is this commonality of properties that forms the foundation of a
vector space. In what follows, these properties will first be considered from
the viewpoint of two-dimensional geometric vectors before moving on to look
at other mathematical objects.
Next frame

Addition of two-dimensional geometric

vectors
Given the pair of two-dimensional geometric vectors: 2 |
a= 2i+3jand b= 4i+ Sj
their sum is defined by adding their respective components as:
a+b = 2i+ 3j+ 4i+ Sj
= (2+ 4)i+(34+5)j
= 61 + 8j
and the result of adding the pair of two-dimensional geometric vectors
together is another y vane. avg
The answer is in the next frame

two-dimensional geometric vector 3

Because the addition of any pair of two-dimensiona! geometric vectors always

results in another two-dimensional geometric vector, it is said that the
collection of all two-dimensional geometric vectors is closed under addition.
Move to the next frame
176 Linear Algebra

|4 Commutativity _
If a and b stand for any pair of two-dimensional geometric vectors, then:
a+b=b+a
The order in which they are added makes no difference. This property is called
commutativity — we say that the addition of two-dimensional geometric vectors
is a commutative operation. A simple example will suffice to demonstrate
this.
Let a = 2i + 3j and b = 4i-+ Sj so that:

The answer is in the next frame

5
Because:

a+b= (21+ 3j) + (41+ Sj)

= (2+4)i+ (3+5)j
= 61+ 8j

and

b + a= (41+ Sj) + (21 + 39)

= (44+2)i+ (5+3)j
= 61 + 8f
=a+b

Move to the next frame

|6 Associativity
If a, b and c stand for any three two-dimensional geometric vectors, then:
a+(b+c)=(a+b)4+c
The order in which they are associated when they are added makes no
difference. This property is called associativity - we say that the addition of
two-dimensional geometric vectors is an associative operation. Again, a
simple example will suffice to demonstrate this.
Let a=i— 2j, b = —3i+ 7j, and c = 41 + 2j, then: -
as. (DC): ne. ee and (a+b)+c=............
The answer is in the next frame
Vector spaces 177

Because:

a+ (b+ c) =i-— 2j + (—3i + 7j + 44 + 2j)

=1-2j+ (+94)
= 21+7j

and

(a+b) +c = (i— 2j — 3i+ 7j) + 414 2j

II —2i+
Sj+ 4i+4 2j
2i + 7j
=a+(b+c)
Move to the next frame

Identity element for addition 8 |

A two-dimensional geometric vector exists that is denoted by 0 with the
property that for any two-dimensional geometric vector a:
a+O=a

The two-dimensional geometric vector that we denote by 0 with this property

is given in component form as............
Can you think what it is? The answer is in the next frame

9|
Because, for example, if a = 2i + 3j, then:
a+0 = (2i+ 3j) + (Oi + Oj)
= 21+ 3j
=a
The two-dimensional geometric vector 0 — the zero geometric vector — is called
the identity element for addition because when it is added to any two-
dimensional geometric vector it leaves the two-dimensional geometric vector
unchanged.
Move to the next frame
178 Linear Algebra

10 Finally, for each two-dimensional geometric vector a there is another two-

dimensional geometric vector which is denoted by —a and is such that:
a+(—a)=0
For example, given the two-dimensional geometric vector:
a—6i-—9] then = A=)...
-- 4...
The answer is quite straightforward and is in the next frame

Because if: z
a = 8i — 9j and a+ (—a) = O where —a = aii + aj, then
a+ (—a) = 81 — 9j + (a1 + aj)
= (8+ a1)i+ (a2 — 9)j
= Oi + Oj
So that:
8+a,=0 giving a;=-—8 and a,—9=0 giving a+2=9
Therefore —a = —8i + 9j
In summary, two-dimensional geometric vectors are closed under addition
and the addition of two-dimensional geometric vectors has the following
properties: z
a+b=b+ia addition is commutative
a+(b+c)=(a+b)+c addition is associative
There is a two-dimensional geometric vector that is denoted by 0 with the
property that, for any two-dimensional geometric vector a:
a+O=a

and for every two-dimensional geometric vector a there is another two-

dimensional geometric vector denoted by —a where:
a+(—a)=0
Now we shall look at another operation, namely, scalar multiplication.
Let’s move on
Vector spaces
179°

Scalar multiplication of a two-dimensional

geometric vector
If a two-dimensional geometric vector is multiplied by a number (it is called a
scalar to distinguish it from the geometric vector), then each component is 12
multiplied by that number. For example:

If a = 2i+ 3j, then

Sa = 5(2i
+ 3j)
= (5 x 2)i+(5S x 3)j
= 10i+ 15j
=.
and the result of multiplying a two-dimensional geometric vector by a scalar is
Enothetn Os, 8.2.0:
Next frame

two-dimensional geometric vector As

Again, we say that the collection of all two-dimensional geometric vectors is closed
under scalar multiplication.
Next frame

Properties 14
Given the two scalars (that is numbers) a and 3, then the scalar multiplication
of a two-dimensional geometric vector has the following properties:
a(a+b) = aa+ab
(a+ P)a=aa+ Ba
(a3) = a(a)
la=a
Don’t be put off by this jumble of equations, let’s look at each one in turn —- we
have plenty of time to see what exactly is going on here.
Move to the next frame
GABOs ee: Linear Algebra

|15 a(a+ b)=aa + ab,

As an example, choose a = 3, a = 2i — Sj, and b = 7i + 4j, then the left-hand
side of the equation is:
a(a+ b) = 3((2i — Sj) + (7i+ 4)
= 3[9i — jj
= 27i — 3j

and the right-hand side is:

aa + ab = 3(2i — 5j) + 3(7i+ 45)
= —6i — 15j + 21i+ 12;
= 27i — 3j =
=a(a+b)
Try one yourself. Let’s choose a = —7, a= 3i + 8j, and b = —Si + 2j. Then:
eit eS |) eer eens
The answer is in the next frame

16
Because:
a(a +b) = —7((3i + 8j) + (—Si + 2j)|
= —7[—2i + 10§]
= 14i — 70j
And again, with a = —7, a= 31+ 8j, and b = —Si+ 2 j:

Again, the answer is in the next frame

17
Because:
aa + ab = —7(3i + 8j) + (—Si + 25)
= —21i — 56j + 35i - 14j
= 141 — 70j
=a(a+b)
So you have shown that with a = —7, a = 3i+ 8j, and b = —Si + 2j, then
a(a+b) =aa+ab
Vector spaces
181

The challenge now is to prove that this equation is true for any value of a and
any pair of two-dimensional geometric vectors a and b. When constructing a
proof we must ensure that each step made can be justified by known rules of
algebraic manipulation. To this end each step will be annotated with the
appropriate rule.

Move to the next frame

Proof that a(a + b) = aa + ab 18

Writing the pair of two-dimensional geometric vectors in their most general
form as a = ai + a2j and b = byi + baj where aj, a2, b; and bz are scalars, then
if a is another scalar:
a(a +b)
= a|(a,i + aaj) + (bii+ boj)| writing a and b in their component form
= a|(a; + by )i + (az + bz)j) from the definition of the sum of a and b
a(d, + b;)i + a(az + b2)j from the definition of scalar multiplication
= ad\i+ abji+ ad2j+ab2j eliminating the brackets
= ad\i+ ad2j + ab\i+aboj_ re-arranging the terms
= a(d\i + dj) + a(bji+ bj) factorizing out a
=aa+ab from the component forms of a and b
It is important that you follow each of these steps and that you appreciate that they
are there to ensure that no unwarranted assumptions are made. The first three steps
follow from the definitions of the addition of a pair of two-dimensional
geometric vectors and the scalar multiplication of a two-dimensional
geometric vector respectively. The next three steps follow from the rules of
algebra and the final step follows from the given forms of the two-
dimensional geometric vectors in their component parts.
Next frame

From the specific to the general 19

In the last frame it was proved that the equation a(a + b) = aa + ab is true for
any scalar and any two-dimensional geometric vectors. This is important. We need
to be able to move away from specific demonstrations with specific scalars and
specific two-dimensional geometric vectors to the realm of general statements
that apply to any scalar and any two-dimensional geometric vectors.
Now let’s look at the second of the properties in Frame 14, namely:
(a+ P)a=aa+ fa
The left-hand side is:

Try and derive the right-hand side

182 Linear Algebra

20
Because:

(a+ S)a
= (a+ B)(a,i + apj) writing a in its component form
= (a+ B)ayit+ (at B)agj multiplying by the scalar a +
= ad\i+ Ba,i+ ad2j+ Ba2j eliminating the brackets
= adi + ad2j + Gaji+ Gazj re-arranging the terms
= a(d1i+ doj) + B(a,i+ a2j) factorizing common scalars
=aa+ Ba recognizing the components of a and b
Notice how the left-hand side is first deconstructed into its component
parts by line 3 and then reconstructed in the format of the right-hand side. If
you are not used to this sort of manipulation it can be a bit daunting at first.
Read through those six lines to make sure you understand why each step was
taken when it was.
Another way to achieve the proof is to work backwards as well as forwards at
the same time. See what you can do with the third equation:
(a)a = a(Ga)
again, the left-hand side is:

Again, try and derive the right-hand side.

. If you get stuck, try working backwards from (aZ)a
and forwards from a(Ga) at the same time

21 a(a)
Because:
(aB)a= (aB)(a1i + aj) separating a into its component parts
= (af)a,i+ (aG)a2j multiplying by the scalar af
= aBa,i+ aaj eliminating the brackets
= a( Gai + Baj) factor out the scalar a
= a(GB[ayi + apj)) factor out the scalar 3
= a((a) recognizing a from its component parts
Move to the next frame
Vector spaces 183

Identity element for scalar multiplication

22
Finally, the fourth equation:
la=a
This equation is telling us that 1 is the identity element for scalar
multiplication and the equation is true because:
la = 1(aji + aj)
=1xqi+1xaj
=d,i-+ Aaj
=a
In summary, two-dimensional geometric vectors are closed under scalar
multiplication and the scalar multiplication of two-dimensional geometric
vectors has the following properties:
a(a+b) = aa+ab
(a+ Bja=aa+ fa
(a)a = a(fa)
la=a
We are now ready to define the mathematical structure that is called a vector
space.
Move to the next frame

Mathematical structures
A mathematical structure consists of: 23

one or more collections of objects

a collection of defined operations acting between objects
a collection of rules that define how objects combine under the operations
The collections of objects that we have just been discussing are the collection
of two-dimensional geometric vectors and the entire collection of real
numbers (scalars). The defined operations are the addition of a pair of two-
dimensional geometric vectors and the scalar multiplication of a scalar with a
single two-dimensional geometric vector. The rules are:
Two-dimensional geometric vectors are closed under addition and scalar
multiplication. Addition and scalar multiplication are governed by the
following rules where 0 is the zero two-dimensional geometric vector:
Addition: Scalar multiplication:
a+b=b+a a(a+b)=aa+ab
a+(b+c)=(a+b)+c (a+ Bja=aa+Ba
a+O=a (aG)a = a(Ga)
a+(—a)=0 la=a be
184 Linear Algebra

This mathematical structure has a name - it is called a Vector Space over the
real numbers. However, as we shall see, the idea of a vector space is far wider
than a mathematical structure just involving two-dimensional geometric
vectors. Three-dimensional geometric vectors, matrices, real-valued contin-
uous functions defined on a closed interval, complex numbers, and more are
all candidates for the objects of a vector space. We shall find that by looking at
vector spaces in general, we shall learn more about specific vector spaces. But
before we move on, we must be more formal with the definition of a vector
space.
Move to the next frame

Vector space
24 A vector space S(V) over the real numbers (scalars) consists of a set V of objects
called vectors (denoted by bold lower case Latin x1, x2, ..., Xn, ... or by
x, y, ...), the set R of real numbers (denoted by lower case Greek a, @2,...,
Qn, --. Or by a, GB, ...)), and two defined operations of addition of vectors and
multiplication of a vector by a scalar which satisfy the following rules (referred
to as axioms): ¢
For every pair of vectors x; and x; in V there is a unique vector in V called
their sum and denoted by x; + xj. Furthermore:
Xj + Xj = Xj + Xj

Xj + (Xj + Xx) = (Ki + Xj) + XK

There is a unique vector denoted by 0 (the zero vector) such that for every
vector x; in V:
xj + 0 = x;
and for every vector x; in V there is another vector in V denoted by —x; such
that:
x; + (—x;) = 0
For every vector x; in V and scalar a in R there is a unique vector called the
scalar product and denoted by ax;. Furthermore:
a(X; + X;) = aX; + ax;
(aj + aj)K = aX + ajX
(ajaj)X = aj(ajx)
Sx
Because the scalars are the set of real numbers R, the vector space over the real
numbers is called a real vector space. It is possible for the scalars to be the set of
complex numbers C in which case the vector space over the complex numbers
is called a complex vector space. In all that follows, we shall only consider real
vector spaces.
Let’s now see how this all works with matrices.
Move to the next frame
eee
a
Vector spaces . Stee 185,

Matrices
Given the pair of 2 x 3 matrices: 25
A= 1 -2 3 ¥ 7 6 -1
i 0 4) and B=( 7, 3 0 )
their sum is defined by adding their respective components as:
A+B=(/ —2 (2 6-1
=—50.4 —2 3 0
= 1+7 —2+6 3+(-1)
—5+(-2) 0+3 4+0 )
-(° te
a eet
=C
and the result of adding two 2 x 3 matrices together is another ............
The answer is in the next frame

Because the addition of two matrices always results in another matrix -

26
provided the addition can be defined; be aware that only matrices with the same
number of rows and the same number of columns can be added together.
So, just like two-dimensional geometric vectors, the set of 2 x 3 matrices is
closed under addition.
Let’s move on again

Scalar multiplication 27
If a matrix is multiplied by a scalar, then each component of the matrix is
multiplied by that scalar. For example:
1 —2 3
If A = & 0 ‘) th en

1 2.3
SA =5
-5 0 4
EN ORO ee LAS
~\5x(-5) 5x0 5x4
SY iets
ets
=>
and the result of multiplying a 2 x 3 matrix by a number is another ............
Next frame
Se a es
Giada eka Linear Algebra

28 |
Because the multiplication of a 2 x 3 matrix by a scalar produces another 2 x 3
matrix we say that:
The set of 2 x 3 matrices is closed under ............

The answer is in the next frame

29 multiplication by a scalar

Because multiplying a 2 x 3 matrix by a scalar always results in another 2 x 3

matrix. 4
Move to the next frame

|30 Notation
To save space and to make our deliberations applicable to general rather than
specific matrices, we shall denote a typical n x m matrix A with the notation:

A = (aj) for 1 <i<nand1<j<m

In this notation aj is the matrix element in the ith row and jth column.
Using this notation the last two results can be written as:
A+ B= (aj) + (bij) forl1<i<2and1<j<3
(ai + bi) adding corresponding elements
= (cy) ° Cj = ay + bi
=¢
and
aA = a(aj) forl<i<2and1<j<3
(aaj) adding corresponding elements
= (dj) Ci = aij
=D
That is, all 2 x 3 are closed under addition and scalar multiplication. Next
we shall look at the axioms for addition.
Move to the next frame
Vector spaces
187°

GF

Axioms for the addition of matrices 31 |

6 tativit

A+ B= (aj) + (bj for1<i<2and1<j<3

(ag + bg) adding corresponding elements
= (by + ag) addition of numbers is commutative
= (by) + (a4)
=B+A
Now you see how it works, try one for yourself.

P ‘ativi

A + [B+ Cj = (ag) + [(by) + (G)] for1<i<2and1<j<3

The answer is in the next frame

32
Because:
A+{(B+C)
= (ay) + [(by)+(cy)| forl1<i<2and1<j<3
= (ay) + (by + cy) adding corresponding elements
= (ay + by + cy) adding corresponding elements
= ((ayj + by) + cy) addition of numbers is associative
= (ay + by) + (cy) from the definition of the addition of two matrices
= {(ay) + (by)| + (cy) from the definition of the addition of two matrices
={A+B)+C
Next frame

The identity element for addition 33 |

Recall that there is a unique vector that we denote by 0 with the property that:
x+0=x
The 2x3 matrix that we denote by 0 with this property is given in
component form as............
Can you think what it is? The answer is in the next frame
188 Linear Algebra

34 : e 0 iD

Ta 2a
Because, for example, if A = ( )then:
Coa)
Aicose ees eae
pees ra 000
tier:
ule Pate 7a
=,\
The matrix 0 is called the null matrix and the general n x m null matrix has
zero for all its elements.
Move to the next frame

Finally, for each matrix A = (aj) there is another matrix which is denoted by
—A = (—aj) and is such that:
A + [—A] = (ay) + (—ai)
= (aj — aj)

So we see that matrices satisfy the first four axioms governing the operation
of addition within a vector space, namely:
A+B=B+A
A+ {B+C]
= [A+B]}+C
A+O0O=A
A+ [-A] =0
We now need to show that they also satisfy the axioms dealing with the
operation of scalar multiplication.

The working is quite straightforward and starts in the next frame

|36 Scalar multiplication

To check your memory, can you remember what the four axioms are that
govern the operation of scalar multiplication?
They are:

The answer is in the next frame

Vector spaces ; 189°

a(Xj + Xj) = aXj + aX; 37

(a4 + aj)X = ajX + ajX
(ajaz)X = aj(ajx)
ibe,

In terms of matrices they read as:

a{A
+ B] = aA+ aB
[a+ BJA =aA+BA
[af]A= o[BA)
labfA =A
Now it’s your turn to demonstrate that they are satisfied. Firstly:
a{A + B] = aA +aB because af/A+B]=............
The answer is in the next frame

ofA + B] = a[(ay) + (by) for1<i<2and1<j<3

= a(ai “i bij
)
(a [aij + bi
|)
= (aay + abj)
(aay) + (bj)
= a(ay) + (bj)

Because:
a{A + B]= a[(ay) + (bj)| ‘from the notation for a matrix
= a(aj + bj) definition of the sum of two matrices
= (a[ay + by}) definition of the scalar multiple of a matrix
= (aaj + abj) from the distributive rule of algebra
= (aaj) + (abj) definition of the sum of two matrices
= a(aj) + a(bj) definition of the scalar product of a matrix
= aA +aB from the notation for a matrix
Did you manage all of the steps?
Next, the second axiom {a + {JA = aA + aB. Here,
la+ BIA =... 0.00000.
The answer is in the next frame
re
ee
os.) Beene Linear Algebra

for 1 Sic 2-and 1 <7 <3

aaj) + B(ai)
= aay) + B(ai)
=aA+ GA

Because:
[a + BJA= [a + ] (aj) matrix notation
= (fa + f]a;) scalar multiplication 2
= (aay + Baj) distributivity
= (aaj) + B(aj) matrix addition
=a(aj) + B(aj) scalar multiplication
=aA+ BA matrix notation

Next, the third axiom [a/]A = a[GA]. Here,

The answer is in the next frame

40 [aB|A= [0,3] (aj) for 1 <i<2and1<j<3

Because:
[a|A= [af] (aj) matrix notation
= ([aG]a,) scalar multiplication
= (aaj) removing the brackets
= (a|[Gaj]) factorizing a
= a(Baj) scalar multiplication
= a[B(a;)| scalar multiplication
= a[GA] matrix notation

Finally, the fourth axiom 1A = A. Here,

The answer is in the next frame

ey
Vector spaces 497

1A = 1(a;;) forl<i<2and1<j<3

=5)(1<ay)

= (ay)=A

And that completes the demonstration that 2 x 3 matrices form a real vector
space. Notice that all that has been said for 2 x 3 matrices can equally be said
for general n x m matrices, provided that for any specific vector space of n x m
matrices the values of n and m are fixed values to ensure that addition can be
defined.
Next frame

Polynomials
Just to make sure that you are clear about what constitutes a real vector space, 42 |
we shall show that the set P,, of all nth degree polynomials form a real vector
space S(P,,). In this vector space the vectors are of the form:

DP, = 40 + ax + a2x? +... + ayx"

and addition and scalar multiplication are defined as one would expect:
Pr +p = (do + a1x + a2x? +... + nx") + (bo + Dix + b2x? +... + bnx")
= [do + bo + [ay + b;|x + [a2 + b2)x? +...+ [an + b,x"

and
= 2 n
ap, = a(do —— G1X s- AaX” wt nx )

= Odo + AG, X + Ad2x2 +... + aanx"

From these two definitions of addition and scalar multiplication, it is clear
that nth degree polynomials are ............ under addition and scalar
multiplication.
The answer is in the next frame

4s |
Because the addition of two nth degree polynomials is another nth degree
polynomial and the product of a scalar and an nth degree polynomial is
another mth degree polynomial.
: 192 aca Linear Algebra

Axioms governing addition

Commutativity
Given the two nth degree polynomials
P, = (do + 41X +... + Gnx") and q,, = (bo + bix + box* +... + byx”) then

The answer is in the next frame

aa
Because:

PrtGy = (do + aiX+...+ 4x") + (bo + bx +... + dp")

= [do + bo] + [a1 + bi]x +... + [an + Dy]x”
= [bo + ao] + [by + ay|x +... [Dn + Qn]x”
(bo + byx +... + Dyx") + (do + 1X +... + anX")
= 4+ Pn

Associativity

Given the three nth degree polynomials

P, = (do + a1X4+...+
yx"), Gq,= (bo +b1xX+...+bnx") and
n= (Cotcxt+...+¢C,x") then
Be ei a cee

The answer is in the next frame

45 (P,+4,) +¥n

Because:

P, + (q,, +¥n)
= (do + 1X +... + dnxX") + ((bo + Bix +... + bp”) + (Co +010 +... + CnX"))
= (do + 1X +... + nx") + ([bo + Co] + [bi +1]x +... + [Dn + Gy]x”)
= ({do + [bo + Col} + {a1 + [bi +c1]}x +... + {Gn + [bn + Cn] }x”)
= ({4o + bo + co} + {41 +b +14 e1}X4+...+4
{Gy + Dp + Cy}x")
= ({[a0 + bo] + co} + {[a1 + bi] + cr}x+... + {[an + Bp] + cn} x")
I (lao + Do] + [a1 + By]x +... + [an + bn]x”) + (co + 1X +... + px")
((do + 41X + ...GnX") + (bo + Bix +... + Dnx")) + (Co +X +... + nx")
(P, +q,) +P n

Next frame
Vector spaces 193

The identity element

46
The nth degree polynomial that corresponds to 0 is:
0=0+0x+...+0x"
because, given p,, = (do +aix+...+ a,x") + (0+ 0x+...+ 0x") then

Next frame

P, 47
Because:

P, + O = (do +aiX4+...4+ a,x") + (0+0x+...+0x")

= ((do + O) + (a; +0)x+...+ (dn + 0)x")
= (do + ayX+...+
yx")
=P
Furthermore, if p,, = (do + a\X +...+ a,x") then:

Next frame

—p,, = —do — 4X —...— Anx"

Because:

P,, a (—p,,) = (do FAX +... + AnX") ir (—do — ax — —AyXx")

= [ao — ao] + [a1 — ay]x +... + [Qn — ay) x"

=0+0x+...+
0x?
—0
Thus we find that the set of nth degree polynomials are closed under addition
and scalar multiplication and that they satisfy the four axioms that govern
addition. All that remains is to demonstrate that they satisfy the four
remaining axioms governing scalar multiplication.
Move to the next frame
 Linear Algebra

49 Scalar multiplication
The four axiom are, for vectors x and y and scalars a and (3:
a(x+y) =ax+ay
(a+ B)x = ax + Bx
(a3)x = a(x)
Ix =x
When the vectors are nth degree polynomials, these four axioms take the
form of:

Next frame

90 a(p, “ipq,,) = ap, + oq,

(a Le B)Pn = ap, + SP,

(a8)p, = o(6p,,)
1p, =P

To show that the first of these axioms is satisfied let

P, = (do + a1X +... + Gnx") and q, = (bo + bix +... + Dnx")

then:
OL Dart Gy.) irae
a
Next frame

51 Because:

a(p, + q,) = a[(do + a1xX +... + nx") + (bo + bx +... + Dnx”)]

a([do + bo] + [a1 + Bix +... + [an + bn]x”)
aldo + bo] + ala, + by ]x +... + alan + by]x”
= [ado + abo} + [aay + ad2|x +... + [aay + aby]x"
= (ado + ad4X +... + A,X") + (abo + abyxX +... + abnx")
= a(do + 4X +... + 4X") + a(bo + bX +... + byx")
= ap, + oq,

By now you should be getting the hang of these demonstrations where

every move has to be explicitly written. If you are still unsure, then go back to
Frame 42 and work through them again, making sure that you know why each
line has been entered.
The second axiom is

Next frame
Vector spaces =. 15

Because:

(a+ B)p,, = (a + B)(ao + a,x +... + a,x")

= (a+ B)ag + (a+ B)ayx+...+ (at B)anx"
= Ado + OG4X+...+ AAnX" + Bag + Gayx+...+ Bayx"
= a(do + @\X+...+4nX") + B(ag + a,x +... + anx")
= ap, + Pp,
The third axiom is

Next frame

a( Sp, 93

Because:

(aB)p,, II= (a@B)(ao + aX +... + nx")

= aBdo + aBayx+...+ aBayx"
= ado + adyX+...+ AdyX" + Bao + Bax + ...3Anx"
= a(do + AyX +... + anX") + B(do + A4X +... + AnXx")
= ap, + OP,

And, finally
tr eee ree
Next frame

54
Because:
1p,, = 1(ao + a1X +... + Gnx")
=1xadapt+1x ayxt+...4+1 x apx"
= dg +ayX+...+
yx"
=P,
enabling us to conclude that the set of nth degree polynomials forms a vector
space.
Let’s move on again
196 Linear Algebra

Subspaces
55 Let S(V) be a real vector space formed from the set of vectors V. A subset U of
V is said to form a subspace of S(V) if it forms a real vector space in its own
right. The subspace is then denoted by S(U).
To amplify, this means that S(U) is totally self-contained in the sense that
all sums and scalar multiplications of vectors in U result in further vectors in
U. So, for example, we have already seen that the set P, of all nth degree
polynomials of the form:
P,, = 40 + aiX +... + 4x"

forms a real vector space S(P,). This is true when n=m orn=m—1 for any
arbitrary value of m. So that P,, and P,-; form real vectorspaces S(P,) and
S(Pm-1) respectively. However, an element of P,,-; can be thought of as an
element of P,,, where the last coefficient is zero, that is:

Gg to oe ek Sa a ee ik ot OX”
Consequently, P;,_1 is a subset of P,,. Therefore S(Pm_1) is a subspace of S(Pm).

Important:
The use of the word vector when discussing vector spaces is historic, coming as
it does from a discussion of geometric vectors. From here on in whenever we
are discussing elements of a vector space, we shall refer to them as vectors, be
they geometric vectors, matrices, polynomials, functions continuous on a
closed interval, or whatever.

Next frame

Ordered n-tuples of real numbers

56 The set of all real numbers is denoted by the symbol R whose elements can be
represented by points on a line. You are also familiar with ordered pairs of real
numbers (x, y) whose elements can be represented by points in the plane — in
two-dimensional space. The set of all ordered pairs of real numbers is denoted
by R’.
Following on from this notation, R® is used to denote ............

Can you guess what is denoted by R??

The answer is in the next frame

7)
Vector spaces 197

all ordered triples of real numbers 57

This is the set of all ordered triples of real numbers (x, y, z) whose elements
can be represented by points in the three-dimensional space coordinated by
x-, y- and z-axes. To enable these ideas to be extended, we shall henceforth
denote an ordered pair by (x, X2) and an ordered triple by (x, x2, x3).
Then, following on from this notation, R” is used to denote all ordered
n-tuples real numbers of the form:
(hig 8a oXs ea)
It is only natural to then say that these can be represented by points in
n-dimensional space despite the fact that we are unable to demonstrate this by
drawing. However, the notion of an n-dimensional space in analogy with two-
and three-dimensional spaces is a useful one.
Now try the following exercises before we progress further

F| Review exercises
1 Show that the set of all ordered triples of real numbers, R”: 58 |
Mee (ey es ka)
where:

X+y = (1, X2, X3) + (V1, Y2, Y3)

= (X%1 +1, X2 + ¥2, X3 + 3)
and
OX = a(X1, X2, X3)
= (QX2, AX2, AX3)

form a real vector space S(R°).

2 Show that the set of ordered pairs of real numbers R?:
X = (X1, X2)
where:
x+y = (%1, X2) + (M1, 2)
= (x1 + Y1, X2 + ya)
and
ax = a(X1, X2)
= (ax2, aX2)
form a subspace S(R’) of S(R°).
The solutions are in the next frame
Ss UE
 Linear Algebra

Solutions

539 1 First, it is clear that the sum of two ordered triples is another ordered triple
and that the scalar multiple of an ordered triple is another ordered triple.
Therefore, ordered triples are closed under addition and multiplication by
scalar. Consider now the axioms:

Addition

X + y = (X1, X2, X3) + (V1, V2, Y3)

= (X%1 +1, X2 +2, X3 + y3)
= (yi +1, Y2 +X2, ¥3 +3)
= (Yi, Yar ¥3) + (1, X2, X3)
=y+x
+ ly + Z] = (41, X2, X3)+ [(Y1, Ya ¥3) + (Z1, Ze, Z3)]
= (X1, X2, X3) + (V1 + Z1, Yo + Z2, V3 + Z3)
= (X1 + [v1 + 21], X2 + [yo + 22], X3 + [v3 + Zs])
= ([41 + yi] + 21, [X2 + y2] + Za, [x3 +3] + 23)
= (41 +1], [2 + yo], [x3 + y3]) + (Z1, 22, 23)
= [(X1, X2, X3) + (Vi, ¥2, Y3)] + (Z1, Z2, Z3)
=([xk+y]+z
The zero element is 0 = 0(x1, x2, x3) = (0, 0, 0). Furthermore,
xX + O = (X1, X2, x3) + (0, 0, 0)
= (xX; + 0, x2
+ 0, x3 +0)
me X2, X3)

—X = —(X1, X2, X3) = (—X1, —X2, —x3) so that

XK + (—X) = (%1, X2, X3) + (—X1, —X2, —X3)
= (%1 — X1, X2 — X2, X3 — X3)
=0

Scalar multiplication

a(x + y) = al(x1, X2, x3) + (V1, Ya, ¥3)]

= a(X1 + 1, X2 +72, X3 + Y3)
I (alx1 +1], [x2 + yz], afx3 + y3])

= (Jax; + ay], [ax2 + aye], [ax3 + ays)])

= (aX1, AX2, ax3) + (a1, ay2, ay3)
= a(X1, X2, X3) + a(V1, Y2, Y3)
= AX 4 ay

3@
Vector spaces 199
(a + B)x = (a + B)(x1, X2, x3)
= ((a + B)x1, (a+ B)x2, (a + B)x3)
= a(X1, X2, X3) + B(X1, X2, X3)
= AX + ay

(a3)x = (a3)(X2, X2, X3)

= ((a@f)x1, (aB)x2, (a3)x3)
(a(GX1), a(x2), a:(3x3))

= a(3(X1,
a( X2, X3))
= a(x)
1X = 1(%1, X2, X3)
= (Le X EL x5, 1 x X3)

= (x1, x2, x3)

=X

Therefore the set of ordered triples R’ forms a real vector space S(R°).
Each ordered pair of real numbers can be thought of as an ordered triple
where, for example, the third component is zero:
X = (X1, X2) = (x1, X2, 0)
Consequently the set of ordered pairs R? is a subset of the set of ordered
triples R*. Furthermore, by repeating the solution to the previous
example with x3 = 0 it is easily shown that R* forms a real vector space
where the zero vector is 0 = (0, 0). Consequently, R? forms a subspace
S(R7) of S(R?).
Notice that the ordered pairs can be represented by points on a plane
but, because a vector space must contain the zero vector, only those
planes that pass Giough the origin (thereby containing the point (0, 0))
form subspaces of S(R”) .

Let’s move on
200 Linear Algebra

Linear combinations and spanning sets

60 Given the set U of vectors x1, X2, ..., Xn then if aj, a2, ... a», are n scalars, the
construct:
n

QyX 1 + a2K2+...+ Qy,Xn = SS OjXy

i=l

is called a linear combination of the n vectors Xj, X2, ..., Xn. Furthermore, if set
V forms a real vector space §(V) and U is a subset of V then the set W of all
linear combinations of the vectors of U also forms a vector space S(W). That is,
if two arbitrary linear combinations w; and wz in W are given as:
n n

w= So aux: and w2 = 5 Bx
1 i=1
then their sum w) + W2 is alsoa............ of vectors in W.

The answer is in the next frame

61
Because:
n n

wit we = > ax t+ > Xi

j=1 j=1

=) lei + Gx:
i=1
n

= 3: yiX; another linear combination of vectors x1, X2, ..., Xn

i=1
where 4 = a; + (i

so W is closed under addition. Also, the scalar product yw, isa............ of

vectors in W.

The answer is in the next frame

Vector spaces 201

linear combination 62

Because:
n
AW, = A yy jXj
i=1
n
= > AQGX;
al
n

= ae 6;x; another linear combination of vectors x), X2, ..., Xn

j=1
where 6; = Aq;

so W is closed under scalar multiplication.

In addition the zero vector is given as............
Next frame

63

Because:

I £ a

As a result, W forms a real vector space S(W).

Notice that each vector in U can also be written as a linear combination of
the vectors in U. For example:
XK, =1xx,+0xx2+0xx3+...+0XXn

and

Next frame
ne ah elle I ai Mt tlie Bi nh ik lh RT
i Linear Algebra

x37=0xx,+0xX2+1xxX3+...4+0XXp

So that set U is a subset of W. Because every vector in W can be written as a

linear combination of those vectors just in U, we say that the vectors in U span
the vector space S(W) and that U is a spanning set of S(W).
Let’s look at a few examples.

Example 1

Show that the vectors x; = (1, 1) and x2 = (0, 1) span S(R’).

The general vector of R” is of the form (a, b). If x; and x2 span S(R*) then it
will be possible to find values for a and ( such that: se
a(1, 1) + G(0, 1) = (a, b) thatis (a, a+ Z) = (a, b)
That is, it will be possible to solve the simultaneous equations:
a=a
a+P=b

This is possible because:

Next frame

65 ‘

Because:

—= (fi oe 6,0)
so that and so
a+ PB=b a+ B=b B=b-a

Example 2

Show that the vectors x; = (1, 1, 1) and x2 = (1, 2, 0) do not span S(R?) but
do span a subspace.
The general vector of R? is of the form (a, b, c). If x; and x2 span S(R°) then it
must be possible to find unique values for a and ( such that:
al) it) BO 270) = (4a; b,'c)* that is (a46 "a4 20; a) = (a rc)
That is it must be possible to solve the simultaneous equations ............
Next frame

3
Vector spaces 203

66

Because equating components of (a + 3, a + 23, a) yields:

[1]: a+fB=a
(2): a+28=b
[3]. a=c

However, [1] x 2 — [2] =............

Next frame

a=2a—b=c 67

Because:

[1] x 2 — [2] = (2a + 28) — (a+ 28)

==
=2a—b
=. itom 15)

The equations are not consistent without imposing the additional condition
that c = 2a — b. Therefore, the two vectors x; = (1, 1, 1) and x2 = (1, 2, 0) do
not span S(R°). However, they do span a subspace of S(R*) formed of vectors
of the form:

Next frame

(a, b, 2a —b) 68

Example 3

Show that the vectors x; = (1, 1) and x2 = (—1, —1) span a subspace of S(R?).
The general vector of R? is of the form (a, b). If x; and x2 span S(R*), then it
will be possible to find unique values for a and 3 such that:
a(1, 1)+ 6(-1, —1) =(a, b) thatis (a— 6, a— 8) =(a4, b)
That is it will be possible to solve the simultaneous equations:
a—-GZ=a
a—-B=b

This is only possible when ............

Next frame
204 Linear Algebra

69 r dexD

This means that x; = (1, 1) and x2 = (1, —1) span a subspace of vectors of
the form (a, a) which consists of all the points that lie on a line in the x-y
plane, through the origin, and inclined at a 45° angle to the axes.
Notice that either one of x; Or X2 will span this subspace on their own. For
example:
a(1,1)=(a,a) and —a(~1, —1)=(4, 4)
which demonstrates that spanning sets are not unique.

Example 4 =

Does the vector (1, 3, 2) belong to the subspace spanned by:

(a) (0, 1, 1) and (1, 1, 0)

(b) (1, 0, 1) and (0, 1, 0)?
Next frame

70 (a) Yes
(b) No

Because:

(a) The subspace spanned by (0, 1, 1) and (1, 1, 0) contains vectors of the
form a(0, 1, 1) + @(1, 1, 0) = (G, a+, @). That is where the second
component is the sum of the first and third components. The vector
(1, 3, 2) is of this form and so it belongs to the subspace spanned by
(Omt, 1) and (1, 1,0).
(b) The subspace spanned by (1, 0, 1) and (1, 1, 0) contains vectors of the
form a(1, 0, 1) + G(0, 1, 0) = (a, 6, a). That is where the first and third
components are the same. The vector (1, 3, 2) is not of this form and so
it does not belong to the subspace spanned by (1, 0, 1) and (0, 1, 0).
Let’s move on

Linear dependence and independence

71 _ Ifa vector x can be written as a linear combination of the vectors x1, X2, ..., Xn
so that:
XK = a1 Xj4 + O19.%9 +... + On Xn

then x is said to be linearly dependent on x, X2, ..., Xn.

=
Vector spaces
205

Example 1

In the vector space of second degree polynomials, to show that the quadratic
P2(x) = 4 — 3x + 2x is linearly dependent on the three quadratics:

qo (x) =1—x+x?
r2(x) = 1- x?
So(x) = x — 4x?
it is only necessary to write p,(x) as the linear combination:

P2(X) = aq,(x) + Br2(x) + y2(x)

and then show that a unique solution for a, 3 and 4 is possible.
Writing p,(x) as a linear combination of q>(X), F2(x) and s2(x) yields:

p2(x) = 4 — 3x +4 2x?
= aG,(X) + Br2(x) + yS2(x)

Next frame

P2(x) = 4 — 3x + 2x? 72
= (a+ B) + (-a+7)x + (a — 8 — 4y)x?

Because:
p2(x) = 4 — 3x + 2x
= a(1—x+x*) + B(1 — x”) +(x — 4x”)
= (a+) +(-a+7)x
+ (a — B— 4y)x?
This gives rise to the three simultaneous equations ............
Next frame

73

Because two polynomials in x are equal if the coefficients of their respective

powers of x are equal and equating coefficients of powers of x yields:
a+ 6=4
=ery=-3
a-—-6G-4y=2

These three equations can be written in matrix form as............

Next frame
 Linear Algebra

Because
a+G=4 1lxa+1x6@+0x7=4
—a+y=-3 canbe writtenas -—1xa+0x@4+1xy=-3
a—-p—4y=2 . lxa-1x6B-4x7v=2

the determinant of the coefficient matrix is............

= Next frame

Because:

1 1 0)
= alt Cal = ries in 1,|=1-3=-240
LY =i —4 :

Since the determinant of the coefficient matrix is non-zero, the three

equations are consistent and, therefore, have a unique solution. This means
P»(x) can be written as a linear combination of the vectors q,(x), r2(x) and
S2(x) and so is linearly dependent on them.
4

Example 2

In the vector space of second degree polynomials, show that the quadratic
P2(x) = 4 — 3x + 2x? is not linearly dependent on the three quadratics:

qo(x) =1—x+x
r2(x) = —2 + 2x — 2x”
S2(x) =x — 4x?
it is only necessary to write p,(x) as the linear combination:

P2(X) = aGy (x) + Gr2(x) + yS2(x)

and then show that a unique solution for a, 6 and ¥ is not possible.
Writing p,(x) as a linear combination: ;
Da 4
= Ay (x) + Br2(x) + yS2(x)
I
ee + &
i)
+

Next frame
Vector spaces 207

P2(x) = 4 — 3x + 2x? 76
= (a — 28) + (-a+ 28y)x+ (a -— 26 - 4y)x?

Because:
p2(x) = 4 -—3x+ 2x?
= a(1—x +x?) + B(—2 + 2x — 2x?) + 7(x — 4x?)
= (a — 28) + (-a+28+-)x + (a — 28 — 4y)x?

This gives rise to the three simultaneous equations ............

Next frame

a-—-2pB=4 77
—a2By = —3
a—260-4y=2

Because two polynomials in x are equal if the coefficients of their respective

powers of x are equal and equating coefficients of powers of x yields:

pdr
ee at
a—-28-—4,=2

These three equations can be written in matrix form as ............

Next frame

78

Because
a—-2@G=4 lxa-2x6+0xy7y=4
—a+26+7=-3 can be writtenas -lxa+2x@+1xy=-3
a—26-—4y=2 lxa-2xB-4xy=2

the determinant of the coefficient matrix is ............

Next frame
208 Linear Algebra

79 |0|
Because:
Thule? a0
ei: 4 ade 1 \-cal7 1j)=-6+6=0
Oy ee
Since the determinant of the coefficient matrix is zero, the three equations are
inconsistent and, therefore, do not have a unique solution. This means p,(x)
cannot be written as a linear combination of the vectors q,(x), r2(x), and s2(x)
and so is not linearly dependent on them.
Let’s move on

Linear independence
80 Given the linear combination:
aX + 9X9 +... + On Xn =0 ; 7

that is:

4X1 + 2X2 + aj_yXj_-1 + A{Xj + Qj41XKi41 +... + anX, =O

then:

ajiXj = —(a1X1 atheoe tale Qj—-1Xj-1 + O441Xi41-.. + AnXn)

so that:
1
xj = — ms (a,X1 + a2X2+...+ Qj-1Xj_-1 + Qj41Xi41--. QnXn)

PPOWICE Bec aesnenn

Next frame

1 Because division by 0 is not defined. This means that, provided a; #0, the
dependence of x; can be explicitly given as:
1
= 7. (a1X] +a2K2+...+ aj;-1Xj_-1 + O441XKi41 --- AnXn)
i
Vector spaces 209

If a; = 0, then this is not possible and this statement is true for any value of
the subscript i from 1 to n. This gives rise to the following test for linear
independence.
If the linear combination:
Q1X1 + a2%2+...+anX;, =0

implies that a; = az =... =a, = 0, then the vectors x;, X2, ..., X, are linearly
independent of each other — no one of the x; can be written as a linear
combination of the rest.

Example 1

To show that in the real vector space of ordered triples the three triples
(1, 0, 1), (1, 1, 0) and (0, 1, 0) are linearly independent of each other, we first
write them as a linear combination equated to zero as:

Next frame

a(1, 0, 1) + B(1, 1, 0) + (0, 1, 0) = (0, 0, 0) 82

This gives rise to the three simultaneous equations ............

Next frame

a+Z=0 83 |

Because:

a(1, 0, 1) + B(1, 1, 0) + 7(0, 1, 0) = (a, 0, a) + (8, B, 0) + (0, 7, 0)

=(a+6, B+a, a)
= (0,
0, 0)

and so
a+B=0
B+a=0
a=0

PRGTOT ORG Ge cds Saws o'sAi < ee eee Te et eee ere re

Next frame
LL —
210 Linear Algebra

a=0, 8@=0andy7y=0

Because a = 0 from the third equation, so 6 = 0 from the first equation, and
then 7 = 0 from the second equation.
Therefore, the three ordered triples are linearly independent of each other
and no one of the ordered triples can be written as a linear combination of the
other two.
Let’s move on

Basis vectors and dimension

85 As we saw in Frame 70, spanning sets are not unique. ee ue to show
that each of the following collections of vectors spans S(R”):
(ajetieO; LATS, 0) and: (0-1-0)
(b) (1, 0, 0), (1, 1, 0) and (1, 1, 1)
(cad, 0, 1)7(0; 1, 0) and.(1, 1, 8
we write:

(a) a(1, 0, 1) + @(1, 1, 0) + 7(0, 1, 0) = (a, b, c), which gives rise to three
simultaneous equations which give further rise to a coefficient
determinant:+

Next frame

Because:

a(1, 0, 1)+ 8(1, 1, 0) + 7(0, 1, 0) = (a, b, c)

yields:
(a 478, 0+ 4, a) = (a) B, 6)
that is:

i ihe Qa a LO 10
Le oo | a|=[b] where Jo 1 1/=[} 9)=1
Om y (é Om O

expanding about the third row.

mr)
Vector spaces 211

This means that the three simultaneous equations are consistent and a
unique solution to them can be found. Consequently, the general vector
of R° can be written as a linear combination of the three vectors (1, 0, 1),
(1, 1, 0) and (0, 1, 0) which, therefore, do span S(R3).

(b) a(1, 0, 0) + G(1, 1, 0) + 7(1, 1, 1) = (a, b, c) gives rise to three simulta-

neous equations which give further rise to a coefficient determinant

Next frame

1 87 |

Because:

a(1, 0, 0) + 6(1, 1, 0) + (1, 1, 1) = (a, b, c)

yields:
(a+f+y7,
8+ 7) = (4, b,c)
that is:

{vena lhoest| a rie p 4;

epee a B where JO 1 1)=(4 aie
Ome Ome y ey ga Mt

expanding about the third row.

This means that the three simultaneous equations are consistent and a
solution to them can be found. Consequently, the general vector of R? can
be written as a linear combination of the three vectors (1, 0, 0), (1, 1, 0)
and (1, 1, 1) which, therefore, do span S(R°).
Hence spanning sets are not unique.

(c) a(1, 0, 1) + B(0, 1, 0) + 7(0, 0, 1) + (1, 1, 1) = (@, B, c) gives rise to three

simultaneous equations

Next frame
. _- wecior soa has the Ciresiot: 1. oie eee Oe SN een
212 Linear Algebra

88 y a+déd=a
B+6=b
a+ y+é=Cc

Because:
a(1, 0, 1) + (0, 1, 0) + 7(0, 0, 1) + (1, 1, 1) = (a, B, c)
yields:
(a+ 6, B+6,a+7+56)
=(a4, D, c)
that is: 7
a+é=a
B+é=b
a+ y+é=c

This gives three equations in four unknowns, so the information given is

insufficient to solve the equations. However, by giving any one of the
unknowns a specific value, then values can be found for the other three.
For example, giving 6 a specific value the equations become:
a =a’ where
a’ = a—6
B=bD'
where b' = b—6
a+7y=c
where c =c—6
These equations now give rise to three simultaneous equations which give
further rise to a coefficient determinant ............
Next frame

89

Because:

a=a’
A=y
at y=c

becomes

Oss Q a’ it Oe @ wt
O>- Lael B}={b | where |0O 1 O =| he
it Key a ay G iO al:

expanding about the first row.

?
Vector spaces 213

This means that once an arbitrary value is given to 6 the three

simultaneous equations are consistent and a solution to them can be
found. Consequently, the general vector of R* can be written as a linear
combination of the four vectors (1, 0, 1), (0, 1, 0), (0, 0, 1) and (1, 1, 1)
which, therefore, do span S(R3).
Next frame

We have just seen that S(R*) can be spanned by a set of three vectors as well as 90
by a set of four vectors. However, the spanning set of four vectors is not a
linearly independent set. For example, if:
a(1, 0, 1) + A(0, 1, 0) + 6(0, 0, 1) + (1, 1, 1) = (0, O, 0)
then a solution is found from:

a+é6=0
B+é6=0
a+y+é6=0

with a possible solution 1=a=6=-6 and y=0O. That is, not all are
necessarily zero because they are linearly dependent. The question then arises
as to the minimum number of vectors required for a spanning set.
Next frame

Minimal spanning sets 91

A minimal spanning set of a vector space S(V) is a spanning set B of linearly
independent vectors and is called a basis for the vector space S(V). The number
of vectors in the basis is called the dimension of the vector space. Just as
spanning sets are not unique, neither are bases, but the dimension is; every
basis will have the same number of elements.
Let the set of vectors e), €2, €3, ..., €y be a basis for vector space S(V) so that
every vector of V can be uniquely written as a linear combination of the basis
vectors €1, @2, €3, ..., €n. That is, if x is any vector in V then:
X = Qe; + ad2€@2 + 4303 +... + An€n

The numbers 4}, d2, 43, ..., dn are called the coordinates of x with respect to the
basis e;, €2, €3, ..., €, and the basis vectors constitute a coordinate system. The
vector space has the dimension n, where n is the number of basis vectors.
Next frame

Example 92
Show that each of the following sets of vectors forms a basis for R? and find
their respective coordinates:

(a) (1, 0, 0), (1, 1, 0), @, 1, 1)

(b) (1, 0, 1), (0, 1, 0), (0, 1, 1)

(c) (1, 0, 0), (0, 1, 0), (0, 9, 1)

214 Linear Algebra

Solution

(a) Ur 11,0,0), (11, 0), Cll Dspars S(R?) then unique values can be found
for a, 3, and y so that the general ordered triple (a, b, c) can be written as:
a(1, 0, 0) + A(1, 1, 0) = 7(1, 1, 1) = (a, B, ©)
giving rise to three equations with coefficient determinant

Next frame

93

Because:

a(1, 0, 0+ B(1, 1, 0) + 4(1, 1, 1) =(e+6+7, B+% 7)

lyin Re

Therefore:

a+B+y=a t { “1 a
B+y=b thatis |O 1 1 = | 5p
y=C RU ee | Cc

where the coefficient determinant has the value:

4 a Ag
Uae 3 Af=a abe
Oo | aa
Since the coefficient determinant has a non-zero value, the three equations
do have a unique solution and are, therefore, a spanning set for S(R®).
Furthermore, if (a, b, c)=(0, 0, 0), the three equations become:

a+@6+7=0
B+7=0 with solution a=B=7=0
m0
Therefore the three vectors (1, 0, 0), (1, 1, 0), (1, 1, 1) are

Next frame

94 linearly independent

Therefore, being a linearly independent, spans set, they constitute a

Seen for S(R°). .
Next frame
Vector spaces 215

95
And the coordinates of (1, 2, 3) against this basis are ............

Next frame

Because:
a+@6+y7=1 seers
B+y=2 hassolution B=2-—y=-1
y=3 a=1-6-y=-1
And now for the next one

(b) If (1, 0, 1), (0, 1, 0), (0, 1, 1) spans S(R?), then values can be found for a, 97
B, and ¥ so that the general ordered triple (a, b, c) can be written as:
a(1, 0, 1)+ (0, 1, 1)+7(0, 1, 1) = (4, b,c)
giving rise to three equations with coefficient determinant ............
Next frame

98

Because:

a(1, 0, 1) + B(0, 1, 0) + 7(0, 1, 1) = (a, B+a, a+7)

= (a, b, Cc)

Therefore:

= a QpcOul a a
B+y=b thatis 101 1)[6]=|[0

where the coefficient determinant has the value:

111
0-11 =15 =!
001
16. > Linear Algebra

Since the coefficient determinant has a non-zero value the three equations
do have a solution and, therefore, are a spanning set for S(R*). Furthermore, if
(a, b, c) = (0, 0, 0) the three equations become:
a=08+a=0a+7y=0 with solution a=G=y=0

Therefore the three vectors (1, 0, 1), (0, 1, 0), (O, 1, 1) are............
Next frame

99 linearly independent

Therefore, being a linearly independent, spanning set they constitute a

pi Tee for S(R°). 2
Next frame

100
And the coordinates of (1, 2, 3) against this basis are ............
Next frame

Because:
a= 1 &

6+a=2Z2 has solution ae i)4=—0;7

= 2.
a+ty=3

And now for the last one

102 (c) If (1, 0, 0), (0, 1, 0), (0, 0-1) spans S(R3), then values can be found for a,
G, and y so that the general ordered triple (a, b, c) can be written as:
a(1, 0, 0) + B(0, 1, 0) + (0, 0, 1) = (a, b, c)
giving rise to three equations with solution ............
Next frame

2
Vector spaces
217

103

Because:

a(1, 0, 0) + B(0, 1, 0) + (0, 0, 1) = (a, f, ¥)

= (a, b, c)
Therefore:
a=ad
B=b
y= c
Since the three equations do have a solution, the three vectors form a
spanning set for S(R’). Furthermore, if (a, b, c) = (0, 0, 0) the three equations
become:
Ge)
pe
0)

Therefore the three vectors (1, 0, 0), (0, 1, 0), (0, 0, 1) are............

Next frame

linearly independent 104

Therefore, being a linearly independent, spanning set they constitute a

Mi han 3 for S(R?).
Next frame

105
And the coordinates of (1, 2, 3) against this basis are ............
Next frame
218 Linear Algebra

106 r a= 1 Ge 277S3

Because the coefficients of (1, 2, 3) are the same as the coordinates against this
basis, that is:
1(1, 0, 0)+2(0, 1, 0) +3(0/07 1) = (1, Z, 3)
this basis is called the standard basis for R*. Note that if ordered triples are
identified with geometric vectors, then the standard basis for R? is identified
with the Cartesian triple i, j, k that is:
is (is O00)
j = (0, 1, 0) *
k=0,092)
Next frame

Solution spaces
|
107 Consider the simultaneous equations:
—x;, + 4x3 = 0
3X1 + X2 — 5x3 =0

By letting x3 = a these equations have a solution that can be represented by

the components of the ordered triple
(X1, X2, x3) = (a ietekaceheiersiar tale pao RIES
3:torr Fy dO OOO ONO D)

108 (4a, —7a, a)

Because:
The first equation yields:
X, = 4x3

and substitution into the second equation yields:

X2 = —7X3

Therefore, the components of any ordered triple of the form:

(4a, —7a, a) =a(4, —7, 1)
will provide a solution to the two equations.
We say that the solution space of the equations is a real vector space spanned
by the ordered triple (4, —7, 1). In graphical terms this can be represented by
the points on a straight line passing through the origin and the point
(4, —7, 1).
7)
Vector spaces
219

Try another. The solution space of the equations:

3X1 + 2x2 —x3 =0
4x; —X2 + 2x3 =0
— 7x3 =0
X; + 8x2

consists of the components of ordered triples of the form

Next frame

(—0.3a, a, 1.1a) 109

Because:
3X1 + 2x2 — x3 =0 [1]
4X1 — X2 + 2x3 =0 [2]
X1 + 8x2 — 7x3 =0 [3]

[1] x 2+ [2]:
6X1 + 4x2 — 2x3 = 0 bie oe:
4x; — X2 + 2x3 =0 [2]

yields
10x; + 3x2 =0 so x; = —0.3x2

Substitution into [1] yields

—0.9x2 + 2x2 — x3 =O thatis x3 = 1.1x2

Checking with equation [3]:

X1 + 8x2 — 7x3 gives —0.3x2 + 8x2 —7.7x2 =0

Therefore, the solution is given by the components of the ordered triple

(—0.3a, a, 1.1a).
That is, the solution space of the equations
3X1 + 2x2
— x3 =0
4x; —X2+2x3 =0
X1 + 8x2 — 7x3 =0

is the subspace of S(R*) spanned by (............ a Hee te oeAe Se ee )

Next frame
c73 ie Linear Algebra

110 (—0.3, 1, 1.1)

Because:
Every solution is given by the components of
(—0.3a, a, 1.1a) = a(—0.3, 1, 1.1)
each one of which is contained in the subspace of S(R*) spanned by
(.0:3,015.1.1),
Notice that this subspace has (—0.3, 1, 1.1) as its basis and is, therefore, of
dimension:

Next frame

111
Because:
The basis only has 1 element, namely (—0.3, 1, 1.1).
Let’s try another just to make sure. The solution space of the matrix equation:
Le22 =3. X1 0)
tees 4 X2 = 0
in iao | X3 0

is a subspace of S(R*) spanned by (............ A asseeah ee)

The answer is in the next frame
Vector spaces
221

(Az, sana) 112

Because:

Loe 24 ess XxX] 0

lia X2 | = | 0] can be written as the three equations:
a5 A X3 0
X1 + 2x2 — 3x3 =0 [1]
X1 + 3x2 + 4x3 =0 [2]
2X, + 5X2 + x3 =0 [3]

[2] — [1]: x2+7x3=0 so x2 = 7x3

Substituting into [1] yields x; — 14x3 — 3x3 =0 so x; = 17x3

Checking in [3]: 2x; + 5x2 + x3 yields 34x3 — 35x3 + x3 =0. Therefore, the
solution is given by the components of the ordered triple (17a, —7a, a).
Consequently, the solution space is a subspace of S(R*) spanned by
(17, —7, 1)

Notice that the three ordered triples (1, 2, —3), (1, 3, 4) and (2, 5, 1) are
linearly dependent.

Be aware: The solution space to Ax = 0 is a vector space (called the null space
of A). In general, however, the solution space to the equation Ax = b where
b #0 is not a vector space because it typically does not include the zero
vector.

Summary
1 Vector space 113 |
A vector space S(V) consists of a set V of objects called vectors, a set of
numbers, and two defined operations of addition of vectors and multi-
plication of a vector by a scalar which satisfy the following axioms:
V is closed under addition of vectors and:
Xj + Xj = Xj + Xj
Xj + (x; + Xx) = (Ki+ xj) + Xx
xj +0=x; where 0 in sin V
x; + (—x;) =0 where — x; is in V whenever x; is in V
V is closed under multiplication of a vector by a scalar and:
a(Xj + Xj) = ax; + ax;
(aj + aj)X = ajX + aX
(ajaj)X = aj(ajX)
ix.=x >
 Linear Algebra

If the scalars are real numbers, the vector space is called a real vector space
and if the scalars are complex numbers, the vector space is called a
complex vector space.
Subspaces
Let S(V) be a real vector space formed from the set of vectors V. A subset U
of V is said to form a subspace of S(V) if it forms a real vector space in its
own right.
Ordered n-tuples
An ordered n-tuple consists of a collection of nm numbers in strict order
(X1, X2, X3, ++) Xn)
The set of all ordered n-tuples is denoted by R”.
Linear combinations
Given the set U of vectors xX), X2, ..., X, then if a1, a2, ..., @, are n scalars
the construct:
n
ayXy + A2%2 +... + AnXn y OGXj
—

is called a linear combination of the n vectors X, X2, ..., Xn.

Spanning sets ;
If set V forms a vector space and U is a subset of vectors x1, X2, ..., Xn Of
V, then the set W of all linear combinations of the vectors of U forms a
vector space S(W). Furthermore, U is a subset of W and because every
vector in W can be written as a linear combination of those vectors just in
U, we say that the vectors in U span the vector space S(W) and that U is a
spanning set of S(W).
Linear dependence and independence
If the vector x can be written as a linear combination of the vectors
Ka, %2) 6.7 Kus that:
X = 1X1 + Q2X%2+...+
AnXn

then x is said to be linearly dependent on xj, X2, ..., Xn. If it cannot be so

written, then it is said to be linearly independent of x1, x2, ..., Xp.
Test for linear independence
If the linear combination:
a1X1 + agxX2 =E metate + AnXn = O

implies that a; = a2 =... =a, =0, then the vectors x1, x2, ..., X, are
linearly independent of each other.
Basis and dimension
A minimal spanning set of a vector space S(V) is a spanning set B of
linearly independent vectors and is called a basis for the vector space S(V).
The number of vectors in the basis is called the dimension of the vector
space.
Vector spaces

9 Coordinates and coordinate systems

If the set of vectors e€;, €2, ..., @, is a basis for vector space S(V), then if x
is any vector in V:
x = ay e; -+- a2e2 ss Sa es Ane€n

The numbers aj, a2, ..., dn are called the coordinates of x with respect to
the basis ej, @2, ...,@, and the basis vectors constitute a coordinate
system.

4 Can You?
Checklist 6

Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames

e Define a vector space and develop proofs in general terms?

Yes L] AY L] C] CJ No
Dw
e Demonstrate how different mathematical objects can form
vector spaces? GD°&
Yes ee ola) L] a gies No
e Define a subspace of a vector space?
Yes few| L] LJ LJ L] No
e Manipulate ordered n-tuples within the context of a vector
space?
Yes i L] L] L 4 No
e Define a linear combination of vectors?
Yes CL L] L] L) L) No
e Define a spanning set of a vector space?
Yes la? Ghee ie ble et “NG
e Distinguish between linear dependence and linear
independence?
Yes ie J ci L] L] No
e Apply the test for linear independence?
Yes C L] L L] CJ No
e Define the criteria for a basis?
Yes phos 3Eals) ot jet Wal ashes. No
e Deduce the dimension of a vector space?
= an Ogg Fl i ST a
224 Linear Algebra

& Test exercise

|
1 15 1 Find the value of a,x; + a2xX2 + a3X3 in R? when: )

X= (2, —1, 0) a,y=-—3

x2 = (0, —3, 4) and a2=2

x3
= (1, 2, 5) a3 =
2 Prove that (a3)x = a(Gx) where x = p, = a+ bx + cx” and a and £ are real
numbers.
3 Show that the set of all complex numbers R with typical element 2, = Xx + Iyx
where i* = —1 when coupled with the real numbers R form a_real vector space
S(£).
4 Let V be the set of real valued functions of a real variable denoted by f, g, h, ...
Define addition and scalar multiplication as:
(f + g)x = f(x) + 8(x) and (af)x = af(x)
Show that V when coupled with the real numbers R forms a vector space.
5 Show that the set of ordered pairs of real numbers (x, y) where x + y = 0 forms a
subspace of S(R*).
6 Find the subspace of the vector space formed by all second degree polynomials
spanned by 1, x+ 1, (x— Ne
7 Determine whether or not the collection of polynomials:
&R

1, Fnac ESR
is linearly independent. If the set is linearly dependent, write the third element
listed as a linear combination of the other two.

se lk a) Oa CO ah *aie TE IOs SCONE Ohl HI tO) x0)

8 Show that the matrices|,) 1 9 , esos tis
, HeOnc , era sake

form a basis for the vector space of all 2 x 3 matrices and define the dimension
of the vector space.

Further problems
116 1 Find the value of ip, + a2p, + a3p; in S(P3) when:
p, = 1x42? a, =4
eas), ety es and a2=-—5

Pp; =4+ 6x a3 =2
2 Solve for x:
9(3¢5,.—7).— 3x = (2,..—1,.0)
2
Vector spaces 225

3 In R? let x = (X1, ¥1) and y = (yj, y2). Define addition and scalar multiplication
as:
x+y =(0, X2+y2) and ax = (ax, ax2)
Does this define a vector space?
Let addition and scalar multiplication of positive real numbers be re-defined as:
xX+y=xxy and ax=x*
where a is a real number and x and y are positive real numbers. Show that with
these re-defined operations the positive real numbers R+ form a vector space.
Show that the set V of all numbers of the form a+bV3 where a and b are
rational numbers when coupled with the real numbers R from a real vector
space S(V).
Given a real vector space S(V) show that the zero element 0 is unique.
How many linearly independent vectors are there among the collection:
(1, 0), (0, —3),.(4, 5), (0, 6), (—2, 0), (5, —1)
Test for linear independence:
1S0oy,
2 eax 224 42) 8 Foe 1077) 2s

; 1 at Rede nt ea la ' ;
Show that the matrices |. 0 0 |1 0 |
1 1/47 linearly independent.

ye A Ca: fa PP os
10 Show that the matrices |) 9|, | 9) | 1 0 |1 1|5pan the vector space of
all 2 x 2 matrices.
11 Which of the following vectors belong to the subspace of S(R*) spanned by
(1, 0, 2) and (1, —1, 1)?

(a) (0, 0, 0)
(b) (3, —2, 4)
(c) (2, 0, 1)
(d) (1, S, 3)
12 Which of the following vectors belong to the subspace of S(P2) spanned by
1—x+x? and 1+x-—x??
(a) 14+x+2x?
(b)) l=2x—2x*
(c) —3 + 4x — 4x?
(d) 5+5x— 2x?
13 If S(V;) and S(V2) are two vector spaces, show that the set that contains those
elements that are common to both V; and V2 forms a vector space.
 Linear Algebra

14 Find the subspaces of S(R*) spanned by:

(a)-(1; 2, —2,-=1),%3, 00,3) and(—1"47070)
(b) (4, 3, 2, 1), (—2, 1, 0, 5) and (2, 4, 2, 6)
15 Show that (1, 0, 1, 1) and (1, 1, 0, 1) spans the same subspace of S(R*) as
(—1, —4, 3, —1 and (1, 2, 1, 2).
16 Let S(V) be a subspace of S(R3) spanned by (1, 2, —3), (0, 4, 0), (1, 0, 1),
(—5, 9, 5) and (3, —3, —3). Find a basis for S(V) and determine the dimension
of S(V).
17 Let S(V) be a subspace the vector space formed from all 2 x 2 matrices that is
1Hls\0 feb l Or Lind) Drak : :
spanned by (j are ara Gecli J and ({ o»):Find a bastsfor

S(V) and determine the dimension of S(V).

18 Let S(V) be a subspace of S(P3) the vector space formed from all third degree
polynomials that is spanned by (1 — x)(1 +x), x(1 — x)(1+x), 1+ 2x — x? - 2x3
and —5x + 5x°. Find a basis for S(V) and determine the dimension of S(V).
19 Find the coordinates of (2, 6, 9) with respect to the basis (1, 0, 0), (1, 1, 0) and
(1, 1, 1) of a subspace of S(R?).
20 Find the solution space for each of the following:
(a) xX, —3x2+x3=0
5X1 + X2 — 3x3 =0
3X1 + 7X2 — 5X3,= 0)

=3*. O81 X1 0
(an eee Pd Pe poet ae
WRF ae: X3 0
(c) —2x, +X2-—x3—-—xX4
=0
2x1 + 4x2 — 3x4 =0
4x, + 3x2 + x3 —2x4=0
5X2 — X3 — 4x4 =0

oe © ee aman ahye 0
2. (88? to-Fosakthlel sx bt slid
Oy eta 8 a ecg | xa |eal
ster arayNiG. 0

@
Linear
transformations

Learning outcomes
When you have completed this Program you will be able to:
Define and identify a bijective function
Recognize an operator and define a linear transformation
Identify some specific linear transformations
Define and identify a bijective linear transformation
Derive the inverse of a bijective linear transformation
Appreciate the effect of linear transformations on a basis
Obtain the matrix associated with a linear transformation
Appreciate the relationship between an invertible linear
transformation and the invertibility of its associated matrix
228 Linear Algebra

Functions
|1 You are familiar with the notation for a real function such as:
3
f(x) =x
with the diagrammatic form:

Xx

This indicates that the function f can be considered as a precess that accepts
the real number input x and processes it to produce the real number output
f(x).
The collection of permitted input values of x is called the domain of the
function and the collection of output values of f(x) is called the range (or
co-domain) of the function. In this example, where f(x) = x, the domain is the
entire collection of real numbers R as is the range.
For the real function whose output is given as:

f(x) = vx
both the domain and the range are............
The answer is in the next frame

2 the non-negative real numbers

Because:

As we are restricted to real numbers, the square root is only defined for
positive real numbers (and zero) so the domain is restricted to the non-
negative real numbers. Furthermore, the surd sign in \/x indicates the
positive square root so the range is also restricted to the non-negative real
numbers.
Next frame

|3 One-to-one correspondence
For the real function with output f(x) = x3, there is said to be a one-to-one
correspondence between domain and range elements because to each domain
element there corresponds a unique range element and vice versa — to each
range element there corresponds a unique domain element.
The domain and the range of the real function with output f(x) = x? are not
in one-to-one correspondence because, for example, f(2) = 4 and f(—2) = 4so
while to each domain element there is a unique range element, it is not the
case the other way round, as is shown in the example given where the single
range element 4 is associated with two domain elements 2 and —2.
4
Linear transformations
229

For which of the following outputs from real functions is there a one-to-one
correspondence between the domain and range elements?

Next frame

Because:

(a) The domain is the entire collection of real numbers R but the range is
restricted to —1 < f(x) < 1. To each domain element there is a unique
range element, but not vice versa. For example, f(0) = f(x) = 0. In fact,
to each range element, there corresponds an infinity of domain
elements.
(b) The domain is the entire collection of real numbers R but the range is
restricted to R* — the positive real numbers. To each domain element
there is a unique range element and vice versa. To each range element,
there corresponds a unique domain element x = In(e*) = Inf (x).
(c) The domain is restricted to —5 < x < 4 which restricts the range to
—125 < f(x) < 64. To each domain element, there is a unique range
element and vice versa. To each range element, there corresponds a
unique domain element.
(d) The domain is the entire collection of real numbers R as is the range. To
each domain element there is a unique range element, but not vice
versa. For example, f(0) = f(1) = f(—1) = 0 because
f(x) =x8 —x =x(x+1)(x- 1)
If a function possesses a domain and a range that are in one-to-one
correspondence with each other, then the function is said to be a bijection.
Therefore, the functions given by (b) and (c) above are bijections whereas the
functions given by (a) and (d) are not bijections.
Next frame
230 Linear Algebra

Operators and transformations

|5 An operator acts in much the same way as a function except that the name is
reserved for processes acting on functions. So, for instance, the notation:

indicates the differential operator D which operates on function f(x) to produce

the derivative f’(x) so that:

The answer is in the next frame

Because:

If f(x) = x° then D(f(x)) = D(x°) = (x°)' = 5x’.

In this example the differential operator has transformed one function into
another. Here the domain of the. operator is the collection of differentiable
functions as is the range, but there isn’t a one-to-one correspondence between
them. For instance:
D(x° + C) = 5x* where C is any constant.
Therefore there is no one-to-one correspondence between the range and the
domain so this operator is not a bijection.
a /2
Similarly, if /(f(x)) = | f (x)dx defines an integral operator J, then:
x=0

Next frame
Linear transformations 231

7
Because:
n/2
If f(x) = cosx, then J(f(x)) = J(cosx) = | cos
xdx = [sin
x]” :aa &
1) _

Here, this integral operator has transformed an integrable function into a

number. However, there is no one-to-one correspondence between the range
and the domain of J because, for example, if f(x) = 2/7, then:

y2jn)=[ ithe= ax=|

anit
—
7 x=0
= 1= J(cosx)

So two different inputs produce the same output; therefore, J is not a

bijection.
In what follows, we shall consider a special type of operator called a linear
operator or, more commonly, a linear transformation.

Next frame

Linear transformations
A linear transformation L is an operator that acts upon the elements of one 8 |
vector space S(V;) to produce elements of another vector space S(V2) while
preserving addition and scalar multiplication. That is:
If S(V;) is a vector space where x; and x2 are elements of V; and if L is a linear
transformation, then for real numbers a; and a2:
L(ayX1 + a2X2) = aL (x1) + aL (Xz)

where L(x,) and L(xz2) are elements of V2, which forms another vector space
S(V2). For example, the integral operator J where:
m |2
(f(x) = | f(x)dx
x=0

is a linear transformation from the vector space of integrable functions over

the interval [0, 7/2] to the vector space of the real numbers because:
m/2
J(aufi(x) + a2f2(x)) = [lie + arfz(x)] dx
am/2 n/2
=| orfi(x)ax+ | _cafi()
x=0 =

x/2 x/2
= ay |BE doe+ a2 | flayax

= a4] (fi(x)) + a2) (f2(x))

232 Linear Algebra

Is the differential operator defined by D(f(x)) = f(x) a linear transformation?

VesorG.y sete
The answer is in the next frame

9 Ves

Because:

D(axfi(x) + a2f2(x)) = farfi(x) + a2f2(x)]’

= arf] (x) + a2f5(x)
= a1D(fi(x)) + a2D(fa(x))
Let’s try another. Is the operator M acting on the vector space of ordered
pairs defined by Mx = M(x, X2) = (X1X2, 3x2) a linear transformation?
Y¥eS: OF NOs oc
tbe aoe

The answer is in the next frame

10
Because:

M(x + y) =M|(x1, x2) + (v1, ¥2)]

= M(x1 +1, X2 + y2)
= ((X1 + ¥1)(X2 + y2), 3X2 + 3y2)
whereas

M(x) + M(y) al M(x, X2) +M(y1, y2)

= (X1X2, 3X2) + (y1y2, 32)

= (X1X2 + Wi1y2, 3X2 + 3y2)
F ((X1 + ¥1)(X2 + 2), 3X2 + 3y2)
Therefore, M(x + y) # Mx+My and so the operator M is not a linear
transformation.
Next frame

|11 Some specific linear transformations

What follows is a list of some commonly-met linear transformations from R?
to R’.

The identity transformation

L(x) = x that is, for example, L(x1, x2) = (x1, x2)

The zero transformation

see = 0 that is, for example, L(x, x2) = (0, 0)

<
Linear transformations
233

Reflection in the x-axis

L(x, X2) = (%41, — X2)

Reflection in the y-axis

L(x, X2) al (—X1, X2)

Reflection in the line y = x

L(X1, X2) = (X2, X1)

Dilation or contraction

L(xX1, X2) = r(X1, X2) Dilation if r > 1 and contraction if 0 <r< 1

Counterclockwise rotation about the origin through @

L(X1, X2) = (X1 COS O — Xsin9, X1 SiN 6 + X2 COs 8)

There are many different types of linear transformation, but we are only
going to consider a restricted subset of them. We are going to restrict ourselves
to bijective linear transformations and the notation will indicate both the
domain and the range thus:
10 V1 ag V2

Next frame

Bijective linear transformations 12

To show that a linear transformation L : V; — V2 is a bijection we need to
show that:
if L(x) — L(y) then x=y
which is the same as saying:
if x Ay then L(x) 4 L(y)
That is, there is a one-to-one correspondence between domain and range
elements. For example, consider the transformation L from R? to R?
(L :R?-R’) defined by:
L(x) = L(X1, X2) = (2X1 — X2, X1 — 2x2)
We need to show first that this is a linear transformation. You can do this
now. Take it easy and one step at a time:
L(ayX + agy = L(ay(x1, X2) + a2(¥1, y2))

The answer is in the next frame

 Linear Algebra

" L(aiX + a2y) = aL (x) + a2L(y)

Because:

L(a,X + azy)
= L(ay (x1, X2) + a2(¥1, ¥2))
= L((a1%1, a2X2) + (21, a2Y2))
= L(a1X1 + azy2, a1X2 + a2y2)
= (2(a1X1 + a2y1) — (1X2 + a2y2), 1X1 + a2y1 — 2(a1xX2 + a2y2))
= ((2a1x1 — a4 X2] + [202V1 — a2y2], [a1xX1 — 204X2] + [a2y1 — 2a2y2))
= ((2a1%1 — aX], [a1X1 — 2a1X2]) + ([2Za2v1 — azy2] + [a2zy1-— 2a2y2])
= 04(2X1 — X2], [x1 — 2X2]) + a2([2y1 — yo] + [v1 — 2y2])
= a4L(x1, X2) + a2L(y1, y2)
= a,L(x) + a2L(y)
which proves that L is a linear transformation. To show that L is a bijection,
we consider the consequence of. demanding that:
L(x) = L(y) that isL(x1, x2) = L(1, yz)
it
L(x1, X2) = L(y1, v2) then (2x1 — x2, X1 — 2x2) = (2y1 — y2, yi — 2y2)

The answer is in the next frame

2X1 — X2 = 2y1 — y2
and
X4 — 2x2 =hy1 — 2y2

Because:

If (2x1 — X2, X1 — 2X2) = (2y1 — y2, yi — 2y2) then by comparing elements of

the ordered pairs
2X4 —xX2= 2y1 V2,

and

X1 = 2X2 =yi — 2y2

Therefore, from this we can deduce that:

Next frame
Linear transformations
235

15

Because:
2X1 — X2 = 2y; —y2 [1]
X1 — 2X2 =y — 2y2 [2]
4x1 — 2x2 = 4y; — 2y2 ZCtRI

3x1 = 31 2(1] — [2] so x1 =y1

Substituting in [1] yields:
2X1 —X2 = 2x1 —y2 SO X2=y2

This demonstrates that if L(x) = L(y) then x = y and so L is a bijection.

Let’s try another. To show that the linear transformation L : P2 — P defined
by:
L(a2x? + a,x +.ay) = aox” + a,x + a2
(interchanging the first and third coefficients)
is a bijective linear transformation, we first need to prove that is it a linear
transformation:
L(ayX1 + a2X2) = L(ay(a2x? + a1X4a0) + a2(b2x* + byx + bo))

Next frame

L(oyX1 + 02X2) = a4 L(: X1) + agL(x2) 16

Because:
L(a1X1 + a2X2) = L(a1(a1x? + a1Xx + ao) + a2(b2x* + byx+ bo))
= L((aa2x* + aya + 010) + (a2b2x? + arb,x+ a2bo))
= L([a1a2 + agb2]x” + [ara1 + a2b;)x + [140 + a2bo))
= [a1
dao+ azbo]x? + [a1a1 + a2b]x = [a1a2 + a2b2]
= 01 d0X? + a141X + aa2 + a2box” + a2b1x + arbz
= a1 (aox” + a,x + dz) + a2(box” + bi x + bz)
= a, L(x) + a2L(x2)
So L is a linear transformation. To prove that it is a bijection, we let
L(x1) = L(xXz)so that:

Next frame
236 Linear Algebra

17 : x) =X2

Because:
If L(x,) = L(x2), then L(a2x? + a,x + do) = L(b2x? + byx + bo). That is:
Aox? + aX + dz = box? +b1x+ bp

Equating coefficients tell us that a9 = bo, a; = b,; and az =b2. That is

Xi = Xo
This proves that if L(x; = L(x2) then x; = x2 and so L is a bijection.

Next frame

|18 Invertible linear transformations

If L is a bijective linear transformation between two vector spaces where
L(x) =y
then an inverse transformation L~! exists where:

L(y) =x
Note: If the linear transformation L is not bijective, then no such inverse exists.
For example, we have already seen that the linear transformation L ‘R?R?
defined by:
L(x1, X2) = (2X1 & X2, X1 — 2X2)
is a bijection. To find the inverse transformation we first let
L(x1, X2) = (a, b) where (a, b) = (2x, — X2, X1 — 2x2)
Now, L~!(a, b) = (x1, X2) so to find the form of the inverse transformation we
need to find x; and x2 in terms of a and b. This we do from the equations:
2X1 —X2 =a pity e 4x, — 2x2 = 2a 2 x [1]
PE eS ae Fi heea beac eos EP denial71
Subtracting yields 3x; = 2a — b giving x, = seca) Substituting in [2]:
s

X1 —2x2=b so x1 — b= 2x2. Thatis“4? _p = 2x, sox, <2”

Therefore:
2a—b
ic1 (a, ») == ( Orea-—2b ) that isso 1 Teel (1, 2) ma 2X Ao
= (A —Xo
3Xp Ky — xX 2
).

2
Linear transformations
237,
Let’s try another. Given the bijective linear transform L :-R?—R? defined by:
L(X1, X2) = (2X1 — x2, X1 — 2x2)
then the inverse transform is given by:

Next frame

L-*(x1, X2) = (u -2, =) 19

Because

Given L(x1, X2) = (x1 + X2, 3x2) = (a, b) where L~'(a, b) = (x;, x2) then:

and so x2= 3 and x; =a~—-. Therefore:

69,5) = 3 3

i-(a/pys
‘ a 3° 2)
(a2,eR 2). he Thats reite LsiZ)x2) = (x1 see 3

And another, just to make sure.

Given the bijective linear transform L :R°?—R° defined by:
L(X1, X2, X3) = (X1 + X2 + X3, X1 + X2, X1)
then the inverse transform is given by:
04; Sens haeeey Ree tees
fy re Al Pe ee ee )

Next frame

L~1(x1, X2, X3) = (x3, X2 — X3, X1 — X2) 20

Because
Given L(x1, X2, X3) = (X1 + X2 + X3, X1 + X2, X1) = (a, b, c) then:
Xi +X2+%3=a x1, =C
X1 +X2=b. Therefore x27 =b—c. Hence:
xT=C X3=a-—b
L7 (a, b,c), = (x1, X2, X3) = (c, b—c, a—b). That is:
L-*(x1, X2, X3) = (X3, X2 — X3, X1 — X2)
Let’s move on.
Next frame
(238 Linear Algebra

Linear transformations and bases

21 Consider the linear transformation L :R?—R* defined by:

U(x = L (x1, 42) = (2X1 = 3X2, 4X3)

Now consider the effect of this transformation on a specific ordered pair. Let
Ke Ai kt) = 2
Then

Next frame

22 L(2, 3) = (=9,. 8)

Because:
L(x) = L(x1, X2) = (2x1 — 3x2, 4X1). So that:
TA2513 EZ Bets x3; 4x2)
= (4-9, 8)
=(—S, 8)
If we now consider the transformation of the vector (2, 3) when it is given in
terms of the standard basis for R’, namely:
e; = (1, 0) andi e=(0, 1)
Then x = (2, 3) = 2e; + 2e2 = 2(1, 0) + 3(0, 1) so that:
L(2, 3) = L(2e, + 3e2)
= 2L(e;) + 3L(e2)
= 2L(1, 0) + 3L(0, 1)

Next frame

23 L(2, 3) = (—S, 8)

Because:
L(2, 3) = L(2e, + 3e2)
= 2L(e1) + 3L(e2)
= 2L(1, 0) + 3L(0, 1)
= (2, 4) + 3(-3, 0)
= (4, 8) + (—9, 0)
= (=, 8)
Linear transformations 239°

This last result obtained via the basis vectors can be replicated by a matrix
equation. We write the ordered pairs x and y as column matrices denoted by X
and Y respectively:
2 a
x= (3) and v= (8 )

and construct the 2 x 2 matrix from the effect of the transformation on the
basis vectors, namely:
Lil, Oye (2,4)
L(O0, 1) = (—3, 0)

from which the matrix L = (a ie) can be constructed where each ordered

pair on the right-hand side becomes a column of the matrix then the operator
equation

L(x) =y
has an equivalent matrix equation LX = Y. That is:
MS \ fa \ [=O
4 0 eh dlls GR

Indeed, the matrix L is said to be the matrix of the linear transformation L.

Let’s try another ordered pair.
Let x = (x1, X2) = (—12, 4) and again L(x) = L(x1, x2) = (2x; — 3x2, 4x1) so
that:

Next frame

TAt hei) 24
Because:

L(x) = L(x1, X2) = (2x1 — 3x2, 4x1). So that:

L(-12, 4) = (-—24-—12, —48)
= (-36, —48)
Also, recalling that:

ZL 0) 5 Lda oveeer aneerne Hint CS e 33, to be constructed.

HO} (3; 0) .
Then:

Next frame
re
 Linear Algebra

LG )G)-Cs)
The linear transformation L(x) = y can be effected between elements of the
two vector spaces by each element being written as a column vector and the
transformation being operated via the matrix equation LX = Y where:

t=(4 0) 2 -3

which is formed purely from the effect of the transformation on the basis vectors.
This means that as soon as we know how the basis vectors are transformed,
then we know how any vector will be transformed. ia

Example 1
The matrix that is associated with the linear transformation from L -R2R?
against the standard basis defined by:
L)(x) = L(x1, X2) = (5X2, 7X1 — X2) is L=
Next frame

Because:

L(x =L (24, x2) = (Sx2, 7X1 — X2) and so:

L(e,) = L(1, 0) = (0, 7) and
L(e2) = L(0, 1)= (S, =aEy

giving the matrix:

eet)
As a check to verify correctness with x = (—1, 3), it is found that:
Lixe lL], aye es and LX =

Because:
L(x = L(x1, X2) = (5X2, 7X1 — X2) so L(—1, 3) = (18, —7 — 3) = (15, —10)

Tae 4) =(88)
Linear transformations
241

Example 2

The matrix that is associated with the linear transformation from L -R?—R°
against the standard bases defined by:
L(x) = L(x, X2, X3) = (—x3, Xj + X2) is

Next frame

L=($ 0 a 28

Because:

L(x) = L(x, X2, X3) = (—x3, X1 + X2) and the effect of the linear transforma-
tion on the standard basis for R? is:

L(e;) = L(1, 0, 0) 7 (—1, 1 + 0) aa (0, 1)

L(e2) = L(O, 1, 0) = (—0, 0+ 1) = (0, 1)

L(e3) = L(O, 0, 1) = (—1, 0+ 0) = (—1, 0)

giving the matrix:

0 0 —-1
rs € 1 0 )
As a check to verify correctness with x = (1, 2, —S):
1b Qe ay Oe ee and LX =
Next frame

x(5, 3) and x= (3)

L(= 29

Because:

= L(x1, X2, X3)

L(x) = (5, 3)
= (—X3, X1 + X2) so L(1, 2, —5S)
and

Next frame
242 Linear Algebra

Invertible linear transformations and inverse

matrices
30 Having seen the equivalence between the operator equation
L(x) = y
and the matrix equation
LX =Y
It is evident that there is an equivalence between the inverse transformation:

x = L~*(y)
and the matrix equation:
X=L"Y
For example, consider the linear transformation from L -R?=R? defined by:

L(X1, X2) = (X1 + X2, X1 — X2)

Set against the standard basis for R?, the matrix that is associated with the
linear transformation L is:

Next frame

Because:

The standard basis is e; = (1, 0) and e2 = (0, 1) and:

L(e:) = L(1, 0) = (1+0, 1-0) = (1, 1)
L(e2) = L(O, 1) = (0+1, 0-1) =(1, —-1)
This gives rise to the matrix L = (; = )that is associated with the linear
transformation L.

The inverse transformation is:

Next frame for the answer

2
Linear transformations
243

nd _ (*A1+X%2 X1— X2 32
L~"(X1, X2) ( 7? 5 )

Because:

Letting L(x1, x2) = (a, b) where a = x; + x2 and b = x; — x2, then

L~'(a, b) = (x1, X2). This means that:
Xj +A2=a
yielding x; = ge and x2 = ate! so that
Xj —-X2.=b 2
Ieoig: b) = (S. >)

That is:
Aix X41 =~?)
Lie ap) = ( ow * 2

The matrix that is associated with the inverse linear transformation

Xie XD eK 52)
Ea, X2) - (
2h © Z
1S;

Next frame for the answer

ae eee 33
~\ 1/2 —-1/2

Because:

The standard basis is e; = (1, 0) and e2 = (0, 1) and:

This gives rise to the matrix (Vee

172 ibe
|) that is associated with the

e ‘ ‘ 2 Xz, +X 2X —X2
inverse linear transformation L~'(x;, x2) = ( =e ).

: ey i/2 12 : pees ath ag ;

Comparing the matrix Lie i with L = fh 2 ),it is easily shown
that:

Next frame
i
244 Linear Algebra

34 ; aK § Gs 1/2 )
L/2t 1/2

Because:
it eet!
Since Mi ( )then |L| = —2 and so:
1 =1

Ptah oT (iets)
This is the same matrix that is associated with the inverse linear
transformation
I(x, Xx
x2) = (7 2 x4 1 2)
Pyke ae?
which permits us to say that:
If the linear transformation given by L where L(x) = y has the associated
matrix equation LX=y, then the inverse transformation L~! where
x = L~!(y) has the associated matrix equation X = L"'Y.
Next frame

Some specific linear transformations

%

35 What follows are some commonly met linear transformations from R” to R’.

The identity transformation

L(x) = x that is, for example, L(x1, x2) = (x1, 2)
Thisas tepresented. Dy the matix Ge. oes=
The answer is in the next frame

™ (1 al

Because:
L(x) =x and so:
L(e;), = L(1,,0)= (1,0 and
1
L(e2) =L(0, 1) giving rise to the identity matrix I = € i)

This has the inverse I"! =.......:....

Next frame
7
Linear transformations 245

ce ==—(
Because:

Up
fe |) and so jf .= 1 therefore I Tw+7(6
t= (5 Fat1)=(6Wal1)=1

As you would expect.

Therefore, the inverse of the identity transform is the identity transform.

The zero transformation

L(x) = 0 that is, for example, L(x,, x2) = (0, 0)
This is represented by the matrix ............
Again, the answer is in the next frame

0- &

Because:
L(x) = 0 and so:
L(e;) =L(1, 0) = (0, 0) and
0 0
L(e2) = L(O, 1) = (0, 0) giving rise to the null matrix 0 = 6 6

This has no inverse because ............

Next frame

39 |
Because:
1
The inverse matrix is not defined because |0| = 0 and 0) is not defined.

Therefore, there is no inverse transformation to the zero transformation.

Indeed, an inverse would not be expected because the linear transformation is
not bijective.
L(x,) = 0 and L(x2) = 0 even when x # x2

Reflection in the x-axis

L(x, X2) = (%, — X2)
This is represented by the matrix ............
Again, the answer is in the next frame
nS EEE EEE
 Linear Algebra

Because:

L(x1, X2) = (x1, — X2) and so:

ie; = (1, 0) 41. 0}-and
1 a
L(e2) = L(O, 1) = (0, — 1) giving rise to the matrix L = (s )

hesinversesmatt sol. — =a ee

Next frame

Because:

se AL o ee le ae fil Os ye -
L=(4 2°.)and [Lj = 1 and so L =3(6 DEAG Zr

Therefore the inverse of a reflection in the x-axis is a reflection in the x-axis.

Reflection in the y-axis

L(X1, X2) = (—%1, Xq)

THs is represented DY 106 Matix oo oy
Again, the answer is in the next frame

uaa)-1 0

Because:

L(x1, X2) = (—%1, X2) and so:

L(e,) = L(1, 0) = (-1, 0) and
-1 0
L(ez) = L(O, 1) = (0, 1) giving rise to the matrix L = (0 is

The inverse matrix L~! =............

Next frame
Linear transformations 247

43
Because:

ih ee to mm AE Ee COR PY A at Sa
L=(% |) and Li =—1 and so L =4(4 =( ‘dete

Therefore the inverse of a reflection in the y-axis is a reflection in the y-axis.

Counterclockwise rotation about the origin through 0

L(x1, X2) = (x1 cos — x2 sin 8, x; sin @ + x2 cos 8)
This is represented by the matrix ............
Next frame

eee ms Ae}
sin@ cosé

Because:

L(x, X2) = (x1 cos 6— x2 sin 6, x; sin @+ x2 cos @ and so:

L(e,) = L(1, 0) = (cos@, sin @) and
L(e2) = L(O, 1) = (—sin@, cos 6)
giving rise to the matrix
Pe cos@é —siné@
TAN sin!” E08 8

The inverse matrix L7! =............

Next frame
248 Linear Algebra

ee
p= (96% 9 ) —6| sin[-0

Because:
cos@ —siné
)and |L| = cos?@+ sin*@=1
sin@ cosé
and so

cos@ —sin@\
L i=
sind cosé

1 cos@ sind
~ cos26+sin?6\ —sin@ cosé
cos@ sing

Ee —sin@ cosé

However, since
cos|[—6] = cos 6
and
sin|[—6] = —sind@
we see that:

Eee Cae rae el

sin{[—@] cos[—6]

Therefore, the inverse of a counterclockwise rotation about the origin through

6 is a counterclockwise rotation about the origin through —0, namely a
clockwise rotation through 0.

Summary
46 1 One-to-one correspondence
The domain and range of a function are said to be in one-to-one
correspondence if, to each domain element, there corresponds a unique
range element and vice versa — to each range element, there corresponds a
unique domain element.
2 Bijection
If a function possesses a domain and range that are in one-to-one
correspondence, then the function is called a bijection.
Linear transformations 249

3 Operators and transformations

An operator or transformation acts in much the same way as a function
except that the name is reserved for processes acting on functions.
Linear transformations
A linear transformation is an operator that acts upon the elements of one
vector space to produce elements of another vector space while preserving
addition and scalar multiplication.
Bijective linear transformations
A linear transformation is a bijection if there is a one-to-one correspon-
dence between its domain and range elements.
Invertible linear transformations
If L is a bijective linear transformation between two vector spaces where
L(x) =y which has an associated matrix equation LX =Y, then an
inverse transformation L~! exists where:
L~‘(y) =x has the associated matrix equation X = L"'Y

4 Can You?
Checklist 7 47
Check this list before and after you try the end of Program test.
On a scale of 1 to 5 how confident are you that you can: Frames
e Define and identify a bijective function? Ga)» (4)
Yastint hele Sul deheplelgnisl tell i eNO
e Recognize an operator and define a linear transformation? (5) to
Yesnow (ebett ls an lelver rtd inelsdleren sie
e Identify some specific linear transformations? G1)
Veg ee lhel edlinl anole ie iviletated| Caen’
e Define and identify a bijective linear transformation? G2) to
Yes EC) oO Clenoledenelen ae
e Derive the inverse of a bijective linear transformation? to (20)
es aif lee OT5) Ve TES Es} % 1a). * Ne
e Derive and appreciate the effect of linear transformations on a
basis? GD » G4)
vy aS We ae) Ce se, 0d a |
e Obtain the matrix associated with a linear transformation? to (29)
Yes Gi vey J. No
e Appreciate the relationship between an invertible linear
transformation and the invertibility of its associated matrix? to (45)
Yes eee) Co LL.” No
 Linear Algebra
250

Test exercise 7
|48 1 Which of the following are linear transformations:
34R?
(a) L:R3—R? where L(x1, X2, X3) ae
= (%41 — X3, 0, x2 )
(b) L :R3—R? where L(x1, X2, X3) = (—%3, 9, |x2l)
(c) L .R3—R? where L(x1, X2, X3) = (3X1, 1+ 5x3, 2x2)

2 Which of the following linear transformations is bijective:

(a) L :R°R?* where L(x1, x2, X3) = (%1 — X3, X1 + X2, X1 + X2 — 3)
(b) L: £ + £ where L(x + yi) = (x — yi) where x, y are real and i? = —1
(c) L :R3R? where L(x1, x2, x3) = (1 + X3, X2)
3 Find the inverse of each of the following linear transformations:
(a) L :R?—R? where L(x1, X2) = (x1 + 2X2, X1 — 3X2)
(b) L :R3—-R? where L(x1, x2, X3) = (1 + X3, X1 — X2, X1 — X2 + X3)
(c) L : Pz — Pp where L(a2x? + a1X + do) = (dox* — aX + a2)
4 Find the matrix associated with the linear transformation for each of the
following: ;
(a) L :R?—>R? where L(x, x2) = (2x1 = 3x2, 4x1 + 5x2)
(b) L :R3—R*® where L(x1, x2, X3) = (%1 + X3, X1 — X2, X1 — X2 + X3)
(c) L:R3—-R? where L(x, x2, 3) = (X41 + x3, X1 — X2)

5 Which of the following linear transformations is invertible:

(a) L :R3R? wher® L(x1, X2, x3) = (x1 — 2x2, x1 + 3x3)
(b) L :R2—R” where L(x1, x2) = (x1 + 2X2, 5x1 + 10x2)
(cee ‘R3-—R? where L(x1, X2, X3) = (%1 +.X3, X1 — X2, X1 — X2 + X3)

Further problems
49 1 Which of the following are linear transformations:
(a) L: Py — Py where L(p) = p(x)°
(b) L:P, — Py, where L(p) = 2p"(x) + p(x)
2 Show that L : P2 —R is a linear transformation where:
1
L(?-+bx +c) = | (ax” + bx +c) dx
x=-1

2
Linear transformations
251

3 Which of the following are linear transformations:

(a) L: Py — P, where L(p) = x*p’(0) + p(0)
(b) L: Py — Py where L(p) = xp’(O) + p(0)
(c) L: Py — Py where L(p) = xp/(0) + 2

Let L: P2 — P3 be a linear transformation for which L(1) = 1, L(x) = x? and

L(x”) = x3 — x. Find L(ax? + bx +0).
Given the linear transformation L -R?—R? where L(1, 0,0) =(1, —3, 2),
L(0, 1, 0) = (0, 4, —0) and L(O, 0, 1) = (2, 0, 0), find L(1, 1, 1).
Describe the geometric effect of each of the following linear transformations:
ake ‘R?—R? where L(x1, X2) = —(3x1, 2x2)
(b) L ‘R?—R? where L(X1, X2) = 6(X1 + X2, X1 — X2)

Find the matrix associated with the linear transformation

L :R°=R* where L(x1, x2, 3) = (x1 +3, X1 — X2, X1 — X2 + X3)

with respect to each of the given pairs of bases:
(a) From (1, 0, 0), (0, 1, 0), (0, 0, 1) to (1, 0, 0), (0, 1, 0), (0, 0, 1)
(b) From (1, 0, 0), (1, 1, 0), (1, 1, 1) to (1, 0, 0), (0, 1, 0), (0, 0, 1)
(c) From (0, 0, 1), (0, 1, 0), (1, 0, 0) to (1, 0, 0), (0, 1, 0), (0, 0, 1)

Find the linear transformation L :R’— P2 from the standard basis for R* to the
2 -2
basis 1, x, x* for second order polynomials for which L = (: 2 is the
Lew
associated matrix L.

[Hint: use the column elements to find the effect of the transformation on the
basis for R”].
Show that if L is an invertible linear transformations then so is its inverse L~'.

10 Find the condition satisfied by a1, a2, 3; and (2 that ensures that the following
linear transformation is invertible:

L(xX1, X2) = (a1X1 + 2X2, 81X1 + 2X2)

11 Find the inverse transformation L~! where:

L(X1, X2, X3) = (X1 + X2 — X3, X1 — X2 +.X3, 2X1 — 3x2 + 4x3)

 Answers

Test exercise 1 (page 24)

tae 3 (Det 2 Cee (hy) 2 ayes
Ww
dN]
(b)
x= 6.5 p=—-105 3 x=3, y= =2,z2=5° 4 2='1, y=3, z=-5

Further problems 1 (page 25) |

5) 10 3 2 11
1 ease 2 x== 3 -x=-3 4 AS 5 x=3 Xx

3
7 x=-75,y=-14 8 Rs 9° x= 21910, 2 y=2,.2=3
Ll 226,97 =7,2=5 12) 2=2,7 ——2, 2—0° 13 "25, S0,2=2
3 1 35 l l
14 x =-— a — 5
te 158 x =-—6 16 x=4
— we" Ty Fx
x=5 »yo=4
ak 18 *x
x=-——, is )¥ 30
=

19 x=-14,y=-1 20 x=-0.5, y=0.25 21 (a) Inconsistent (b) Inconsistent

(c) Consistent

Test exercise 2 (page 58)

1 (@)4 (b) 18 2 x=2,y=-1,z=4 3 x=3, y=-2,z=-1
4 k=30r-25 § x=3, 1.653, —6.653

Further problems 2 (page 59)

1 (a) 144 (b)0 2 (a)0 (b) 666 3 x=5,y=4,z=-2
4 x=25,y=3,z=-& § x=2.286, y=1.095,z=-3.476 6 40r-14
7 Sor-2.7 8 (a) Oor+v2 (b) (a—b)(b—c)(c—a)(a+b+o)
9 x=lorx=—-54+V34 10 x=-1.5 11 2(a-—b)(b-c)(c—a)(at+b+c)
12 i2=5.2 13 (at+b+c)*(a—b)(b—c)(c—a) 14 2o0r-16/3
15 (x—-y)(y—z)(z—x)(x+y+z) 16 x=-3 0or+v3
(2M; + M2)W 70 llr
De = ——————_
Mi (M; + 2Mp) 18 4, = 0, io2 = 2, iz
iy 23=3 20 @ = —12%
or ——
12

Test exercise 3 (page 88)

5 9 8 10 Se he:
eat Gane 6 ) (b) (7 6 -6 t)
21 45 -19
2 (a) ee ut a) (b) & se ( 48 0 %)
-17. 35 -37
ste Bae,2 ee tee
2 AD | AO WA a 13 701 be AP
5 4 6 Wer ea
35031. ald 74h =15, 4.75
4 |a\=0 5 (a) 105 (b) Ce wie et 1 2.5
Rd) ki LPO

@
Answers
253

W) +4 —5 X\ —7
7 1 -3 1 ‘ X2 = 10 8 Ky = 23 Xo = —33' X2=5
ey 5 3 X3 2

9 (b) X1 = —]; x2 = -3; X3=4

Further problems 3 (page 89)

11 -
CRG ee ets a7)°0 (fetoat tah (2546)
(d) -I 3 (a) ies . (b) $ ey)
2 (a)C (b)A (c)B 2 -1 7 el 3
- - 2 -3 -1 x) 2
(c) ( a9 4 kKk=-2 S)da) 4 Ai e2ils{ xm] =13
, a OSs Soe! x3 5

1 -2 -1 3 xX, 10
PE foe” Kon © al Aaa, 8 righ is. bee
(b) 1 Oy ed hee TN | eae 3 oe A aeSee he |

0 -1 3 1 X4 —7
7 x=), y=2;z=-1 8 i, =2;i2=1;i3=0 9 no unique solution
10 4 = 2) x%2= 15; %3=-35 11 f=05;5b=1.5713—10
12 4, =3.0; i =2.5;i3=-4.0 13: i; = 2.26; ip = 0.96; iz = 0.41
LA 2 Ong 0s = —20
15 ey oN peta bee where Z = 2122 + 2223 + 232)
Zi Zi Zz

Test exercise 4 (page 153)

1 (a) solutions unique (b) infinite number of solutions 2 x; = —4, x2 =2,
x3 = —-3 3 Xi nD, Loe e Kate 4 X, = —3, x2 = 4, x3 = —2
1
5 My aL, Xo Aptos 6 Ay =1, 42 = -2, A3 = 3. n-| |
1
il 5
ui ei. a 1b Pe {ts ES 1S
vid) | | 8 bello ee Z, 7 M=| > 1|iM a) vA
3 —1
M~!AM = ~ a 8 fi(x) = _10 e* 4 =se; fa(x _>SC tteas
0. «5 ye

9 SCRE et Me RE
bP as 3/5 3 3
+ 1
(x)= -=700s V5xpe
35sign ex te3 cosh
2x + r—sinh 2x

-8 7.196
ett |‘ (b) aig
Further problems 4 (page 154)
1 x, =1,%2.=-4,x%3=3 2 (a) 1 =3,xX2=1,%3=-4 (0) 1 =4,%2=-2,
x3=-1 3 (a) x1 =4, X%2 =2, x3 =S,X%=3 (b) 1 =S, X2=-4,%3 =1,
X4=3 (C) X%1 =3, Xo = —2, X3 = 0, % =5
254

(a) xx7=—3,
% =1,%3=3 (Dee 5, Xo. = 2, X39 == —-1 TS) Xi 4h Xo = 3,
r 3 1
x3=-l, Xy»=-—2 5 (a) Ni 25, \2 = 7; X1 = [jl | (b) Nea

5 iit 5 1

Ar 4
X%1 =
(d) A1 = 4, A2 = —6; 0, =9
(e)' Ay = 1, 22 =3323

heels i:‘i UIE

7 1 0 1 H
X= HL 1 (£): As rae Lee aL ed

ELL
() A= =k =3, 772X= - |-At2 (h) As —2, 497;
0 10 =

@) fils)=4(See); fla) =Glee)

(0) filx) =2 (e** —e); fx) = GC + 9")
(©) fils)=5(Se — 36); flo) = Fe Se+20
f3(x) = 4e*— se + ne (d) fi(x)= 3e-% — e* + 2e”*;

fo(x) = —e7* Seed +56 f3(x) =e * —Se* — 3a 7 210,07 13

%

R= PF by S85 bee 9 | ems YAP 5 —2, x2= 5) Bye

9 il
10 a x cosh V2x + —- sinh V2x+5adie Ge ee sinh os
@) A= 5/2 5 5V7
2 6 2 tt
x cosh /2x ——=sinh V2x + cosh V7x +—=sinh V7x
fa(x) = 5V2 5V7
R) 5 1
(b) ec: ae >,cosh 2x + = sinh 2x;
1 i s) dy,
Pie sie E08 yon an 2V2x + et 2X. + g sinh 3

(c) nh (x)=
—
1 cosh x + 1
—
12
Sh
sinh x + — V3x —_—
1 Nees
——_—— ]
V3x

13 7 5 Tee, 5
+ Gg cosh 3x + Taq Sinh 3x; (x)= Te coshx — 7g sinh x + 75cosh v3x
ea: 13 a ZS 5
— Fagg sinh v3x+ 4g Cosh 3x + 77; sinh 3x; AB) = Tg cosh — 7, sinhx
7 OBS te & 65 a
Eien ee
Answers
255

9 ae
(d) fix) = ~ gcosx + Fsinx +77
Cchad are aqiin V3x + Z cosh 7x
3 , : 15 15 = =.
+
sinh V7x; f2(x) = 16 605% 3 sinx — 55.0 V3x + sin v31
20V7
81 oF, : 3 3 =
+ —cosh V7x + sinh V7x; fa(x) = — 3 COSx +$sinx +3 cosh V7X
80 40/7
1
4 apes
sinh /7x
4/7

Test exercise 5 (page 173)

1 AB = 2i — 5j, BC = —4i + j, CA = 21 + 4j, AB = V29, BC = V17, CA= V20
4 (a) -8 (b) —2i—7j-—18k

Further problems 5 (page 173)

we ey als
1 OG==(10i+2j) 2 3,4, 5); (1,2, —3); 6 = 98°42’

3 moduli: /74, 3/10, 2/46; direction cosines: wot (3, Tn, as Se

V74 3V10
1
—=(3,-1,6);
7H. )s sum =10i 4 BC—AB, —AB, —BC, AB—BC

1 Mp al Sachs

Test exercise 6 (page 224)

1 (—2,5,28) 6 ax*+fx+7
7 Linearly dependent: (x + 137 =4x1+4+4(x-1)+(1-—x)* 8 Dimension =5

Further problems 6 (page 224)

1y.12—.7%+29x7 2 x= (>, ~ -21) 3 Yes 7 Two
8 Test for linear independence: 1 — 2x, 2+ x-—3x?, —4+2x, —8+6x?, 10x”, —3
11 (a) Yes (b) Yes (c) No (d) No 12 (a) No (b) Yes (c) Yes (d) No
14 (a) (a, b, 2c,c) (b) (4a — 2b, 3a+b, 2a, a + Sb)
f : Led 1 0 Ves Nie eee ,
16 (1,0, —1), (0, 1, 0): dimension 2 17 & oh " y (; a dimension 3
18 1-—x?,x-—x?: dimension 2 19 (2, 4, 3)
20 (a) Subspace of S(R3) spanned by (1, 1, 2) (b) Subspace of vector space of those
1
3 x 1 matrices spanned by (-s) (c) Subspace of S(R*) spanned by (—1, 2, 2, 2)
«!
(d) The 4 x 1 zero matrix

Test exercise 7 (page 250)

1 (a) Yes (b) No (c) No 2 (a) Yes (b) Yes (c) No

3 (a) L-) :R2->R? where L~"(x1, x2) = oe x1 es)

where ites, X2, X3) = (X1 — X2 +3, —X1 +X3, X2 — X3)
(b) iAas -R3=R?
(c) L~! : Pz — Pz where L~!(a2x? + a,x + ao) = aox® — a,x + a2
256 Answers

elite
4 otal =) me (1 =I | @u=(] as a
heey
5 (a) Not invertible (b) Not invertible (c) Invertible

Further problems 7 (page 250)

1 (a) No (b) Yes 3 (a) Yes (b) Yes (CQ. Nor 4 ax*?+bx*—ax+.c.,.5 * (3, 1, 1)
6 (a) Stretching by a factor of 3 in the x; direction and a factor of 2 in the x2, followed
by a reflection first in the x; axis and then in the x2 axis.
(b) Stretching by a factor of 6V2 in both the x; and x2 directions, followed by a
clockwise rotation of 45 degrees.
ihe Tiies| ie 2 ari pi
ya eWAP NES 2c lehabeideyi aac feyguid O(a) CORE
oyEire
lee lap terse! leet .
8 L(a, b) =2a—2b+2bx+ (a+5b)x2 10 02162 4 arf
X, +x —5
11 hase oy X2, x2) = (, a Ni — 3x2 + X3, a 2 “2 +x)
 Index

Addition and subtraction of matrices 66 Direction cosines 171

Addition of two-dimensional geometric Double suffix notation 64
vectors 175
Addition of vectors 160 Eigenvalues and eigenvectors 120
Adjoint matrix 104 Elementary operations 94
Adjoint of a square matrix 75 Elementary row transformations 107
Associativity 176 Equal matrices 65
Augmented coefficient matrix 98 Equal vectors 159
Augmented matrix 82 Equating coefficients 6
Axiom 184 Equivalent matrices 94
Evaluation of a third order
Basis 238 determinant 37
Standard 218 Existence of solutions 14, 99
Basis vectors 210
Bijective linear transformation 233 First order ordinary differential
equations 128
Cartesian unit vectors 167 Free vector 160
Cayley-Hamilton theorem 127 Function 228
Characteristic determinant 121 Co-domain 228
Characteristic equation 121 Domain 228
Cayley-Hamilton theorem 127 Range 228
Characteristic vectors 120
Characteristic values 120 Gaussian elimination 82, 111
Clockwise rotation 247
Cofactors 74
Homogeneous linear equations 121
Column equivalent matrix 94
Column matrix 63
Commutativity 176 Identity element 177, 183
Comparison of methods 120 Identity transformation 244
Components of a vector 163 Integral operator 230
Consistence of a set of equations 49, 98 Inverse matrix 76
Contraction 233 Invertible linear transformation 242
Inverse method 103
Invertible linear transformation 236
Determinant 29
Inverse matrices 242
Characteristic 121
Minor 36
Properties 53 Latent roots 120
Square matrix 73 LCM 3
Determinants of third order Line vector 160
Evaluation 37 Linear combinations 200
Diagonal matrix 71 Linear dependence 204
Diagonalization of a matrix 133 Linear equations 2
Differential operator 230 Homogeneous 121
Dilation 233 Solution 79
Dimension 210 Linear independence 204, 208

257
 Index

Linear transformation 230 Transpose 70

Basis 238 u Unit 71
Clockwise rotation 247 Upper triangular 96
Dilation or contraction 233 Matrix algebra 91
Identity transformation 244 Matrix equation
Matrix of 239 Characteristic equation 121
Reflection in the x-axis 245 Eigenvectors 120
Reflection in the y-axis 246 Matrix notation 65
Zero transformation 245 Matrix transformation 145
Lowest common multiple 3 Rotation of axes 147
Mathematical structures 183 Minimal spanning sets 213
Matrix 63 Minor 36
Addition 66 Modal matrix 133
Adjoint 75 Multiplication of matrices 66
Adjoint 104
Augmented 82 Non-existence of solutions 17
Augmented 98 Non-singular matrix 92
Characteristic values 120 Null matrix 72
Cofactors 74
Column 63 One-to-one correspondence 228
Column equivalent 94 Operator 230
Determinant 73 Differential 230
Diagonal 71 Integral 230
Diagonalization 133 Linear 231
Double suffix notation 64 - Order of a matrix 63
Eigenvalues 120 Ordered n-tuples 196
Elementary operations 94
Equality 65 Polynomials 191
Equivalent 94 Position vector 160
Inverse 76 4 Post-multiplication 69
Latent roots 120 Product of a square matrix and its
Linear transformation 239 inverse 79
Minor 74 Properties of determinants 53
Modal 133
Multiplication 67 Rank of a matrix 93
Multiplication 66 Reflection in the x-axis 245
Non-singular 92 Reflection in the y-axis 246
Notation 65 Right-hand rule 169
Null 72 Rotation of axes 147
Order 63 Row equivalent matrix 94
Post-multiplication 69 Row matrix 63
Pre-multiplication 69 Row transformation method 107
Rank 93
Row 63 Scalar 158
Row equivalent 94 Scalar multiplication 66, 179
Scalar multiplication 66 Identity element 183
Singular 73 Properties 179
Singular 92 Second order ordinary differential
Skew-symmetric 71 equations 138
Special 71 Set of equations
Spectral 133 Consistency 98
Square 71 Existence of solutions 99
Subtraction 66 Solution 103
Symmetric 71 Uniqueness 99
Transformation 145 Uniqueness of solutions 99
<
Index 259

Simultaneous linear equations with three Uniqueness of solutions 99

unknowns 8, 19, 40 Unit matrix 71
Pre-simplification 10 Unit vectors 167
Simultaneous linear equations with two Upper triangular matrix 96
unknowns 5
Solution by equating coefficients 5, 6
Vector 157
Singular matrix 73, 92
Addition 160
Solution by equating coefficients 6
Components 163
Solution by substitution 5
Components 167
Solution of a set of linear equations 79
Direction cosines 171
Solution of sets of equations 2, 103
Equal 159
Comparison of methods 120
Free 160
Gaussian elimination method 82, 111
Line 160
Inverse method 103
Position 160
Row transformation method 107
Sum 161
Triangular decomposition method 114
Types 160
Solution spaces 218
Vector addition 175
Spanning set 200, 202, 204
Associativity 176
Minimal 213
Commutativity 176
Special matrices 71
Identity element 177
Spectral matrix 133
Vector representation 159
Square matrix 71
Vector space 184
Standard basis 218
Axioms 184
Subspaces 196
Basis vectors 210
Sum of a number of vectors 161
Dimension 210
Systems of first order ordinary differential
Matrices 185
equations 128
Ordered n-tuples 196
Systems of second order ordinary differential
Polynomials 191
equations 138
Solution space 218
Standard basis 218
Test for linear independence 208
Subspace 196
Transformations 230 Vectors in space 169
Transpose of a matrix 70
Triangular decomposition method 114
Types of vector 160 Zero transformation 245
 ‘yw a\

:
pf

H
eB
*)j

‘ : 2,

Fi

re 2
'
{i ( Th
f

» Le ! i
: ie i : ’ .
i ie . bt
<2 f ;
7 if¢ue
vi
» t 4

2 wi j he! . “= ) werk

Fuad : ~~
. ae es y 4 bphar yey rz2e acilR P
Pap ’ : -
Hi eye Nol rik
hee ~
eT wu P one fk : wriiser »
‘; ae :
a. TEV ier rer A , he °
é 4 i 7 fs wir a4
-_"
fi é STVTISIUIG
- f } J
TNO ; rT os iiten: lake

Yt!
Lis 7760
JESS {3s teres) et. sien inte:
- Go 9 Metronet in)
ee : pony ; ™ bi fe 1. ahenar sisupe
;
{j BIS was Hiabrietc
. ” . i] pa
orl arid c ; : ural 7
/ a sila (ual oro. nen slaids a Lin —s : vel ssompuor
» a : a eat rowelr to rettraye-s to oud
; 7 ™ Off” palenvetiC Cintierisite
pen tiiertsbyo teil toetrateye
- P hl=e ine i ; get paphiah + a
Bot antileepy
- » percent tony ‘pills bain JAttpretiifepstibys ino hrwease to eimoreye.
“A

maltion) Getter. 1A0-, BEL’ erpatieups— 5

‘- fontnitiplcation &2 >
 Using the same innovative and proven approach that h Ss
LINE AR made the authors’ Engineering Mathematics a worldwi ¢
bestseller, this book can be used in the classroom or as an iin ©
ALGEBRA depth self-study tutorial. Its unique programmed approa in
K. A. Ctr oud _ patiently presents the mathematics in a step-by-step fashi n
with additionsby ' together with a wealth of worked examples and exercises. |It
Dexter Booth also features Quizzes, Learning Outcomes, and Can You?
checklists that guide you through each topic and reinforce
learning and comprehension. Both students and professionals ©
alike will find this book a very effective learning tool and 4
reference. q

Distinctive Features |
® Uses a unique and proven “programmed” approach that takes you through
each topic in a step-by-step fashion with a wealth of worked examples and exercis
© Serves as a "personal tutor", allowing you to proceed at your own rate of learning —
and master each topic before going on to the next. ©
© Provides many opportunities for self testing through numerous Quizzes, Learning2b
Outcomes, and Can You? checklists. 4 a)
® Equally effective as a classroom texthook or a self-study tutorial. : i| |

ee
of ™

About the Authors

K.A. Stroud formerly was a professor in the Department of Mathematics at Lanchester
Polytechnic (now Coventry University) in the United Kingdom. He originated the hugely
popular Engineering Mathematics (ISBN 978-0-8311-3327-6) and Advanced Engineering
Mathematics (ISBN 978-0-8311-3169-2), both available from Industrial Press.

Dexter J. Booth formerly was principal lecturer at the University of Huddersfield, UK.
He is the author of several mathematics books, including with Stroud, Engineering
Mathematics, Advanced Engineering Mathematics, Differential Equations (ISBN 978-0-
8311-3187-6) Vector Analysis (ISBN 978-0-8311-3208-8), and Complex Variables
(ISBN 978-0-8311-3266-8), all available from Industrial Press.
a ee

ISBN 978-08311-3188-3

||
iiii
a_FeUeST HE = a