1.

INTRODUCTION

we

student
student amiliar
is fam with matrices and their properties. In the coming chapters w
he
The that
Before doing
rank and using this we solve the systems of linear equations.
of the properties of matrix (without
define

ome
proofs).
5.2. MATRIX

ur
the form oy
Definition. A sysiem of mn entries from a field F arranged in
entries between [ ]
rdered set of m roWs, each row consisting of an ordered set ofn
lered set
r
or
).
is called a matrix of order or type mx n.

mn entries constituting the mxn matrix is called an element of the matrix.

Each of
Thus we write a matrix,

a12 an
a21 a22 a2n
A [4 lmxns 1Sism,1sjsn

am2 amn) m xn

call the entries as scalars.

In relation to a matrix, we

gives rise to many matrices.

Note. 1. A set of mn entries
numbers, then the matrix A is
2. of the matrix A are real (complex)
1f the elements
numbers, R (C).
said to be defined over the field of real (complex)
5.3. SOME TYPES OF MATRICES

and then A is called a square matrix.

Definitions 1. 1f A =[a;lmxn m=n,

is sometimes called as a n-rowed matrix A or

A square matrix A of order nxn

Sumply a square matrix of order n.

of n elements a (i,j=1,2,3,..,n).
IfA is a square matrix of order n, it consists
r e the elements a, a;2 ,....., Ai, Constitute the ith row and the elements a1 j, a2 js.... ., Anj

of the ith (first suffix) and the jth

Onstitute the jth column. Thus aj is the element
row

Oumn (second suffix) of the square matrix A.

97
 Mathemati3

Sc. ular matrix

of B.
r e c t a n g u l .

Book c a l l e d a

Text
A matrir
is
98 not a
square
squ m a t r i x .

which is row
matrix calleda
2. A is matrix.

ixm column

oforder called a then AA ic

is called
jor i*j,
marir

. A is =0
nx1
oforder ay
4 . A m a t r i x

=1 for i=jand
such
that a
Ivn
5. 1f A=la, a zero matri
denoted by 1. A 1s
called
or q
then
matrix. I is and Í,
unit 0Vi
a
=

such that
6. 1f A=la,
]mn
by O. diagonal elemen
matris. Ii is denoted called a
null
for i=j is
elements ay
7. Each ofthe vectors
in F .
Amn are egarded as
8. The
ows
of vectors of F"
are
regarded a s
9. The
columns of Amxn
5.4. EQUAL MATRICES
said to be equal if and onk .

=l4,] and B=lb;J

are
matrices A
Tiuo
every i and j
Definition.

order) and (ii) aj =bj for

type (or
) A and B are of the same
PRINCIPAL DIAGONAL
MATRIX.
DIAGONAL ELEMENTS
OF A SQUARE
5.5.
elements aj of Afor
which i=j ie
Definition. Ifa matrir A=la;]nxn, the
elements ofA. 1The line along which the diagonal
al
af1,22,,n )are calledthe diagonal
elements lie is called he Principal diagonal of A.
5.6. THE TRANSPOSE OF A MATRIX.
Definition. The matrix obtained from any given matrix A, by interchanging is

rows and columns is called the transpose of A. It is denoted by A' or A'.

f A=loglmen then the tramspose of A is A=®,lnxm where by =aji.
Also (A') =A.
Note. A, B are matrices. Then A = B> A' =
B.
5.7. THE cONJUGATE OF A MATRIX.
Definition. The matrix obtained
by the from any given matrix A, on replacing its
corresponding conjugale complex mumbers is called the elemenio
denoted by A. conjugate of A anu
Thus if A =[ay lmxns then A
of dij.
=[b;]nxm where b =
ajj , the conjugate complex 1mben
nun
Note. 1.
If the matrix A is defined over
R, then A A.
2 3i 2-5i
=
2. ( ==A-
A
e.g.
A-i 0 2+5i 1
4i+3]2x3 i0 -4i+3
 99
5.8. THE CUI
I'RANSPoSE OF A
nition. The
efinition. The
conjugate of the MATRIX
0se
ranspose of Aand is denoted by A transpose of the matrix A is called the
conjugate
Thus A= = (
A) where A is the
transpose of A.
Now A=layl mxnA A" =lbjl
n*m where
b= aj
,-5 3-i -2
e.g.A=
e.g A =3+i 1-i
0 1+i
4-il23 22i 4+i3»2
4+il3x2

Note. () A" =(A')=(A' ) (a®° =A

(ii) (KA)" =
KA" where K is a
complex number
(iv) (AtB)" = A* +B
(AB)° =B°A
5.9. DIAGONAL MATRIX

Definition. A square matrix A is said to

be a
diagonal matrix
diagonal elements is zero. if each of its non

Thus if A
=[ajlnxn is a diagonal matrix, then 4 =
0 for i#j.
If d,d,..,d are the diagonal elements of a n - rowed diagonal matrix, then we
write the diagonal matrix A as: A =
diag (d1.dz,.d)
A diagonal matrix whose diagonal elements are all equal is called a scalar matrix.
5.10. TRIANGULAR MATRICES

Definition. A matrix A=la; mxn is said to be an upper triangular matrix if

lower triangular matrix if aj =0 for every i<j.
" for every i> j and a

2-3 0
0 4 2 1 matrix.
is an upper triangular
eg
0 0 0
0 -6

0 0

7 0 0 0
5 3 0 0 0 matrix.
0 is a lower triangular
and-4 6 0 0
1 -8 5 0

2 0 4 1 6
 A Text B0ok of B.SC. Vidu
100

5.11. SYMMETRIC MATRIX

A =[4j] is said to be symmetrIC y aj =

ai fo
Definition. A squae matrir ever
i and j.
A =A
Thus: A is a symmetric matrix
SKEW SYMMETRIC MATRIX

Definition. A square matrir A =[4;] is said to be skew symmetric if aijaji

a=~a.
fo
every i and j.
Thus A is a skew symmetric matrix A =-A
Note. Every diagonal element of a skew symmetric matrix 1S necessarily zero sin.
since

ah
hb is a symmetric matrix
e.g.

a -b
and -a 0 is a skew symmetric matrix.
b -C

Note. O,nxn and I, are symmetric matrices.

5.12. COMPLEX MATRICES
HERMITIAN MATRIX.

A square matrix A such that A' A is called Hermitian matrix. This

=
a can also be
written as (A")=A .

eg. If A= 1+3 then A=4 1-3i

then A +3
7 7

and AT= 413 7

l-31
and A =,.
1+3 7 then A' = and Ais Hermitian.
Note. 1. The elements of the
leading diagonal of a Hermitian matrix must be real.
2. A Hermitian matrix over the field of real
numbers is nothing but a real
matrix. symmetri
3.
Anecessary and sufficient condition for a matrix A to be
SKEW HERMITIAN MATRIX Hermitian is that A° =A:

A
square matrix A such that
A'= -A is called a skew hermitian matrix. This c
-

also be written as (A')=

-A.
 -3i 2+i
Let A-2+i hen «3
-2-i 2-1
-3-2+
and A
and -A--3i
2+i -i
2+i
2+i which
1.The
1. elementsof the leading
The elements shows A =-
skew
2. A ermitian matrix over diagonal
Skew - Her must be all
the field of zero or all are
are purely
purely imaginay
imaginary.
real numbers is othing
nothing but aa real skew
symmeuicmatrix.

real
ssary and
3. A necessary sufficient condition for a
matrix A to beSkew
Skew Hermitian
Hermitian is1 that
--

A=-A

TInitary Matrix : A
Square matrix A such that
(AT =AT
Tha is ( =A (A)' =I or A"A =I is called a
unitary matrix.

is a unitary matrix.
e.8

Note. 1. Every real symmetric matrix is Hermitian. Similarly a real Skew - Symmetric
matrix is Skew - Hermitian and a real orthogonal matrix is unitary.

matrices generalize symmetric,

2. Thus Hermitian, Skew Hermitian and unitary
-

matrices respectively.
Skew Symmetric and orthogonal
then iA is a Skew Hermitian matrix.
3. If Ais a Hermitian matrix
-

as the sum ofa

Hermitian matrix and a
matrix is uniquely expressible
4. Every
square
Skew Hermitian matrix.
-

A+iB where A is real and symmetric

can be written
as
5. Every Hermitian matrix
and B is real and Skew - Symmetric.
matrix.
then iA is a Hermitian
Hermitian matrix
6. IfA is a Skew -
MATRIX
5.13. SUBMATRICES OF A columns or both are deleted
some
rows or

Definition. Let A be a
matrix. Ifsome
submatrix of A.
matrix is calleda
Om A, then the resulting
matrix A.
submatrices of a
n u s we can have many DELTA.
CALLED KRONECKER
5.14 0 , i j IS
1i-j:8y
matrix,
then ay =y
a unit
eg. If I, nxn is
FIELD the field
oVER A matrix is
over

S15. THE SSET OF

THE SET MATRICES
field F, we say
that the

belong to a field F.
matrix over
lf all ethe elements
eleme ofa of all
matrices

. From nOW the set

FrO Y DN We
çonsider
 102 MATRIX
BYA SCA
m u
m ullt
tiippthyiinng
g
e
e v
v eerr
y
elem
by K
MULTIPLICATION OFA ob
o bt
taai
inne
edd
by
by by
kA or
AK.
demoted

5.16. ma
m at
tr i r
.."k,
matrix
The
K
and is
A
Klalmx
|a,1m KA
Let A be
a
product
of A by and IKa,
Imxn
the
=[Ka,Im*n
called
is
a
Calar, KA
then
]mvn.
Thus: IfA=l4,
negative ofA.
Properties. calied the
= -A.
(-1)4 scalars.

O4 = 0
(null matrix). are
) where K,.K2
(K^A)
=
(K,K, )A =K2
(1) R(K2A)
KA is symmetric.
symmetric
(ii) A is symmetric.
KA is skew
skew symmetric
(iv) Ais 0if K 0.
=0 A
=

(vi) KA
(v) (KAY= KA'
null matrix.K
=
K2
(vii) K,A =KaA and Ais not a

ADDITIOON
MATRICES OR MATRIX
5.17. SUM OF

be two matrices
Let A =la, Imxn.B ={b,),m»n
4
where =a, +b,. is called the sum of the m a t r c e s and

Thematrir C=lc,),mxn
B
of Aand B is denoted by
A+
B. The sum

and l,,lm..l4, mes "l, lm..

Thusla,Imn *1b, lm. *l4, ,lm..
matrces and be the natrix adkdaton
Properties. Let M be the set of all mn

(composition) in M. Then

() Closure. A,Be MA Be M
(i) Commutativity. A,Be M A B-B A

(iüi) Associativity. A. B.Ce M (A B)C = A (B C)

Thus (A+ B)+C =A (B C) A B.C
(iv) ldentity. o (null matrix). A M0- A AO A
Thus the null matrix is the additive
dentity in M
(v) Inverse. A ¬M A+(-A)
(-A} A =O, =

-A is called the additive

inverse of Ain M.
(vi) A,BE M and Ke F~K
(A+B) KA KB =

(vii) (A+B=
A'+B for A,B e M.
Note If A,
B, Ce M
.
A+B=0B= - A or A =
3. A -B 2 A B A+C
B-CA =C-B or =

B C
B= C- A
5 , 1 8 . 1 1

1fA. B
are hvo matrices of the
HCTION)OF TWO MATRICES. 103
same
=l4j lmx n.B =lb,l type
Thus
(order), then
mxn then A+(-B) is taken
taken A-B

A-B=l«,}-1b,]=la-b,}and
as
« .

tay-bylayl-lb,]
b, l and
rties. (A -B) =
A'-B,
Noe
Note If A 3 are
and
matrices ofK(A-B):=
the same
KA -KB where
KEF
able for addition or type
r subtraction. (order), then we say that A and B
heorem. Every square matrix
5.19. Theorem. Eve are

nda
skew
symmetric
- symmetric m cun be
matrix in one
and only
expressed as the sum ofa
of symmetrie
5.20. MATRIX MULTIPLICAT
one
way (uniquely). symmeric a

Let A =[4l mxn=jlnep

A =ljl mon B =

.The matrix
2a% C =[¢;l where
a bbj isis called
called
aroduct of the matrices A and B in that order.
t h ep r o d u c t
nep
Cij2
k=1
.

fo the
product AB, the mairixXA IS called the
1Cthe number of columns ofA is equal to the prefactor and B the post factor.
number of rows in B, then the matrices
are said to be conformable for multiplication in that order.
Note.1. IfA, B are matrices, then

AB may be possible and

BA may not be
possible.
(i) BA may be possible and AB may not be possible.
(ii) AB, BA may both be possible.
(iv) Even if AB and BA are possible, they may not be equal.
(v)AB and BA are possible when A, B are each of order nxn. In this case also they
may not be equal.
not commutative. If A and B are matrices such that
(vi) Multiplication of matrices is
AB= BA then we say that A and B commute.

.IfA, B are conformable for multiplication and k,I are scalars,

(k/) (AB)
(kB) and (kA) (B)
=

tnen (kA) B = k (AB) = A

multiplication is
associative i.e. if A, B, C are matrices,
1. Theorem. Matrix
then (AB) C A
(BC). =

The Associative law for the product

of any
Law. Hence the
Cor
. Generalised
Associative
using
mathematical
induction.

Tumber aof tm be proved to be true

particular
manner
of bracketing,
rices can
independent of any
produetf
While
any y n u m b e r of
matrices is
is not changed.
der ofthe matrix factors MATRICES

52
.22. POWERS
OF
SQUARE

SITIVE INTEGRAL AA.

defined as
is
Let A be matrix.
Then A
a square
 =AA' A
associative
law, =

Now. by the we
write A^A
that
A (AA) =AA
so

AA
=(AA)A
= is a positive
integer
where m
A"
A "-A
=

AA=
have
Similarly'we
(A")"=A
where m,n are positivei ntegerg
A"A" = A+"and
Further we have

Note: 1" =I,0" =0 is distributive w.rt. addi

W.r.t. additi
matrices
Theorem. Mulriplication of
5.23.
AB+AC and (B+C) A =BA+CA.
matrices i.e. A(B+C)=
Al, =1mA A.
=
then
5. 24. Theorem. 1fA is a matrir of order mxn,

AB =
0 does not necess
5.25. A, B a r e matrices such
that AB =0. Then essare
must be O.
mean that atleast
one the natrices A and B
of

0 0
e.g. Let A=
egLet and B=
A0 022 1J2x2
Now AB= = O even though A # O and B #0
0J 2x2
5.26. TRACE OFA SQUARE MATRIX

Definition. Let A =lay) nxn. The trace of the square matrix A is defineda
a and is denoted by 'tr A'. Thus tr A
=2Gi =q1ta22 t...+an
i=l
i=l
Properties. If A and B are square matrices of order n and a is
() tr (7.A) = 2. ir A
any scalar, the
(i) tr (A +B) = r A +tr B (ii) tr (AB) = tr (BA)
5.27. Theorem. IfA is mxn matrix and B is
Cor.
a nxp matrix, then (AB) =BA
(ABC...Z) =
Z'Y"...C'B'A'
5.28. 1. IfA is a
square matrix such that
A =
A, then A is called
2. IfA is a
square matrix such that A" 'idempotem
=
0, where m is
A is
called'nilpotent'. lfm is least positive intege, tne
a

'nilpotent' of index, m. positive integer such that A" 0, then A is

=

3.
cau
1fA isa square matrix
such that A^
4. 1f AB=A
and BA B =I, then Ais called
5.29. ADJOINT then A and B are
=

involutory.
OF A
SQUARE MATRIX idempotent.
Let A
=la;lnxn be a matrix. The
A, is called the matrix
adjoint
transpose of the matrix of Aand is denoted by the
[b;)nxn
where bij is the or of ak
cofactor
formed by the symbol
cofactor of A. adj A. Thus adjoint of oAIsthe
 11 A21*. Anl
105
adj A =| A2 A22 A2
.e.
********** where Aj is
An A2n An the cofactor of
aij
2 -1 4

e.g.
If A =|4 2
-3, then adj A=|-11
7 6
-5
1 0 22
L2 -3 8
I,
=In,adjO=0. adj (KA)
K"=adj A if Ais of order and K isis
=

5.30. Theorem. If A=l4jlnxn then A(adj A)= and K ascalar.

scalar. n a

5.31. INVERSE OF A MATRIX (adj A) A=|A|l,n

nefinition. Let A be any
matrix, then a
of A.
inverse matrix B, if exists such
called that AB= BA I is =

Eor AB. BA to be both

defined and equal, it is
aatrices of same order. Tnus a necessary that A and B are both square
non-square matrix cannot have inverse.
Definition. A matrix is said to beinvertible, if it possesses inverse.
5.32. Theorem. Every invertible matrix
possesses a unique inverse.
Note. 1. The inverse of Ais denoted by A.

Thus AA =A-'A =I Also A A =AA =I-(Al=A

»The inverse of a matrix is invertible and the inverse of the inverse of the matrix is
itself.
2. Since I, I, = I,, we have = I i.e. the inverse of a unit matrix is itself.

3. IfA is an invertible matrix and if A = B, then A" =B

sufficient condition for a square matrix to

5.33. Theorem. The necessary and
POssess inverse is that |A|#0.

Note. If |A|#0, then A' =(adA)

MATRICES
SINGULARAAND NON-SINGULAR f |4|#0 then A is
be singular if | A|=0.
matrix is said to
einition. A sauare

aid to be
non-singular possess
inverses.
Thus thee
Thus only
only non-singular
s
matrices

then AB,
the product
is also
non-singular,

No A , B are
non-singular,
non-singular.

product of non-singular matrices is

also
 Book of B.Sc. Miau. B=C
A Text then
= AC,
106
non-singular
and AB n
such AB I, the
that 41R.
Theorem. 1fA is m a t r i c e s of
order
S.34.
square
and B
are

5.35. Result. IfA

A =0.
BA =I. A)=(adj A)
then A (adj
S.36. Result. IfA is
singular,
non-zero
square
matrices
is a
matrix, both
zero matrir

of
S.37. Result. Iftheproduct
matrices.
them must be singular
=0
then |adj A|
5.38. Result. If|A|=0, also symmetric.
then adj A is
symmetric matrix,
5.39. Result. IfA is a

A' =
(adj A).
be square matrix, then adj
5.40. Result. IfA a

non-singular matrix,
then det (A ") det A
5.41. Result. IfA is a

5.42. Result. 1fA is a square matrix of order n, adj A|=|A|".

Note. If A is non-singular, then adj A is also non-singular.
5.43. Theorem. 1fA and B are two non-singular matrices of the same type, then

(adj AB) = (adj B) (adj A).

5.44. Theorem. If A,B are invertible matrices of the same order, then

0 (AB)=B'a- () (a'1=(a-l}
5.45. Result. If A is a square matrix of order and B is
n a
non-singular matri
oforder then there exists one and only one matrix X such that A= BX and thern
n,
exists one and only one matrix
Ysuch that A= YB.
5.46.RIGHT AND LEFT DIVISORS
Definition. f A,B are square matrices
A#0 then A is called a left zero of the same order such that AB = 0 and
zero divisor
divisor. If AB =0 and
B+0 then B is called a rigt
5.47. ORTHOGONAL MATRIX

Definition. A square
matrix Ais said to be
A is an
orthogonal matrix A'A =
I
orthogonal if A'A I (0. U. M0)

A'A|=|l|=|A'| A|=1|A|=1
A=t1<|A|#0=A
A'A
is invertible. Also AA
=AA =1>A' 1, if A'A =I.
Thus:A A =

is an
orthogonal matrix A'A
=1= AA'
 10g0ut atrces,
a r e orthoy

ogonal 107
The inverse
5.49. Result. The
inverse of an
matrices, each ofof, order
-ARY orthogonal matrix is then AB and n

TRANSFORMATI O NS
5 4 9 .E L E M E N T A R

OR orthogonal.
R. Interchange of i" and jh
rows
OPERATIONS ON AMATRIX
(i) R, Multiplying every element oftermed
(k). M
h
as
Type I
row by k, termed as
(ii) Ry k). ultiplying every element Type I
sponding element of h of j" row
row, termed
as
by k and
an
adding to the
Type Ill
Similarly we understand
Ci. C, (k), C(k)
lementary
transformations on a
for columns.
matrix do not
50. ELEMENTARY MATRIX, change
ange the order of the
matrix.
nefinition. A matrix obtainea Jrom a unit matrix
o f the elementary I, by subjecting the unit matrix
matrix
tary mati
elementarry is said to be
transformations is called
elementary matrix. The
an
of type 1, 2 or 3
according as whether the elementary
perations performed
erations performed on unit matrix
I,, is a
type 1, 2 or 3 operation,
5.51. SYMBOLS respectively.
1. Ei. Elementary matrix obtained by
interchanging h and h row or h and ji"
columns inI.
2. E (k). Elementary matrix obtained by multiplying every element of th row or h
column with k in I.
3. E (k). Elementary matrix obtained by multiplying every element of " row with k
and then adding them to the corresponding elements of h row in I.

4. E k). Elementary matrix obtained by multiplying every element of h column with

k and then adding them to the corresponding elements of th column in I.

S.52 transformation of a matrix

Theorem. Every elementary row (column) with the corresponding
can be
th obtained pre-multiplication (post multiplication)
by
elementary matrix. denoted as
5.53. Definition.
Definition. Let A belong to M,mxn), we define the rank of A,

rank (A) or linear transformation LA:F

-F".
OP(A)to be the rank of the
linear transformation
between
finite
5.54. Theorem. Let T:V w be a
W respectively
eorem. bases for Vand
dimensional ector spaces, and
let Band y be
ordered

rank (T) = rank ([T!

 A T e x t BOOK
fPana
g
matrix.
108 mxn

an
LetA
be
Theorem.
5.55. then
respectively,
matrices,
nxn

=rank (A)
rank (AQ)
1. therefore
and
= rank (A),
2. rank (PA)
rank (4) maximum
number
of its
of its linea
3. rank (PAQ)= the
ce sgenerag
The rank ofa
matrir equals
the
dimnesion
of subspace
Theorem. matrir is
columns (or)
The rank ofa
ndependent
by its columns.
FOAM
REDUCTION TO NORMAL
5.56.

to the form
[1, by aa fini
by
be reduced o
matrir can
Theorem. Every n o n - z e r o
marx Of order r and..
transformations,
where I, is the unit dr
mumber
of elementary
the rank of the matrix
exist non-singular matri
matrix of rank r, there rica
5.57. Theorem. ifA be a mxn

P and Q such that PAQR=

5.58. Theorem. The rank ofa matrir does not change by pre-multiplication
post-multiplication with a non-singular matríx.
5.59. Theorem. Let T:V-Wand U: W>Z be linear transformations offint
dimensional vector spaces K,Wand Z, and let A and B be matrices such thatth
product AB is defined. Then
i) rank (UT) s rank (U) (0. U. M 08) (ii) rank (AB) < rank (A)
(ii) rank (AB) rank (B) (iv) rank (UT)< rank (T)
5.60. ECHELON FORM OF A
MATRIX.
If all the elements ina row ofa matrix are
if there is atleast one
zeros, then it is called zero row a
a
non-zero eleement in
Row
a
row, then it is
calle.! a
Echelon form or non-zero row.
A matrix is said to
simply Echelon form ofa matrix.
be in Echelon
() Zero rows, if any, must form if it has the
follow non-zero following properties.
rows.
(ii) The first non-zero
element in each
(ii) The non-zero row must be
number of zeros before the
number of such zeros first non-zero
1.
in the next
row.
element in a row is less tna"nthe
Some authors
The rank
ignore the property (ii) to consider a
the matrix. of a matrix in matrix to be in
Echelon form is
equal to the number Echelon forn
of non-zero ro
wS
 OF A MATRIX
INVERSE
narked
5 6 1 .I N V E R S E

109
Weh a v e .r e m a r
that an nxn
compute the rank of matrix is
te the
any matrix invertible
qoWh o wt o

itis
invertible or not. we can
ie if
and
and only
only ifif its rank is n.n. As we
whether

Def. Augmented Matrix

always test matrix
a
atrix to
to determine
62 dete
mxn,
n,
LetA andB
are n
mxp matrices
or (A
B) we mean mx(n+ p) matrixrespectively. By the augmented
mented matrix
mairix [A B]or
lastp
coum are the
columns of B. whose first columns
n
ans are the
columns offA,
define the rank or a matrix
w h o s

We n o r

OF AMATRIX
using the minors of that
1.
MINOR matrix.
LetAbe a matrix of the type mxn defined
matrix
over the field F.
SUch
minimum
m
that 1 S
of m and n.
If we delete fromA,
Let tbe a
positive integer
n-1) columns
columns leavi a certain t (m-t) rows, leaving certain f rows
leaving columns and if t>1,
rowed square matrix whose determinant is a then the remaining
t- ining elements
elements of
ATre Corresponds a system of minors of A to
Thusthere
minor of the matrix A of order t.
each possible value of t. Each single
element of Amay be regarde as a minor of order 1 (t
=1)
OF A MATRIX
2. RANK
IfA is a null matrix, then the rank of A is defined as zero.
Definiton. fA is not a null matrix; then the rank
of the matrix A is the positive
integer.if it possesses the following two properties
() Each minor of order n+1 of the matrix A vanishes (i.e. is zero)
(i) There exists at least one minor of order n of the matrix A which is not zero.
Note.
1. Rank of A is denoted by p(A) and is unique.
2. Every Matrix will have a rank.
. IfA is a matrix of order mxn, then p(A) Sm orn (smaller ofthe two)

minor of order n+1,n+2 etc., is zero.

p(A) =n, then every
p(A)=n
AIS a matrix of order nxn. A is non-singular
6. Rank of
I, =n k <m) is zero, then
order minor (k < m,

4matrix of order mxn. Ifevery kth

P(A) <k. orderk (k<m, k <n) which is
not

of
8. Ais
drix
of order mxn. Ifthere is a minor

zero, then p(A)>k

 exceed the
rank
e r c e e d

c a n n o t

110 M A T R I C E S .

m a t r i c e s

OF two

OFA
P R O D U C T

product
of
3.
RANK

Theorem.
The
rank ofa
m u l t i p l i c a t i o n , then p(
AB)<
either matrix conformable
for
matrices

fA, B
are

OR:

p(AB) Sp (B). The n m

and

mumbers
is
called
an
n - v e c t o r

may be writen in,

number
4. VECTOR
tuple of mp
com po
onne
ennt
tss
may
ordered n- The o."
An row or a
taken as a
Definition.
components.

called its be
are vector may
n-vector
Thus a
of a n in a
vertical line.
line or
horizontal
the v e c t o r is to hbe
said to ove
matrix. to a field F, then
ofa n-vector belong
If the components

the field F.
and K eF. then
f = (a),a2,..G,) multiplication ofa vector.
called scalar
KG=(kaj, ka2, ka,)
- -

. .

then i 52 ( =
+h.4 +b +...4, +h
and 2 =(h. b....b,).
If
S =(,a2,...,)
called vector addition.
K F>1 +5, V,
is aset of r-vectors overa field Fand . .VeV
e a
If V,
KEEV
The set of all n-vectors of a field of F is a vector space over the tield F for the
operations of vectoraddition and scalar multiplication. It is denoted by V,,(F) or simply V,
A vector X, expressed in the form X =k, ky ty ....k, t, Where k, k.. F and
x, X2,x, ¬ V, is called a
linear combination of vectors . 2, ,

5. SUBSPACE OF ANn-VECTOR SPACE V, .

Definition. A
non-empty
subset. S of V,tF) is called a
subspace of V, (F) #1
follows vector addition and scalar
Every subspace of V, (F) contains the
multiplication
6. SUBSPACE SPANNED BY A SET OF
zero
vector.
Definition. A
VECTORS
given set f vectors,
subspace which arises as a
is said to be set of all linear
7. BASIS OF A
SUBSPACE spanned by the combinations
given set of vectors. of
Definition. A set
of vectors
(i) the is said
8. subspace is
Theorem. A basisspanned by the
to be basis of a
a

which span S. of a subspace S,

set and
(i) the
subspace if
can set is L.I.
Note. lt may always be selected from
also be a set oj vecoe
every basis of S, isshown that
vectors in
the same. V,(F) possesses a
basis and that nber
the u
g . D i M E N S I U

The numbe.
er of vectors
Definition.

ion of the subspace in any

basis of
d i n e

COLUMN SPACE
1 0 . R O WS P A C E ,
SPACE AND
a
subspace
ace is called the

LetA be matrix of order m

xnover
OFAMATRIX
a field
F. Every
of consists of
A column
row of A
may be taken as a elements. So cach row
m
devenlumn
ande
every

lumn may
ande a c h c o l
consists of elements n
be w
m-vector. may taken as an n-vector
V,F) and every
column of
A is an Clearly every row n-vector
of A is an element
an

element of V.(F). Or

space spanned by ihe m rows

which is
of the matrix A.
a
subspace of V,(F) is called the row
space
a c e

The space spanned by the n

columns which is
column space of the mxn matrix A.
a
subspace of VF)
V, is called the

dimension or the row space of A is called the

The
row rank of A and the
column space of A is called column rank of A. dimension of
1. EQUALITY OF ROW RANK, COLUMN RANK AND RANK.
Theorem. The row rank of a matrix is equal to its rank.
12. Theorem. The column rank of a matrix is equal to its rank.

RANK OF A SUM.
13. Theorem. Rank of the sum of two matrices cannot exceed the sum of their
ranks. (OR) FA,B are two matrices, conformable for addition, then

rank of (A+B)S rank of A + rank of B.

14. Theorem. IfA, B are two n-rowed square matrices,
then p(AB)2p(A)+p(B)-n. (S. V.U. 93 S)

15. Result. If A be any non-singular matrix and B a matrix such that AB exists, then
show that AB and B have the same rank.

Sol. Since A is non-singular, A exists.

C AB Alc =A (AB)= (A'A) B

I B = A-'c B-Ac.
. (2)
P(C)=p (AB) S p(B)
.
(1) and p (B)
=
p(AC)sp(C)
From (1) and (2), p (B) =
p (C) =p (AB). type n and I is a unit matrix of
matrices of the
same
16
K e s u l t . If A,B are square
p(B-)
, then show that p (AB-) Sp (A-) +
AB-Al+A-IA(B-1)+(A-)

Sol.
Now (AB-)= AB-A+A -I=
p(AB-1)Sp[A(B--1)]+ p(A -
PAB-I) =p[A(B-1)+(A -)]
p{B--1)} p(B-).
=

ut p[A(B-1)] s min. of {p(A), + p(B-1).

p{AB-) sp(A -1)
PAB-D <p(B-)+ p(A - 1)
 A Text B EQUATIONJ

LINEAR u n k n o w n S.
in n
112 NON-HOMOGENEOUS equations
linear

SYSTEM
OF non-homogeneous
17.
C o n s i d e r a system ofm

=2
+a222 t
t
a2
a21-3
.

Dm
+am2.2
t..
+amnn equation.
am1 as the matrix
can be expressed
The system
i.e. AX=B Ihen the linear
A nx1= B mxl n,
mxn

non-singular
matrix oforder
18. Theorem. IfA is a
solution.
unknowns has a unique
AX =
B inn
DETERMINANTS
PROPERTIES OF
19. =det (A)det (B).
then det (AB)
Theorem. For any A,B eMn(F),
only if det (A) =0.
) is invertibleifand
Theorem. A matrix A eMpn

1
Further if A is invertible then det (A)=detA

det (A) =
det (A)
Theorem. For any A EMps(F),
LINEAR EQUATIONS
EQUATIONS
5.63. SYSTEM OF NON-HOMOGENEOUS LINEAR
Consider the system of m equations in n unknowns given by
a11 t12X2 t... tajnd, =bj

**°°°°°°°°°°°°°°°°°°°°°°°°.°°°*°e

amtam2*2 T. +amn'nbn
(i.e.) AX = B

The augmented matrix a21 422 23.2

of the above system is [A|B] =| *******°****.

Findthe ranks of A, [A|B].

m m2 m3*.mn |Dm
Working Rule : The system AX =
B is
infinite) ifand only if consistent i.e. it has a solution (uni1qu
rank of A rank of =

[A/B]
() Ifrank of A rank of =

[A/B]
there exists infinite r<number of unknowns, the =

number of solutions. Giving

unknowns, we may express system is consise f but

the other r arbitrary values the

unknowns in terms of theseto n-
 TErank of A rank of =

[A/BJr =n the system has

of A* rank of unique solution.
(ii) If rank [A/B] the system is
solutions, set (n-r) variables
htain solu
To obtain inconsistent. It has no solution.
any arbitrary value and
and solve for the remaining
unknowns.

t
htus
Thu irrespective of m < nor m =
n, or m
n, we have decided >

system of linear simultaneous equationsabout the existenC

uniqueness
aueness of solu of
solutions of a
DrOcedure is illustrated equations AX=
AX B. =
s
the through following examples.
SOLVED PROBLEMS
Ex. 1. Show that the equations +y
are not consistent
x +z =4, 2x + 5y - 2z =
3, x +
7y - 7z=5
Sol. We write the given equations in the form AX =B, where
|1 11 X
4
A= 2 5-2X=| B =|3
I 7 -7 5

I 1 14
Consider [AIB] | 2 5 - 2 3
=

I 7 -7 5
Applying R > Rz -2R and R R - R, we get
I I 4
LAIB]0 3-4 s
0 6-8 1
i 4
Applying R,>R, -2R,, we get LA/B]00
3 4 -5s
0 0 11|

We can see that rank [A|B] =3, since the number of non-zero rows is 3.
on A, we get from above
Applying the same row operations

A =0 3 4
0 0 00 2.
the rank of A
=

rows is 2 so
Here the number of n o n - z e r o

Here we have rank (A) # rank [A/B].

consistent,
The given system is
not

u the
simultaneous equations x +y +z =6, x

EX. 2. Discuss for what values of , unique solution

H have (i) no solution (ii) a
+
3z= 10, x + 2y + z
=

4y+
(iii) an infinite number of solutions. is
o . T h e matrix form of given
system of equations
 114

10-B
AN=I 3
2

3 10
1 2
[AIB] =|
the
augmented
matrix isis [a
matrix 2 J i 116
We have
22
AIB]~|0
get
-R,and R,
>
R, -R,, we
we get 01 -1 l -6
Applying R R ,
6
2 4
1
[A'B]
-

10
Applying R, ->
R, -R,, we get 0 0 -3 -10

[A/B]= 3,
so thatthey have same s

rank. Case
Then I.theLet
system of equations
then rank is
of A =3 and rank
consistent. Hereofthe number of unknowns is 3 whefor
2 3 ich
solution. This is tnie t
unique
equations will have
a
The system of
Is same as the rank of A.
any value of u. will have a unia
Thus if , #3 and u has any value,
the given system ofequations
solution.
that rank of A =
2 and rank a
Case II. Suppose 2 =3 andu # 10, then we can see

[A/B] 3. Since the ranks of Aand

=
[A/B] are not equal, we say that the system of equations
has no solution (inconsistent).
Case Il1. Let = 3 and = 10. Then we have rank of A = rank of [A/B] = 2.

The given system of equations will be consistent.

But here the number of unknowns =

3 > rank of A.
Hence the system has infinitely many solutions.
Ex. 3. Find whether the
following equations are consistent, if so solve them.
x +y + 2z 4;
2x-y 3z = 9;
+
3x-y-z =?
Sol. The given equations can be written in
the matrix form as |2 -1 3
2
i.e. AX = B 3 1 -
The
augmented matrix [A|B]=2|2 -1 3 49
3-1 -1 2
2 4
[A B-0-3 -1 1
0 4-7 -10 (Applying R2 >R2-2R and R, R,- 3R,)
 3 -1
A
A/B/-/o
0 0 -17 -34 (Applying R,> 3R,-4R.)
KanK
of A
ofA
=
3 and Rank of
Since
T h eg i v e n s y s t e m is
consistent. LA|B]=3. Rank of A
So it has =Rank of [A|B)
Rank =
ofA of Rank
[A/B] = solution.
a
Since
number of
given system has a unique
solution. unknowns,
21
0-3-1
0 0
-34
r+y++2 = 4 (1) -3y-z =l
(2) -17z =-34
in or z
=2
Substituting
ubstituting
z=2
y=-1,z =2
(2)-3y-2=1>-3y
in (1),
=3= y=-1
(3)

Subs we get
r-1+4=4 x=1
=1,y=-1,z =2 is the solution.
T 4. Find the value of n or which the
system of
2, 6x + y * Az - 3 will have infinite equations 3x 4z
+ >
3z=-
=
y
-

that n value. number of solutions and solv

hemwith
Sol. The given system of equations can be written in the matrix form as AX = B

3 -I
41:] [3]
(i.e.) 12
3 =-2
65 -3
3-1 4 3
2-3-2
TheAugmented matrix [4,B] =| 1
6 5 -3
3 -1 4 3
-13-9
Pplying R2 >3R2 -Ri,R > Rs -2R [4,B] -0 7

0 7 A-8-9

4 3
3-
-9
Applying RgR3 -R2, [A,B] ~ 0 700 +5
-13

a+5 0
0 0
If = 5, Rank of A =2 and Rank of [4,B] = 2

Number of unknowns = 3
unknowns
Rank of A number of
=
Rank of [A, B1 +
 and

116 is
c o n s i s t e n t

system
-5,
the given
Hence
when 1 =

solutions.
4
3-1
o 7 -13

If a =-5 the given system

becomes
oll
=-9
(2)
and 7y-13z
3x-y+4z =3 (1)

Let =k* Then

From (2), z

7y =13k -9 y
=
(13k -9)/7
7y-13k =-9
get
Substituting the value ofy (1),
in we

3-(13k-9)+4k =3 3x- -4k+3

3x=

*=(-Sk+4),y = (13k -9),z =k is thesolution.

Ex. 5. Findthevalues ofp and q sothat the equations 2x+3y + 5z =9, 7x+3y+2z =8
2x+3y+ pz = q have

() No solution (ii) Unique solution

(iii) An infinite number of solutions.
Sol. Given equations are
2x+3y+5z =9,7x+3y+2z 8, 2x+3y+ pz =
=
q
The equations can be written in the
matrix form AX= B
[2 3 5
i.e., |7 3 2||
3 pllJ
2 3
9
0 -15 -31
=47 (Applying 2R2-7R|.R3
lo0p-s||=] l4-9
Applying same operations,
-R)
2 3 5 9
[A, B]-0 -15 -31 47
0 0
Pp-5 q-9

We have 2 35
det A=7 32
2 3
 117
2
(3p-6)-3 (7p-4)+5 (21-6)
6p-12-21p +12+75 -15p +75 =

det A =0>p=5
Case I:When p =5,q *9
The rank (A) =
2 and rank
[A, B] =3
The system will be
inconsistent. The system will not
1: When have any solution.
Case p *5, det A+0
The system will have unique solution.
Case IlI : p=5,q =9
rank A 2 and rank [A, B]= 2
Number of variables = 3
The system will be consistent and will
have infinite number of solutions.
s 64, SYSTEM OF HOMOGENEOUS LINEAR
EQUATIONS
Consider a system of m homogeneous linear equations in n unknowns,
namely
G11*1 +a12*2 t
+ann=0
a211 +a22*2 t +a2n*,= 0
a311 ta32*2 t... +agnn =0
.. (1)

*
*****

am1 +am2*2 t + a m n n = 0

We write
0
11 2 . n
21 a22..02
A =| ,X = , 0 |0
(2)
aml m2mnmxn
mxl
Then (1) can be written as AX = O which is the matrix equation.

Here A is called the coefficient matrix. It is clear that

solution of (1) ie., X =|0 is a solution of (2).

=0=x2 =^ =.. =x, is a

This is called trivial solution of AX

=
O.
it has solution.
=Ois always consistent (i.e.)
a
This AX
solution.
hetrivial solution is called the zero
Zero solution is not linearly independent
(i.e.), it is linearly dependent.
 B solul.

Text
Book
ol
independent the
mber of ariable
n u m b

A being
amdn

linearly

118 n u m b e r
af m a t r i r A r

Theorem:
The
rank afthe system
AX= 0
0 .has only
the linear

n-s r being 0)
then
the

det A4
A N = 0

Note
IfA is
a
non-singular
matrix.
(i.e,

solution i land
only if A
is a
non-singula
ne

I. the zerosolution.

a
non-zero

O
p o s s e s s e s

AX = O
Al=
system Equation
2. The
matrix the
Solutions
ofthe
Rule for Finding =

Working
rand rank of[A/B] only trivial
blution
sal..
ofA
=
have
Let rank then

Since all
b's are
zero. r
=

l,. the system ofequations

of variables) soluti.
1)
Ifr={number

(i.e.. zero solution).

have an
infinite
number of n o n - t r i v i a l
tions,
the system ofequations solutions.
(i)
Ifr<n
linearly
independent value
value and solve fors for the
any arbitrary
arbitrary
have n - r
weshall solutions. set variables
)variables
r)
infinite (
Toobtain unknowns, it has a non-trivid
remaining unknowns. number of
of equations is less than the infinite.
If the number AX=0 will be
of the equation
solution. The number ofsolutions the solution is alwa
is less than number of variables,
II. If the number of equations
other than a trivial solution. and sufficient condition
number of variables, the necessary
IlI. Ifthe number ofequations=
that the determinant of the coefficient
for solutions other than a trivial solution is
matrix is zero.
SOLVED PROBLEMS
Ex. 1. Solve completely the svstem of equations
r+-22 + 3w =0; x-2 +3- w = 0; 4x +y- 5z + 8w = 0; 5x - 7y + 2z -w=0
Sol. The given system of equations in matrix form is
[ -2 3
4X= 21 -1 = 0
4 -5 8
5 -7 2
-w

Applying R,-R,, R, - 4R, and

R, 5R,. we get -

A
0 -3
-23
3
-
4
0 -3 3 -4
0 -12 12 -16
Applying , we get A~
-2 3
0
4
-33-4|
-3 3 -4
0-3 3
 -2
Applying R,- R, and R-R2, we get 3
0-3 3
A p p l y i

A ~
-4
0
in the Echelon
form. We ha
This is 0 0 0
A the number
=
of 0
k non-zero rows in
A(=2) is less than the this Echelon
Sin rank number form =2
of
has1 nfinite
number of non-trivial
solutions. unknowns (=4), therefore, the given sysc
system
Number of independent solutions
4-2-2. =
laW, we shall assign arbltrary
values to 2
of these. variables and the remaining 2
found
be
in terms The given variables shall
system of equations is
-2 3x [ equivalent to
0-3 3 -4 y 0
0 0 o
0 0
0 ow
This gives the equations x *y-2z +3w0,
-3y +3z -4w =0
5
Taking z \and w=4, we see that x =-4, y=n-
4
H, z= A, w=p constitutes
3
the general solution of the given system.
Ex. 2. Show that the only real number n for which the
system
x+2y + 3z = x ; 3x + y + 2z = y ; 2x + 3y +z z

has non-zero solution is 6 and solve them, when = 6.

Sol. Given system can be expressed as AX = O where

1- 2 3 0

A31-2
2 3 1-J
2 X =|
z
and O =0

Here number of variables = n =3.

if
ne given system of equations possess a non-zero (non-trivial) solution,
< number of unknowns i.e.,
Rank of A< 3
Rank of A
For this we must have det. A = 0

1- 2
3 1- 2 0
2 3 1- 6- 6-A 6-
1-3 2 = 0
have 3
Applying R,->R, +
R, R, + we

2 3 1-2 |

= 0
(6-1.) 3 1- 2
| 3
 UI
BooK I
A Text
120 3-2- -1 0
(6-2)
C, we get 2 1 -1-
andC,>C, -
C-C,
Applying C,
other are
values are complex
co

6 - 3 ) [-2-2)(-1-2)+1]=0 value and

real
the only
l 6 is
+3)= 0
=
~
(6-.) (A*+ 3 becomes
system
When = 6 , the given

-5 2 3
3 y=0
3 -5 2 37
0 -19 19

R2SR2 +3R),R3 > 5R3 +2R1 19 -19 L:J lo

-s
0
2
-19 19
11
RsR +R 00 0J
-5x+2y+3z =
0 and -19y +19z = 0 >y=z
Since Rank of A < number ofunknowns (Rank of A =2, number of unknowns = 3),

therefore, the given system has infinite number of non-trivial solutions

Let z =k *y=k and -5x+2k +3k = 0 =x=k

X=k,y =k,z =k isthe solution.

Ex. 3. Solve the system 2x-y+3z = 0, 3x +2y+z = 0 and x-4y+5z = 0.
Sol. The Given system can be written as AX =
0

2 1 3
where A=|3 2 1, and O=0
z
1 4 5
Consider A-|3 2
1| (Applying R3 t> R)
2 -1 3

1-4 51
0 14 14
0 7 -7 (Applying R2 -3R and
R3-2R)
4 51
00 14 -14
0 0 0 (Applying 2R,-R,)
 4 5
0 (Applying
0 0 0] 14
dince the number of
non-zero rows is
Here number of variables = 3 2, we have rank (A) =
2
The system will have n-r =3-2 =1 non-zero
solution.
From the matrix, we have y-z =
0y =z
Let y=z=k. Then
-4y +5z =0 >x =
4y-5z 4k -5k - k
=

The solution is given by X =|y|=|k=k1

Ex. 4. Find all the solutions of the system of equations

x+2y-z =0,2x + y+z = 0,x-4y+5z =

1 2 -1
Sol. Let A =|2.1 1 Then
4 5

det A 1 (5 +4)-2 (10-1)-1(-8-1) =9-18+9

= = 0

The rank of A is <3.

Applying R2 > R2-2R and R3>R3 -R1, we get

2 -1
A 0-3 3
0 6 6

Applying R3 > R3-2R2, weget

2-1
0-3 3
A
0 0 0
solutions =
n-r =3-2=1.
2. Number ofnon-zero
Thus rank (A) =

>y=z
FTom the above matrix, -3y +3z =0

Let z =k. Then y=k.

>x=Z-2y
=k-2k =-k
+2y-z =0

T
by X=k|
solutions are given
The