-

series
MATRICES

CAYLEY, a French mathematician, discovered matrices in the year

1860. But it was not until the twentieth century was well-advanced
that engineers heard of them. These days, however, matrices have been
found to be of great utility in many branches of applied mathematics,
such as algebraic and differential equations, mechanics, theory of electrical
circuits und in statistics. Matrices are useful because they
"u.u"-fitting
enable us to consider an array of many numbers as a single object, denote
it by a single symbol, and perform calculations with these symbols in a
u".y form. One of the most important applications of matrices is
lrx "o*puit
the treatment of linear equations. It is, therefore, necessary to learn the
elements of matrix algebra in order to keep up with the fast developments of
19x3 engineering.
3

DEFINITION. A matrix is a rectangular array consisting of elements written

in rows and columns :'':ellV enclosed by a bracket. If the matrix has m
rows and n columi:,s, ,ie matrix is an mxn matrix. A matrix is also
denoted bY a sing-: ;.::ial letter'
Co)x2
Thus,

is a matrix with :: ::;; i :--i :r columns and is called an mxn matrix, (read
"m bY n matri!.'
The numb.: . r - iie called the elements of the matrix, The hori-
zontal lines r. -rL":: ::;-. .:,d the vertical lines are called columns of the
matrix. To lr,r;:l; :i-i :,:-.-:"iiar elements of a matrix, the elements are de-
noted by a l::::::-.,- ; :,r r) tu'o suffixes which respectively specify the
rorvs and the ;: -;r:-r .. I:,':s. r-, is the element in the ith row and the jth
column of A

--''g
TYPES OF MATRICES
number'of rows equals
is a matrix in which the
.Jo'ii*n,'-rt'"
1) SQUARE MATRIX - is called its order. For
the numb", of "Y#';|;;ws
3 is'
a square matrix of order
"-"*pfe' -'t
&rz u"\
[r 8 t fu" ut'l
lo + 5l \ut' ^zz
t, t nl
,tz u"f
Lu,'
is
' elements all' &22' a,33
diagonal of this matrix-contarining the
The
called the leading; or PrinciPal diagonal'
the diagonal ele-
MATRIX is a square tnatrix in which
2) IDENTITY - the off:diagonal elements are zer(,. For exam-
ments are "one"
""l-"il
P'*,;i;"ttitY matrix of order '3 is
--
lr o1 o\
io o ol
io
L
tl
'-n" columns'
is a matrix having only one row and
3) ROW MATRIX -
It is also called a row vector'
Thus,

A=[u, r"

a matrix with "m" rows and

ml;' one col-
MATRD( is
4) COLUMN - vector'
umn' It is also "at"a a cslumn
Thus,

Iu'l
o= \", \

.*l
 matrix for which every term off
DIAGONAL MATRIX - is a square
rhe main diagonal is zero' For example'

[: o tio\ B- [r o', ol
\= o I 'na l: 'jl-i
o o 2\
i

Io "
matrix can also be
lre diagonai matrices of order 3' Thus' an identity
considered as a diagonal matrix'
on the
bt SCALAR MATRIX-- is
a square matrix for which all terms
marn diagonal are equal' For example'

roo $= lo
[rool
I 0l
\= 4 .l I .rnd
m- .iorLoorj
matrix can be consid-
;r: .:a1ar matrices of or'ler 3 Again' an identity
.lS scalar martrl\'
=-: -
I:ROORNULL\1\T'R'IX-isamatrixwhereinallelementsare
-.:-', Thus,
I
nns.
r=" [t i, a zero or null matrix'
lo
TRIANGULAF:
. .lRlX is a square matrix whose elements above
the princiPal : -:
-:. tr -beiow the principal diagonal) are zero' For
L

example,
f-
+ ls o o

n- .rnd B- t8
l

2 0
;ol-
]o 3 4-
I

are uPP;: --: : rsular matrices resPectiveh"

ESUALITY oF l"1A]' E5

Two m::1.;:s i' -- ::: said to be equal if and onlY if

i I th:r ::; , i -- - s;lme order, and
element of B '
ii' :.i;1'': =- :; A is equal to the corresponding

-
 -
ADDtTtoN AND suBrRActtol.')F MATRT.ES

-:- matrices A and B c-an be

ad1e1J.o^1:*'lTTiltl:':?,ili,::'li;
Twom{lcesAanot'carr""":";;;n"OUyaddingorsubtracting
rwo
or rs difference I

same ordet' Their sum

their corresPonding elements' order'
matrices A'lB and C have the same
If one assumes that the
(Commutative)
,
A+B=B+A (Associative law)
B + (A- C) Theref
(A + B)-Q = A + (B - C) =

MlTRlxBY A SCALAR
MULTlpLtcATloN 9t o whene each element
product of a matrix A by a scalar K is a matrix
The,
tru*tntt of A'
is K times trt" "o""lj;;
Thus, TRAI
Ta, -.r T Kb' .. .-1
rc'
l br ",1 --iKa'Kb'Kc4
l Ka' 1

Kl", b2 - ",i
1

the ro
L" -] r-
1
1"e''
holds for such^ products'
The distributive law

K(A+B)=KA+KB

PROBLEM:
Evaluate 2A - 38 whcre

A-
hrr-l
t'o I Ol and $=
f,
lr
1

-2
i
0l
fa- t-
It 4
' t
3-"1
It
L
0 3j
l_
i.e.,
Solution: -t
2:
[r I oi;i lz
lt
1

li
I

2A-38=z\z I -3 n

,j I

\t 4 3J Ir 0
- fo z | [o 3 6-l
2A-38 =lq 2 ol -lt -6 oi
lr 8 r,l [: o,l
if the
rcting

er. [u-u 2-3 2-e

ZA-38 =l+-3 2+6 o-ol
2-3 8-o 6-e_)
il)
Therefore,

f
'0 -l 41
2A-38=',7 8 0l
I

[ement
-i8 -31 j

TRANSPOSE OF A MATRIX
The matrix Ar is called rhe transpose of A obtained by interchanging
the rorvs and columns of A.
Thus,

lr-i[:2
t-_1-
'. I
ifA= 2 0 .thennt=ll I 4l
I' 3,,1

Also,

*T
:f A
It ar i rhen A' = L7
'l
A square n:1r:-*r -{ s s.ard to be symmetric if it is equal to its transpose,
i.e., AT = A.
Thus, the mlxtla

s ,-l
{= a -1 2l is a syrnmetric matrix,
;11_l

-
 [28'l
et=l8 3 2l=A
\o 2 r)
MULTIPLICATION OF M,ATRICES
and onlY if
multiplied in the order AB if
The matrices A and B can be of B' Such
to the number of rows
tf," nu*U", of columns of A' is equal
*utJ""t are said to be conformable'

MULTIPLICATION
PROPERTIES OF MATRIX
and is distributive with respect
a) Matrix multiplication iS associative
to addition of matrices' that is' 1

(KA)B=K(AB)=A(KB)
A(Bq) = (AB)C
, (A+B)C=AC+BC
C(A + B)= CA + CB

that the expressions on left are defined' (K

provided A, B and C are such :he
is any number)
that is' if A and B are
b) Matrix multiplication is not commutative;
matrices such that Uottt eg and
BA are defined' th'en'

AB + BA in general
general' that AB = 0 does not
c) The cancellation la*' is not true' In
necessarilY imPlY A = 0 or B =
0'

d)Thetransposeofapoductequalstheproductofthetransposed
factors, taken in reverse order' So that

(AB)r = BrAr
(BA)r = ArBr
(AB)r + ArBr
(tsA)T + BTAT
 PROBTEM:

3lonoB=[t 3-l ,rno AB and BA.

rra=12 ro_i
lr L5 -2)
Solutiort:

Since the number of columns of A is equal to the number

of rows
a)
of B, then the two matrices are conformable'

ft
AB=l-
rl[r
-ll 3-l
I

L3 l0_l Ls -2)
| 2t lr + 3t5) t1', *-11-?,.-l
AB =
.r(l) + l0(5) 3(3) + 10(-21

Therefore,

t- r
AB=
:: -

b) Since the :.'--:r: :: :,-irnns of B is equal to the number of rows

of A, thei- :-: - ::.:-:"J:s are conformable and BA is possible"

B-d

B-
i
Therefc::

-
PROBLEM:rLEwr'
---1 [r -z 4 and BA'
Ir -2 - '
3lunas=13 5 t\ -l , find the AB
rr ,t =
\_+ z
t-
sJ""* [z 4 oJ

s o t ut i a n : tfff mbers of rows

S ince th e number or "
of B' then the two
::l:1'
matrices ar
":"t;
t iil?ff]'?
1
ruX"nl "t*
-'[t -2
r\\r
+\

AB =L;
Ir -z s 1\

1(4) + (-2Xl) + 3(0) \

--- ,e\/^\ t96
:

[,t,r * (-2lg) + 3(2) r(-2):(-.2]3i !?l\" + 2(1) + s(0) J

AB =
L-l1i; ri;;*
stzl 4(-2) +

Therefore'
I-r
AB=l '
o ,l I

b*:
b),:" .:;:i"ilf,,:i,","
then the two
# :H?'sHTKi: m:
rows of A;
not defined:

PROBLEM:
[r -z tl [rl
u =lo 3
:

IfA= \ , I -tl I+ -s\

\

\-,
L
t 2) t-_

a) AB &BA
e) BBr
Find: f) BTB
b) Nn
e) (AB)r
c) BrA
;; 6z-lr+eI
d) A2

of A is equal to
the
sol'uti::: of'columns
the number
AB is defined since
nomb"r of rows
of B'
-
 [r -z :l[r f
o"=l z 3 -'llo 3l
BA.
[-s t 14 -s_]z)

I t,r, + (-2)(0) + 3(4) t(2) + (-2x3) + l{-s)-l

AB = ztrl + 3(0) + (-tX4) z(z) + 3(3) + 1-r)t-s)
|
L-3(l) + l(0) + 2(4) -3(2) + l(l) + 2(-5)
|

rows __l

;ible.
Therefore,

[r: -rql
AB= l-, r8l ;

I r -tr]
3(0;l
(o) BA is not defined or non-conformable since tlle number of col-
l
umns of B is not equal to the number of rows <lf A.

[, -2:'l-r f
b) ATB lz 3 -r1_ lo 3l
[-: t l+ -5-j
rcr of
AB is Ir -:: -i[r
I o I3i
ATB
l-:
L:
i . - 1,0 + (-3)(a) \Z) + z(3) + r-:y1-S1
ArB = -l : - --rr0) + 1(a) -2(2) + 3(3) + t(-5)
|

_,: . - -1X0) + 2(4) 3(2) + (-tX3) + 2t-s) j

I

Therefore,
i-

,rTo
nD- -
ll
{r

to the

-
I

_T

Ir
-lT
2l Ir -z 3l
c) BTA
I

-10 :l \, 3 -1 |
,t
[-r
I

-s_l
1
")
L4
1 [t a n
Ir 0 4l\, 3 -11
BTA It 3 -5)I l^ 4
L" l, I

1(-2) + o(3) + 4(l) t(3) + 0(-l) + +(z)I Therel

[rrtl*o(2)+4(-3) 2(3) + 3(-1) + (-sx2]

BTA =lzirl+3(2)+(-sX-3) ;@.;i3; . (-sxl)

Therefore,

u,o=[l ?:l
lr -z3 -tli2
ll-' -z3 -l\i
d) e'=lz
l-r r z] lt-r | 2)
L

r?2) + (-2X3) + 3(1) 1(3)+(-2X-1)+3(21

Itttl *?-rx2) 3(-3)
2(-2) + 3(3) + (-lxl) 2(3)
+ 3(-l) + 2(-1)
n, = l2(l) + 3(2) + (-lX-3)
I

-3(3) + lt-l)
+ 2(2)
+2(-3'1 -3(-2) + t(3) +2(r) I

[-rtrl + l(2)

Therefore,
-t
l^, -s llI
A2 = I ll 4 -r1
\-z rt -tj
-t
"l ,

Fi[
i
e) BBr 3:I
-s1
 [, J-
,llr 041
BBT =lo
I
t4
.-1i)
3 -sJ
L -\lL
-l
f
J tttl +2(2) l(0) +2(3) 1(4) + 2(-s)
I
BBT =10(l)+3(2) 0(0) + 3(3) 0(a) + 3(-s)
14fl) + (_5X2) 4(0) + (-5)(3) 4(4) + (_sx_s) I

,',I Therefore,
i

]iaj
a-
t-) 6 -o-l
BBT =J6 9 -1s
I

l
:-
L-6 -15 +t!
T-
[,
) t z1
f) BrB =10 J o. :l_t I

lo -5 + -\r I

Ir I. ,
12
,,{
BrB =
i,
L'
1
03
r-J
-1)
zrzt
I

J',*1 1(2) + 0(3) + 4r-5)

I BrB =
'i , - 1_
-5t(4) 2e) + 3(3) + (-5X;5)
I
J

Therefore,

s) (AE
from (a)
 :s

Therefore,

- lrc -z
(AB)' =
t
_rl
rs
i_,n
using the rel ation (AB)r =
Another method to fihd (ABf is by
and then multiply the two
BrAr. In rhis ,"j"i*, nrd first Br and Ar
matrices.

-
l-', -5 ril I t lt
3l
h) A2-3A*9t=l tt 4 rl-:l z J -1 i
L*7 11 -4 [-r 1 )l
-l
-t a)
Ir 0 ol
*ei o 1 o1 b)

l0 0 I
- e 0
I

0l
T
trl [g el
nz-3Aapt=l
\-t2
l1
-5
4 tl -l u -6 s -3i* oe
oor_l
ol

[-r 11 -r] l-r 3 6l

Therefore,

-6tl
A2-3A +91-- 5o ol
2 8 -3 I

-
 SUPPLEMENTARY PROBLEMS
-

{B)r =
r _r [r
l)LetA=l | -2 3lunos=l-r t
rhe two
t_4 2 s_l l, ol
4
Find AB and BA.

,y,ro=j ' -' lunan=f I 3 5l

L_r r-l l: s r_]

a) find the product AB

i
b) show that A3 = 44

-3 --i
,2 4 _i B- -l _3 _51 i, : it _2 _;
.l 3)Let.r=f-r I .=1_;
!

rr-3i
ri

o
0l
ll -3 _-r : I _4_,

'l Find: a) AB b) B.d ;, BC d) CB e) CAB 0 CrAB

g) (AB)TC : C:C i) 3A + 48 _ 5C
jl C2_B__..:_-,i

.1, If A _ ZA2 _ 9,{ = 0 but

_.iron, rhat A3

A2_2A_9I*0

f
5) IfA=l-
1r
_: b
:--.iD=
[o o
j I 0
o-l
01. calculate rhe products
io l
L:
ol
AB and B-{ Jni :r:-, .i ::;f A2B +B2A=A.

-
$y
€l
d
I
 Hence, by solving the equations simultaneously,

c=1 d=-1
a=-1 ;
r-- 3
2
These are the el ements of the ma trix A-l and, therefore,

[-' 3l
n,r-l --l I
,l
Ir -1_,1

Check:

AA-l = I
-t r
I 3ri-l + ?I 0r-
I 1ll
L

3--1 ., I 0l
--1

1-2+3
,-2+) 7_ - = 0 I
f-
1.r_10
l_ ,_ 0 1,

DEFINIITION: An ei::-:::-.:-, ::'* ior column) operation on a matrix A is

any one of the ::,-- *-:tr: i,perations:
a) Interchar,-:: :: - ; i,r columns) i and j of A. This is called a type
I operat - -
b) ::; -::,_lumn) i of A by a non-zero constant..c,'. This
Multipl-,
: :.:. -; :leration.
is call=:
c) Add ": ::::._r :-; :r column) of A to row (or column) j of A,
wheie I = , l:"i ir ;alied a type III operation.
When a::,r:ni :, :::s:jered as the augmented matrix of a linear
system, the elernenl::-. :- a :::rations are equivalent, respectively, to inter-
changing two eqrairc:s, ::--:l:iving an equation by a non-zero constant, and
adding a muitrple oI .:= :r-laiion to another equation.
 the Gauss-
ch methods like
il*.on"riltl5J,"""i^" t#in*rr" me*'oa for solving I
elimination uno
the Gaussran elrllrura'vr'
Jordan method'
equatlons'
u tyo"* of linear
Soluti
INVERSE
METHOD IN FINDING THE
"^;;'t
PRACTICAL
ntlrnrx
Gauss'Jordan Method
- 1 [rle-']
\e\t\=
IJ
r
'
Procedures:
t^',.1^'.,-",1,
Moving from column t
= largest ab-
potnt Un"
\rrrE elemenr with the
vrv"'--
maximum pivot
1) Search for the
solute value)' anorhcr row (if neces-
-ai"gqr
--.':'num pivot with
, )'
:'i,"JT J :i l": f i:'
1:1": ",y" i::L[' ;',i;;

tu clements of the
p'-:::1"* by dividing all
3) Normali ze themaximun of the maximum Pivot'
' row bY the;;t"
4rF,liminatetheelementsofthepivotcolumnexceptforfuplYot,making
-'
t-t'" equal to zero'
"r"*ents
remembered by
csing the acro-
can easily be
Note that the n193ed1cs
nYm; "S - I - N -
E';

S for search
-
I - for interchmgc
N for normaliP
-
E for elimin*im
-
-
AUSS- PROBLEM.
h'ing -,
Find the inverse of e=]1 ^I
0
uring the Gauss-Jordan method.
_0 1l|
Solution:

LIT' -l
LAlrl = LrlA-' ]

r L] ,_l3llr
r,r-
L
lni rl=l' ol
Lo r_]

Normalize row ( l) by dividin g by Z:

rst ab-
[' -rlr
I 212 o

neces- lz 2l o 1_ R:=rz-ztr
of the i'I 311
212
,.,

naking
[o -r]-r
Normalize rou' ,l :'. *r"iding by -1:
f.
Ir
l- + - I R,=,-*,
L0 r --_

t .
Ir -
Hence. I
Iio LtI o-'l_l
fo.* [*I
which is already in the
'
L-
Therefore.
 PROBLE

Firx

o=
[' 2I using the
Gauss-Jordan
method'

Find the inverse "t

[l i iJ Solution

Solution: _1
_l

\e\
t-
t\ \r\r'\
L

z z\r
f, o 'lo\
rl -\,
-\ , r \o io tJ l'--tz-tr
R'=r3-rt
[et 'r\o
\1 5 p I r
L
N''=r1 -2rz
Ir , ,\, : :\
\;\n i1 -r\-:
o\ i il R,='I3-rz
L
\.. 2 R'=11 -4rr
s 4\_i -o, 'l;\ R,=ri+r,
Ir 1-1
\s -1 1i
;
L; '\o
Ir o l\-', 'o i\
\o 1 -r
L: ; ;io 1r
o

tbe
n )' 'therefore
frm ''l-lf
it is atreadY in -r
Since

t-7 zo -11l
rlo-'=i-l
[o -r
 FROBLEM:
-
[, r i]
rethod. Find the inverse of A = lr 2 3l using the Gauss-Jordan method.

[o t rl
Solution:

[^ t
'] = ['l o-']
[, r rlr o
lottl=lt 2
o-.1
:lo r ol R2=f2--11
lo I rlo ot_] Rl=rt

[r rt zl-r
rlr o ol
r ol
Rr=rr-2rz
lo I
Lo 'lr o o_l R3=f3-fz

lz -rr olol
['lo ot -rz l-l
|o o-r|'-r
']
Normalize row (3) by dividing by -l:
r o-rlz -r ol R,=rr*r.
o t z l-r I ol Rz=rz-2rt
_o o ti.l I _l_]
I o o1t o -11-
io r o jr-r zl
, o o r i-r -t_l
'
which is already in the form
[t ,O-'_]
 row (1) by dividing by 5:

r -l-lo
si
o -l-l
si
|.

2 3l0 r 0l Rz=rz e
r rlt o o] Rl=fr-ft
o alo
sl o -l-l
s
2 3lo
tt r o I

o alr
sI
o -l-l
sl
row (2) by dividing by 2: -1 ,

r 4lo
)t
o +l
)
Rt =ft *fz

r 3lo I o
I

1l I
I

LILI I

o +l
5:
r o -rl5j R3=r:

I _13 __l_ r I
r0 o 2 5
{F
ft

1 3 o I ol f
,q
]

221 B
-f
n method. 0:-l0l
rortr' r-r :1 ;i,.iding Ot
*,
1 1 -l-1
:CI o -
1 I Rr=rr+-]fit.rt
2 5

m point is o ilo+ 0 Rr=rr-at2'.,)

ltti
001,'0 .'i
{
A
+ 4l 'i
t
!
t 0 r-, 'frJ -- I ri
I
82 8l
0 I lr -jj
r< I "l
_l-)t
0 i
i
82- 8l
!
\ tl I
1 I
I
I 00r-0 4
_ _1
+l
L

- .#
- [' oI _l
-r o I 2 t , 0l R1 =rr*rj
i0 o I _,Olt I 0 0l R2=rr*rj
Jo o _r ll I I ol
L0 2l'0 0 0 Ro=14*13
'J
[' o o -rl3 z ,l
lo I o _riz 2
r
I ol
R1 =rr*14
R2=12*14
io 0 I _llt I I ol
[o o o rlr I I
R3=13+14
Ra=r+
'_J
[' o o ol4 3 2 rl
lo I o oi3 3 1l
lo oo or IolzI z
2
z rl I

Lo I I I 1J

Since it is al ready in the form

s-Jordan
[tt o-I , therefore
T'

lq i 2 rl
A*r _
lz : 2 rl
Jr r I rl)

Ir-f'
)T
t_

a,

tr+12

11+12
r,{

";:
 ;:==
SYI
SUPPLEMENTARY PROBLEMS
GA

of the given matrix us.ing Gauss-Jordan

In each case find the inverse tion
method and check the result'

[r z1
B=Lz
,, o=
[l l, t)
jl
Ans.A-'=
, rf-z
* L, _il
.J
B- = +rl
syst

t+ 0 'lo
syst
genl

o ol
[+o -2 on..o-,=\o -+
zra=1 oi the

[o o
L0 0 +l
l
3_l Ponr
has

[r ,3 il tl
[r -z
Ans.A-r=+l I o
I
-l
syst

r)e=lt 2l
"e3
\ I o tl
2
'_ whe

lll- I I r' -6
II ltt -3 -3
z 2 -1 4 I
^ 1t _l -l2 |
+lA=1t n"J.rr
Ans.A-r =!l
-gl_11 -2 1 2

i' 3 3 2
l
t-
L-tt
6 6

i' I z [t -10 9 -:-

rlo -5 o o6
'ioi*.o-,, =mlr
^ io -;
sin=qi -z o
| -zi lo
pr 3 z rJ [+ -s 4