Engineering Mathematics-I(KAS-103T)

Course Outcomes:The students will be able to apply the concepts of elementary

transformations to find the inverse of a matrix, rank of matrix, solve the system of linear
algebraic equations and will able to determine Eigen values, Eigen vectors and diagonalize a
matrix
 MATRICES

This subject is of extreme importance in engineering problems, since a great many physical
phenomena are expressed in terms of linear differential equations. By appropriate
transformations the solution of a set of linear differential equations with constant coefficients
can be reduced to a set of linear algebraic equations which can be solved by using matrix
algebra.J. J. Sylvester was the first to use the word matrix in 1850 and later on in 1858 Arthur
Cayley has discussed this subject in detail. Matrix is used by Economists in the study of inter-
industry economies, input-output tables. Statisticians use matrix theory to study design of
experiments and multivariate analysis. Its study is helpful in electrical and communication
engineering.

DefinitionA matrixis a rectangular array of numbers written inside a parenthesis. A matrix of

m rows and n columns is expressed as A = [aij]m,n and is written as

 a11 a12 a13  a1n 

 
 a21 a22 a23  a2 n 
A  [aij ] mn  (1)
................................... 
 
 am1 am 2 am3  amn 

The individual entries in the matrix are called its elements; the quantity aij in the above matrix
is the element in the ith row and jth column. The subscripts used to indicate the elements will
always denote the row first and the column second. A matrix may be represented by the
symbol (aij), || aij|| or by a single letter A. If the number of rows of a matrix is the same as
number of columns, it is said to be a square matrix i.e. m = n.

A set of mn elements of a field F arranged in the form of a rectangular array having m

rows and n columns is called m × n matrix over the field F. If all the elements of a matrix
belong to the field of real numbers, the matrix is said to be real. If elements of the matrix are
complex number, the matrix is said to be a complex.

Matrices are classified by the properties of their elements aij which are pure numbers
(complex or real), when rows and columns are interchanged or according to their behaviour
when two of the matrices are added or multiplied under special rules.

Special Types of Matrices

(i) Square matrix

If the number of rows and columns of a matrix A are equal to n, then such a matrix is
called a square matrix of order n, or simply a matrix of order n and written as
 a a12 ....a1n 
 11 
 a21 a22 ....a2 n 
A  [aij ]nn   (2)
 ....................... 
 
 an1 an 2 ....ann 

(ii) Row or column matrix

When there is only one row so that m = 1, the matrix is termed as a vector of the first
kind or a prime. As an example, we have

R1  [r11 r12 r13 ...r1n ] . (3)

On the other hand a matrix of a single column of n elements will be called a vector of second
kind or a point.

As an example, we have

c11 
 
c 21 
C1  c31  (4)
 
 
c n1 

(iii) Diagonal matrix

Those elements aij of any matrix [aij] are called diagonalelements for which i = j. The line
along which the above elements lie is called Principal diagonal or the diagonalof the matrix.

A square matrix in which all elements except those in the diagonal are zero is known
2 0 0

as diagonal matrix. For example  0 3 0  is a 3- rowed diagonal matrix. The sum
 0 0 7 
of the diagonal elements of a square matrix is called the trace ofthe matrix. If all elements of
4 0 0

a diagonal matrix are equal, then it is called scalar matrix. For example,  0 4 0 
 0 0 4 
is a scalar matrix.

(iv) The unit matrix or identity matrix

The unit matrix or identity matrix of order n is defined to be the diagonal matrix of order n in
which all diagonal elements are 1 i.e. aii = 1 for all i. It is denoted by Un or In or simply by U
or I. Unit matrix of order 3 is given by
 1 0 0
I 3   0 1 0 
 0 0 1  33

(v) Symmetric matrix

If aij= aji for all i and j in a square matrix [ aij], then it is said to be symmetric matrix. A
symmetrical matrix of order 3 is given by

1 2 3
2 4 5 

 3 5 6  33

(vi) Skew – symmetric matrix

If aij = - aji for all unequal i and j but all the elements are not zero, then that square matrix is
called a skew matrix.

If aij= - aji for all i and j so that all aii = 0, then square matrix is called skew-symmetric
matrix. Examples of these two types of square matrices of order 3 are respectively given
below:

 1 2 3  0 2 3
 2 4 5  and  2 0 5 
 
  3 5 6  33   3 5 0  33

(vii) Upper triangular matrix

A square matrix A = [aij] is called an upper triangular matrix if aij = 0 whenever i >j. Thus, in
upper triangular matrix all the elements below the principal diagonal are zero. For example,

1 2 4 2
0 2 4 5
3 1 5  0
A  and B   1 2 
0 0 4 3
   0 0 4 
0 0 0 6

are respectively 4×4 and 3×3 upper triangular matrices.

(viii) Lower triangular matrix

A square matrix A = [aij] is called a lower triangular matrix if aij = 0 whenever i <j. Thus in a
lower triangular matrix all the elements above principal diagonal are zero. For example,
 1 0 0 0
2 1 0 
0  2 0
A  and B  
4 2 3 0  4 3 
 
5 4 3 4

are 4×4 and 2×2 order lower triangular matrices respectively.

(ix) Null matrix or Zero matrix

The m × n matrix whose elements are all 0 is called null matrix ( zero matrix) of order m × n.
It is usually denoted by 0 or 0m,n. For example,

0 0 0 0 0 0 0
0 0 0 0  and 0 0 0 
 
 0 0 0 0  34  0 0 0  33

are zero matrices of the type 3×4 and 3×3 respectively.

(x) Sub-matrix
A matrix which is obtained from a given matrix by deleting any number of rows or columns
 4 5
is called a sub-matrix of the given matrix. For example   is a sub-matrix of the matrix
 5 6 
given in (v).

Transpose of Matrices

The accented matrix A  [ aij ]  [a ji ] obtained by a complete interchange of rows and

columns in A = [aij] is called transposed matrix of A. The ith row of [aij] is identical with ith
column of [ aij ] . For a rectangular matrix of order 2×3, the transposed matrix is of order 3×2
as given below:

  a a21 
 a a12 a13   11 
 11    a12 a22 
 a21 a22 a23   
 a13 a32 
 
For a symmetrical matrixA = [aij] , we have

A  [aij ]  [ a ji ]  [ aij ]  A

Transposed matrix is sometimes written as AT.

Equality of Matrices

From the definition of the matrix, it is clear that a matrix is merely a square or rectangular
array of quantities. By defining certain rules of operations, a certain algebra may be
developed that has similarity to ordinary algebra but operations are performed on the
elements of the matrices. In matrix algebra two matrices A and B of the same order are said
to be equal if and only if their corresponding elements are identical. Then, we have

A=B

provided that aij = bij for all i and j. (5)

From the definition given above, it is evident that

(i) If A = B, then B =A (Symmetry)

(ii) A = A, where A is any matrix (Reflexivity)

(iii) If A = B and B = C then A = C (Transitivity).

Hence, the relation of equality in the set of all matrices is an equivalence relation.

Remark: The values of two determinants can be equal though they are of different order but
it is not so for matrices.

Addition and Subtraction of Matrices

If A and B are matrices of same order, then the sum of A and B is defined to be a matrix C i.e.

A + B = C,

provided the typical element cijof C is equal to

cij= aij + bij . (6)

In a similar manner we define the subtraction as

D = A- B,

provided that D is of the same order as A and B and its typical element dij is given by

dij = aij – bij (7)

Properties of Matrix Addition: Let A, B, C be three matrices of order m × n

(i) A + B = B + A (Commutative Law)

(ii) (A + B) + C= A + (B + C) [Associative Law]

(iii) Zero matrix of order m × n i.e. Om×n is additive identify because

A + O = O + A = A.

(iv) - A is additive inverse of A because

A + (-A) = O.
(v) Let A be a square matrix of order n, then

(A+ A ) = [aij + aji] and  A  A = [aji + aij] =  A  A .

which shows that A+ A is a symmetric matrix.

(vi) (A- A ) is a skew-symmetric matrix because


(A- A ) = [aij - aji] and  A  A = [aji- aij] = - [aij - aji] = - (A - A ).

Results of (v) and (vi) may be put in the following theorem:

THEOREM 1 Any square matrix can be uniquely written as the sum of a symmetric matrix
and a skew-symmetric matrix.

Proof Let A be any square matrix of order n, then we may write

1 1 (7)
A  ( A  A' )  ( A  A' )
2 2

According to the properties (v) and (vi) the first part and second part of the right hand side of
Eq.(7) are respectively symmetric and skew-symmetric. For uniqueness, we take

A = P + Q, (i)

where P and Q represent symmetric and skew-symmetric matrices respectively.

A  P  Q  P  Q (ii)

Solving (i) and (ii) for P and Q, we get

1
P  A  A and Q  1  A  A
2 2

This shows that the representation given in eq. (7) is unique. Hence the theorem is proved.

Multiplication of Matrices
(i) Scalar multiplication
Multiplication of a matrix by a scalar k results in a new matrix B defined by

k A = B,

where bij = k aij .

Thus the multiplication of a matrix by a scalar quantity yields a new matrix whose elements
are obtained by multiplying the elements of the original matrix by the scalar multiplier. This
is not true for a determinant in which k multiplies only a single row or a single column. This
is demonstrated below for a square matrix of order 2×2. For a matrix A, A   2 4
we
 6 5 
have 6A=  12 24 
but for a determinant
 36 30 

2 4 12 24 2 4 12 4 2 24
|A|  , 6 |A|     .
6 5 6 5 36 30 36 5 6 30

(ii) Matrix multiplication

Two matrices can be multiplied only when the number of columns of the first is equal to the
number of rows of the second. Matrices that satisfy this condition are termed as conformable
matrices. Let A = [aij]m×p and B = [bij]p×n be two conformable matrices, then their product is a
matrix C = [cij]m×n such that
p
cij  
k 1
ai k bk j .

The order of the matrix C will be m × n. This is written as:

C = AB

In the product AB, the matrix A is called the pre-factor and B is called post factor. Also we
say that the matrix A has been post-multiplied by B and the matrix B has been pre-multiplied
by A. As an example, let us consider the multiplication of the matrices A and B given below:

 a11 a12 a13   b11 b12 

   
A  a 21 a 22 a 23  by B   b21 b22 
 a31 a 32 a33   b31 b32 
 

By applying the definition, we obtain

 a11b11  a12 b21  a13 b31 a11b12  a12 b22  a13 b32 
 
AB=  a 21 b11  a 22 b21  a 23 b31 a 21b12  a 22 b22  a 23 b32 
 
 a31b11  a32 b21  a33 b31 a31b12  a32 b22  a33 b32 
 

Example1 Compute AB given that

 3 1 
 2 1 3   1
A   , B  2 
 1 2  1  
  1 1 

Solution

 2.3  (  1) .1  3 .(  1) 2. (  1)  (  1) . 2  3.1   2 1 
AB      .
1.3  (  2) .1  (  1) .(  1) 1. (  1)  (  2) .2  (  1).1   2  6 

It may be noted that A is a 2×3 matrix and B is 3×2 matrix. Therefore the product can be
formed, and the result is a 2×2 matrix. The product BA can also be formed, the result being
3×3 matrix. It is clear that AB  BA, since AB is a 2×2 matrix and BA is a 3×3 matrix. In
general, multiplication of matrices is not commutative. This is true even for square matrices.
Another example

4 4  1 1 0 0
If A    and B    , then AB  
3 3 1  1 0 0

i.e. the product of two non-zero matrices can be a zero matrix.

 1 1  4 4   1 1
BA      0.
1  1  3 3   1 1 

Hence AB  BA.

Example 2 (a) Show that the product of two given matrices can be zero without either of the
matrices being a zero matrix.

(b) Find the product of following two matrices

0 c b  a 2 ab ac 
 
A    c 0 a  , B  ab b2 bc 
b a 0  ac bc c 2 


Solution (a) Let us choose two non-zero matrices A and B as under

1 1  1 1 
A  ,B , then
3 3 1 1

1. (1)  1.1 1.1  1. (1)  0 0

AB     0.
3.(1)  3.1 3.(1)  3. (1)  0 0

0.a2  c. ab  (b).ac [Link]  c.b2  b.(bc) [Link]  c.(bc)  (b) .c2 

 
(b) AB  (c).a2  [Link]  a .(ac)  c.(ab)  0.b2  a.(bc) c.(ac)  0.(bc)  a.c 2 
 2 
b.a  (a). ab  0.(ac) b.(ab)  (a).b2  0.(bc) b.(ac)  (a) .bc  0.c 2 


0 0 0
 0 0 0
0 0 0

Properties of Matrix Multiplication

(i) Multiplication of matrices is associative if conformability is assumed i.e.

A (BC) = (AB) C.
(ii) Multiplication of matrices is distributive with respect to the addition of matrices i.e.

A (B+C) = AB + AC,

where A, B, C are any three matrices of order m × n, n × p, n × p.

(iii) The multiplication of matrices is not always commutative. This has been discussed in
detail earlier and there is no general proof.

(iv) The equation AB = 0 does not necessarily imply that at least one of the matrices must
be zero. This has been illustrated earlier and there is no general proof.

(v) Positive power of a square matrix: The product AA is defined only when A is a square
matrix of order n. We shall denote it as A2. Similarly

A2A = (AA) A = A (AA) = AA2 = AAA = A3.

If m and n are any positive integers, then we have

Am An = Am+n .

(vi) The reversal law in transposed matrices

One of the fundamental consequences of non-commutative law of matrix

multiplication is the reversal law in transposing a continued product of matrices. Let A and B
be two matrices conformable for multiplication. Then we have following theorem:

THEOREM 2 When a matrix product is transposed, then the order of the matrices forming
the product is reversed.

That is

( BA)'  A' B '

and similarly

( ABC )'  C ' B ' A' .

These results can be put in the following theorem.

Suppose A is a square matrix of order n and B is a diagonal matrix of order n. Then

pre-multiplication by a diagonal matrix has the effect of multiplying every element in any
row by a constant and that post-multiplication by a diagonal matrix results in multiplication
of every element in any column by a constant.

(vii) Multiplication by a diagonal or unit matrix

If In is a unit matrix of order n, then we have

A In = In A = A. (8)
Thus unit matrix plays the same role in matrix algebra that number 1 does in ordinary scalar
algebra.

 a11 a12 a13   d1 0 0 

   
Example Let A   a21 a22 a23  and D =  0 d2 0  , then it can be verified
a a32 a33  0 0 d 3 
 31 
that

d1 a11 d1 a12 d1 a13 

 
DA   d 2 a21 d 2 a22 d 2 a23  .
d a d a d a 
 3 31 3 32 3 33 

 d1 a11 d 2 a12 d 3 a13 

 
Similarly AD   d1 a21 d 2 a22 d 3 a23  .
d a d a d a 
 1 31 3 32 3 33 

Adjoint of a Square Matrix

Let A = [aij]n×n be a square matrix given by Eq. (2) and Aij be the cofactor of the element aij in
the determinant |A| of the matrix. Then the transpose C ' of the matrix C = [Aij]n×n i.e.
C '  [ A ji ]nn is called the adjointof A and is denoted Adj. A.

The adjoint of the matrix A given in Eq.(2) is

 A11 A21  An1 

 
A12 A22  An 2  .
Adj. A   (9)
.......... .......... ..... 
 
 A1n A2 n  Ann 

Sometimes the adjoint of matrix is called adjugate of the matrix.

Example Find the adjoint of the matrix

 2 4 3
A   1 2  2 
  3 3 2 

The matrix C and C' are respectively given by

 2 4 3   2 1 2 
C   1 13  18  and C'   4
 13 7 .
  2 7  8    3  18  8 
The matrix C' is adj A.

Properties of an Adjoint Matrix:

THEOREM 3 If A be any n-rowed square matrix, then

(adj A) A = A (adj A) = |A| In , (10)

where |A| is the determinant of the matrix A and In is the n-rowed unit matrix.

This theorem states that the matrix A and adj A are commutative, and their product is a scalar
matrix, every diagonal element of which is |A|.

Corollary 1 If A is a n-squared singular matrix, then

A (adj A) = (adj A) A = 0(null matrix). (11)

A matrix is said to be singularif its determinant is zero i.e. |A|=0, and hence in this case all
elements of A (adj A) and (adj A) A will be zero and hence the null matrix. If |A| ≠ 0, then A is
said to be non-singular matrix.

Corollary 2 It clearly follows from Eq. (10) that |A| |adj A | = |A|n = | adj A| |A|

or |adj A| = |A|n-1 provided |A| ≠ 0.

Inverse of Matrix

Definition If A be any square matrix of order n, then a matrix B, if it exists, such that

AB = BA = In (12)

is called inverseof A. The matrix B is also a square matrix of order n. The matrix B will exist
only when A is a non-singular matrix i.e. |A|  0. If we divide the relation (10) by |A| and
compare it with Eq. (12) ,we observe that

(adjA)
B . (13)
|A|

The matrix B is written as [Link] the Eq. (12), it is obvious that A=B-1. Matrices A and B
and called reciprocalof each other. The inverse of A possesses the following properties:

(i) Inverse of a matrix is unique

Let A be an n-squared matrix and has two inverses B and C, then by definition
AB = BA = I
AC = CA = I.
We have
C = CI = CAB = I B = B.
Hence C = B.

(ii) A-1plays the role of division i.e. if we have

AB = CD,

then pre-multiplying by A-1 the above equation gives

A-1AB = A-1CD

or B = A-1CD.

(iii) Reversal Law

THEOREM 4 When inverse of a product of matrices is taken, the order of the product is
reversed.

That is , (AB)-1 = B-1A-1 (14)

and (ABC)-1 = C-1B-1A-1 (15)

(iv) Negative Powers: If A is a non-singular matrix then the negative powers of A are
defined by raising the power of inverse matrix of A i.e. A-1 to positive powers. That is

A-n= (A-1)n (16)

(v) The operations of transposing and inversing are commutative: Let A be a square
matrix of order n, then we have

AA-1 = A-1A= I.

Taking the transpose of the above equation, we get

( A1 )' A'  A' ( A1 )'  I '  I .

This implies that ( A1 )' is the inverse of A' i.e.

( A1 )' = ( A' )-1 (17)

Example 1 Find the inverse of a square matrix of order 2.

Solution Let A be square matrix of order two given by

 a11 a12 
A .
 a21 a22 

The matrix B formed of the cofactors of the elements of A is

 a22  a21 
B 
 a12 a11 

a a12  1 a a12 
adj A  
22
 , A1   22  (18)
  a21 a11  (a11 a22  a12 a21 )   a21 a11 

Example 2 Find the inverse of

1 0 2
A   2 1 0 
 3 2 1 

and verify that AA-1 = I3

Solution Here |A| = 3, so A is non – singular. Now

1 4 2
adj A    2 5 4 
1 2 1 

1
A1  adj A
3
 1 4 2
 3  
1 0 2 3 3 1 0 0
 
AA 1
  2 1 0   2 
5 4 
  0 1 0  = I3.
 3 3 3 
3 2 1     0 0 1 
 1 
2 1 
 3 3 3 

Various Kinds of Matrices

(i) Periodic matrix: A square matrix A will be called a periodic if

Ak+1 = A (19)

where k is a positive integer. If k is the least positive integer for which this relation holds
good, then k is called the period of A. If k = 1, then we get

A2=A (20)

and the matrix is called idempotent.

Some results

(a) If A and B are idempotent matrices and A, B commute, then AB is idempotent.

(AB) 2 = (AB) (AB) = A (BA) B (associative law)

 = A (AB) B = A2B2= AB.

(b) If A is idempotent, then the matrix B = I – A is idempotent and AB = 0 = BA .

We have

(I-A)2 = (I-A) (I-A) = I2 – IA – AI+ A2

= I – A- A+A = I-A.

AB = A (I-A) = AI – A2 = A-A = 0

Similarly BA = 0.
(c) If A and B are square matrices of order n and AB = A and BA = B then A and B are
idempotent, we have A2 = A.A=ABA = A (BA) = AB=A
and B2 = BAB = BA = B.
Example 1 If A is a diagonal matrix of order 3, find the condition that A is idempotent.

d 0 0   d12 0 0 
 1   
Let A = 0 d2 0  , then A2  0 d 22 0 
   
0 0 d3  0 0 d 32 
  

Therefore the condition is that d i2  d i i.e. d i  0 or 1 .

Hence I3 or O3 are diagonal matrices which are idempotent.

2 3 5
Example 2 Show that A    1 4 5  is idempotent .
1 3 4

2 3 5 2 3 5
Solution We have A   1
2
4 5   1
 4 5 
1 3 4 1 3 4

2.2  (3). (1)  5.1 2.(3)  (3). 4  (3). (5) 2. (5)  (3).5  (5).(4)

 (1).2  4.(1)  5.1 (1). (3)  4.4  5. (3) (1).(5)  4.5  5. (4) 

1.2  (3). (1)  (4) .1 1.(3)  (3).4  (4) .(3) 1.(5)  (3).5  (4).(4) 

2 3 5
  1 4 5   A .
1 3 4

Hence A is idempotent.
(ii)Involutory matrix: A square matrix A is called involutoryif

A2 = I. (21)

As I2 = I, so identity matrix is involutory.

THEOREM 5 A is involuntary if and only if (I+A) (I-A) = 0.

Proof (I+A) (I-A) = I2 –A2 = I2 - I = 0

Conversely, if (I+A) (I-A) = I2 - A2 = 0, then

A2 = I2 = I.

 5 8 0 
Example Show that A  3 5 0  is involuntary.
1 2 1

 5 8 0   5 8 0 
A  3
2
5 0  3
 5 0 
1 2 1 1 2 1

 5 .(5)  8.3  5.(8)  (8).5 0  1 0 0

 3. (5)  5.3 3.(8)  5.5 0   0
  1 0  I .
 5.1  2.3  1 (1) 8.1  5.2  2.(1) (1). (1) 0 0 1 

(iii) Nilpotent matrix: A square matrix A is called nilpotent matrix of order (index) m,
provided it satisfies the relation

Am = 0 and Am-1  0, (22)

where m is any positive integer.

ab b2 
Example 1 Show that  2  is a nilpotent matrix of order 2.
 a ab 

SolutionWe have

a 2 b 2  a 2b 2 ab 3  ab 3  0 0
A  3
2
   0.
 a b  a b
3
a 2 b 2  a 2 b 2  0 0

1 1 3 
Example 2 Show that A  5 2 6  is nilpotent matrix of index 3.
 2 1 3
 0 0 0 
SolutionWe have A  3
2
3 9 
 1 1 3

0 0 0
and A  0
3
0 0 .
0 0 0

This shows that A is nilpotent matrix of index 3.

(iv) Conjugate of a matrix : If A =[aij] is a given matrix, then the matrix obtained by
replacing all the elements by their conjugate complex is called the conjugate of matrix A and
is denoted by A  [aij ] , (23)

where a ij is conjugate of the corresponding element aij. If

1  i 4i 3  4i 
A
4 3i 1  i 

1  i 4i 3  4i 
then A . (i)
4 3i 1  i 

Properties of conjugate matrix:

(a) Conjugate of the conjugate of a matrix A is the matrix itself i.e.

A  A . (24)

(b) Conjugate of the sum of two matrices is sum of their conjugates i.e.

A B  A  B .

(c) The conjugate of the product of two matrices is the product of their conjugate in the
same order i.e.

AB  A B .

(d)The transpose of the conjugate of a matrix is equal to the conjugate of the transpose of the
matrix A i.e. A  '  A' .   (25)

The transpose of the conjugate of the matrix A is called tranjugate of A and is denoted

by Aθ. The tranjugate of matrix A given by Eq. (i) is

 1  i 4 

A  ( A )'    4 i
θ
3  i 
3  4 i 1  i 

It may be noted that (Aθ)θ=A.

(v) Hermitian and skew – Hermitian matrices: A square matrix A = [aij] is called

Hermitianif every i-j th element of A is equal to the conjugate complex of j-i th element

of A i.e.

aij= a ji .

This implies that the matrix A is Hermitianif

A = ( A ) '  Aθ . (26)

A square matrix A = [aij] is called a skew- Hermitian if i-jth element of A is equal to negative
conjugate complex of j-ith element of A i.e.

aij= - a ji .

This implies that a matrix A is skew -Hermitian if

A = - Aθ. (27)

Example 1 The following matrix is Hermitian

l a  ib c  id 
a  ib m x  iy 

c  id x  iy n 

where a, b, c, d, x, y, l, m, n are constants.

Example 2 The following matrix is skew-Hermitian

 il a  ib c  id 
 a  ib im x  iy 

 c  id  x  iy 0 

where a, b, c, d, x, y, l, m are constants.

2  i 3 1  3i 
Example 3 If A=  .
 5 i 4  2 i 

Show that AθA is a Hermitian matrix.

 2  i 5 

SolutionWe have A  ( A )'  3
θ
 i  .
 1  3i 4  2i 

It may be observed that A is 2×3 matrix and Aθ is 3×2 matrix, hence pre-multiplication of A
by Aθ is conformable and resulting matrix B = AθA is a 3×3 square matrix. The matrix B is
given by

30 6  8i  19  17i 
B  A A  6  8i
θ
10  5  5i 
 19  17i 5  5i 30 

30 6  8i  19  17i 
B  6  8i 10  5  5i 
 19  17i 5  5i 30 

30 6  8i  19  17i 
( B )  6  8i 10  5  5i  =B.
 19  17i 5  5i 30 

Properties of Hermitian and skew-Hermitian matrices

(a)The diagonal elements of a Hermitian matrix are real.

Let α  iβ be a diagonal element of a Hermitian matrix, then it must be equal to its conjugate
complex i.e.

  i    i 

This relation implies that   0 .

(b)The diagonal element of a skew-Hermitian matrix is either zero or imaginary

Let   i  be a diagonal element of a skew-Hermitian matrix then its conjugate complex

must be its negative i.e.

  i    (  i  )     i  .

This implies that   0 . Hence its diagonal will be i  or 0 in case  =0.

(c)If A is Hermitian, then (AAθ) is also Hermitian, we have

(AAθ)θ= (Aθ)θ Aθ = AAθ,

Hence AAθ is Hermitian.

(d) If A is a Hermitian matrix then iA will be a skew-Hermitian matrix.

If A is a Hermitian matrix then Aθ = A

and i Aθ   i Aθ = - (iA).

This shows that iA is skew- Hermitian.

(e) If A is any square matrix of order n, then (A + Aθ ) is a Hemitian matrix.

Let A be any square matrix, then

A  A  θ θ
 Aθ  (Aθ )θ

= Aθ + A = A+ Aθ.

Hence (A+ Aθ) is a Hermitian matrix

(f) If A is any square matrix of order n, then (A- Aθ) is a skew -Hermitian matrix

Let A be any square matrix, then

A  A  θ θ
 Aθ  ( Aθ )θ

= Aθ – A = - (A – Aθ).

Hence ( A- Aθ )is skew-Hermitian.

THEOREM 6 Every square matrix with complex elements can be uniquely written asa sum
of a Hermitian and skew -Hermitian matrix.

Proof Let A be a square matrix of order n, then we can write it as

A
1
2
( A  Aθ ) 
1
2

A  Aθ  B  C (28)

1 1
where B = ( A  Aθ ) is Hermitian and C  ( A  Aθ ) is skew-Hermitian by the property
2 2
(e) and (f) of Hermitian matrix and skew-Hermitiam matrix respectively. The uniqueness of
these matrices can be proved as in theorem 1.

(vi) Orthogonal Matrix: A matrix A is said to be orthogonalif

A' A  I . (29)

We know that | A' |= |A| so

| A' A| = | A' | |A| = |A|2  | A |   1 , (30)

If |A| = 1, then A is called unimodular.

Also, since |A| ≠0, the orthogonal matrix A is non-singular.

THEOREM 7 If A, B are n-rowed orthogonal square matrices, then AB and BA are also
orthogonal.

ProofSince A and B are both n-rowed matrices, therefore AB is also n-rowed square matrix.
We have

| AB |  | A | | B |  0 .

Now ( AB )( AB )  ( BA)( AB)

= B ' ( A' A) B  B ' IB  B ' B  I , as A' A  B ' B  I .

Hence AB is orthogonal. Similarly we can prove that BA is also orthogonal.

THEOREM 8 If A is a square orthogonal matrix, then A-1 is also orthogonal.

ProofWe have A an orthogonal matrix, then

A A=I.

Taking inverse of the above relation, we get

( A A)-1 = I

or A-1 ( A )-1 = A-1 ( A-1 ) = I.

Hence the inverse of orthogonal matrix is also orthogonal.

Example 1 If (li, mi, ni), i = 1, 2, 3 be direction cosines of three perpendicular lines, then
prove that

l1 m1 n1 
 
A  l 2 m2 n2  is an orthogonal matrix.
l m3 n3 
3

l1 m1 n1  l1 l2 l3 
   
Solution We have AA'  l2 m2 n2  m1 m2 m3 
l m3 n3  n n2 n3 
3  1

l12  m12  m12 l1 l2  m1 m2  n1 n2 l1 l3  m1 m3  n1 n3 

 
 l2 l1  m2 m1  n2 n1 l22  m22  n22 l2 l3  m2 m3  n2 n3 
l l  m m  n n l3 l2  m3 m2  n3 n2 l32  m32  n32 
3 1 3 1 3 1
 1 0 0
 0 1 0  I 3 .
0 0 1 

Hence the matrix A is orthogonal.

Example 2 Determine  ,  ,  when

0 2  
A      is orthogonal.
   

Solution We have

0 2β γ  0 α α  4 β 2  γ 2 2 β 2  γ2 -2 β 2  γ 2 
 
AA'  α β γ  2 β
 β  β   2 β 2  γ 2 α2  β  γ2 α2  β  γ2 
α β γ  γ γ γ  - 2 β 2  γ 2 α2  β  γ2 α 2  β  γ 2 


When A is orthogonal then A A = I, so equating the above matrix to I , we get

1 1
4 β 2  γ 2 1 and 2 β 2  γ 2  0  β   ,γ   .
6 3

1
Also α 2  β 2  γ 2  1  α   .
2

(vii)Unitary matrix: A square matrix A is said to be unitary if

Aθ A = I (31)

Since | Aθ| = | A |

| Aθ A |  | Aθ | | A |  | A | | A |  1 .

Thus determinant of unitary matrix is of unit modulus. Hence unitary matrix is non -singular.

THEOREM 9 If A, B are both n-rowed unitary matrices, then AB and BA are also unitary
matrices.

ProofMatrix AB being product of A and B which are n-rowed square matrices is also n-
rowed square matrix. Moreover, |AB| = |A || B|  0, hence AB is non-singular. We have

 AB θ ( AB)  ( B θ Aθ ) ( AB )  B θ ( Aθ A) B  B θ IB  B θ B = I . (4.58)

Hence AB is a unitary matrix. We can similarly prove that BA is also unitary.

NOTEA unitary matrix over the field of real numbers is an orthogonal matrix.

1 1 1  i
Example 1 Show that the matrix A   is unitary.
3 1  i 1 

1 1 1  i
Solution Aθ  1  i
3  1 

1 1.1  (1  i ) . (1  i ) 1. (1  i )  (1) . (1  i ) 
so Aθ A  (1-i)1.  (1).(1  i )
3  (1  i ).(1  i )  (1). (1)

1 1  1  i 2 0 1 3 0 1 0
    
3  0 1  i  1 
2
3 0 3  0 1 

1 3 0 1 0
Similarly AAθ   .Hence A is unitary
3 0 3  0 1 

0 1  2i 
Example 2 If N   , then show that (I-N) (I+N)-1 is unitary matrix.
1  2i 0 

Solution We know that a square matrix A is unitary if AAθ=I.

1 1 2 i
The matrix I  N    ,|I +N)|= 1-(-1-4)=6
 1  2 i 1 

and its inverse is given by using Eq. (4.43) as

1 1  (1  2i )
( I  N ) 1  .
6 1  2i 1 

1 1 2 i 
Now IN  
1  2 i 1 

1 1 1  2i  1  (1  2i )
 ( I  N ) ( I  N ) 1  1  2i
6  1  1  2i
 1 

1  4 2  4i 
   A.
6 2  4i  4 

1  4 2  4i 
Then Aθ    2  4i  4 .
6  

1  4 2  4i   4 2  4i  1 36 0 
Now AAθ       
36 2  4i 4   2  4i  4  36 0 36
 1 0
   I2 .
0 1 

This shows that (I-N) (I+N)-1 is unitary matrix