B.A.

(Programme) Semester-II Mathematics

DISCIPLINE SPECIFIC CORE COURSE

ALGEBRA
Study Material : UNIT 1-3

SCHOOL OF OPEN LEARNING

UNIVERSITY OF DELHI

Department of Mathematics
Editor : Dr. S.K. Verma
 Graduate Course

UNIT 1 : Theory of Equations and Expansions of Trigonometric Functions

Lesson 1 : Complex Numbers
Lesson 2 : De Moivre’s Theorem
Lesson 3 : Applications of De Moivre’s Theorem
Lesson 4 : Applications of De Moivre’s Theorem to Summation of Seric
Lesson 5 : Theory of Equations–I
Lesson 6 : Theory of Equations–II

UNIT 2 : Matrices
Lesson 1 : Matrices : Basic Concepts
Lesson 2 : Elementary Operations on a Matrix and Inverse of a Matrix
Lesson 3 : Rank of a Matrix
Lesson 4 : Systems of Linear Equations
Lesson 5 : The Characteristic Equation of a Matrix

UNIT-3 : Groups, Rings and Vector Spaces

Lesson 1 : Groups
Lesson 2 : Subgroup
Lesson 3 : Ring
Lesson 4 : Vector Spaces

Editor:
Dr. S.K. Verma

SCHOOL OF OPEN LEARNING

University of Delhi
5, Cavalry Lane, Delhi-110007
 Semester-II
Paper II: Algebra

Course Objectives: Students will get conceptual understanding and the applicability of the
subject matter, helps students to see how linear algebra can be applied to real-life situations.
Modern concepts and notation are used to introduce the various aspects of linear equations,
leading readers easily to numerical computations and applications.
Course Learning Outcomes: The course will enable the students to understand:
i) Solving higher order algebraic equations.
ii) Become aware of De Moivre’s theorem and its applications.
iii) Solving simultaneous linear equations with at most four unknowns.
iv) Get an overview of abstract algebra by learning about algebraic structures namely,
groups, rings and vector spaces.
Unit 1: Theory of Equations and Expansions of Trigonometric Functions
Fundamental Theorem of Algebra, Relation between roots and coefficients of nth degree
equation, Remainder and factor theorem, Solutions of cubic and biquadratic equations, when
some conditions on roots of the equation are given, Symmetric functions of the roots for cubic
and biquadratic; De Moivre’s theorem (both integral and rational index), Solutions of equations
using trigonometry and De Moivre’s theorem, Expansion for cos nx, sin nx in terms of powers of
cos x, sin x, and cosnx , sinnx, in terms of cosine and sine of multiples of x.
Unit 2: Matrices
Matrices, Types of matrices, Rank of a matrix, Invariance of rank under elementary
transformations, Reduction to normal form, Solutions of linear homogeneous and non-
homogeneous equations with number of equations and unknowns up to four; Cayley−Hamilton
theorem, Characteristic roots and vectors.
Unit 3: Groups, Rings and Vector Spaces
Integers modulo n, Permutations, Groups, Subgroups, Lagrange’s theorem, Euler’s theorem,
Symmetry Groups of a segment of a line, and regular n-gons for n = 3, 4, 5, and 6; Rings and
subrings in the context of C[0,1] and ℤn; Definition and examples of a vector space, Subspace
and its properties, Linear independence, Basis and dimension of a vector space.
References:
1. Beachy, J. A., & Blair, W. D. (2006). Abstract Algebra (3rd ed.). Waveland Press, Inc.
2. Burnside, William Snow (1979). The Theory of Equations, Vol. 1 (11th ed.) S. Chand &
Co. Delhi. Fourth Indian Reprint.
3. Gilbert, William J., & Vanstone, Scott A. (1993). Classical Algebra (3rd ed.). Waterloo
Mathematics Foundation, Canada.
4. Meyer, Carl D. (2000). Matrix Analysis and Applied Linear Algebra. Society for Industrial
and Applied Mathematics (Siam).
Additional Readings:
i. Dickson, Leonard Eugene (2009). First Course in The Theory of Equations. The Project
Gutenberg EBook (http://www.gutenberg.org/ebooks/29785).
ii. Gilbert, William J. (2004). Modern Algebra with Applications (2nd ed.). Wiley-
Interscience, John Wiley & Sons.
 UNIT 1

LESSON 1

COMPLEX NUMBERS

1.1 Complex Number

Definition : An ordered pair (x, y) of real numbers x, y is called complex number.
Illustration : (1, 6) (, –2), (0, 1), ( –1, 0), ( 2,1) , (1/2, 3) are complex numbers.
Two complex numbers : (x, y) and (p, q) are said to be equal if x = p, y = q.
In symbols, (x, y) = (p, q)  x = p and y = q.

1.2 Addition of Complex Numbers

If  = (a, b) and  = (c, d) be to complex numbers, we define their sum  +  to be the complex
number (a + c, b + d). Symbolically
(a, b) + (c, d) = (a + c, b + d)
Illustrations : (3, 7) + (4, –1) = (7, 6)
(2p, q) + (0, –2q) = (2p, – q)

1.3 Multiplication of Complex Numbers

If  = (a, b) and  = (c, d) be two complex numbers, we define the product  to be the complex
number (ac – bd, ad + bd).
In symbols (a, b) (c, d) = [(ac – bd), (ad + bc)]
Illustrations : (i) (2, 4) (–1, 3) = [2(–1) –4.3, 2.3 + 4 (–1)]
= (–14, 2)
(ii) (1, 0) (2, –3) = [(1.2) – 0.(–3), 1.(–3) + 0.2)]
= (2, –3)

1.4 Division
If , , be any two complex number (  0) then by definition    = –1. We shall also use the
symbols /, to denote   .

1.5 Usual Notation for Complex Numbers

We shall now see at to how we can express a complex number (a, b) in the usual way as
a + ib. The mapping a → (a, 0) associates to each real number a, a complex number (a, 0) whose first
number is ‘a’ and whose second number is ‘0’. This mapping is compatible with addition and
multiplication. That is, if ‘a’ and ‘b’ be any two numbers, then
a + b → (a + b, 0) = (a, 0) + (b, 0)
ab → (ab, 0) = (a, 0) (b, 0)
The above relations mean the if we have any relation involving real numbers, then it remains true if
each real number ‘a’ in the relations is replaced by the complex number (a, 0). Because of this important
property we can write any relation involving complex numbers.

1
 Also (0, 1) (0, 1) = (–1, 0) = – 1
which suggest that we may write (0, 1) = i
If we agree to write ‘a’ for (a, 0) and i for (0, 1) then
(a, b) = (a, 0) + (0, b)
= (a, 0) + (0, 1) (b, 0)
= a + ib
Thus we find that if we set up the convention of writing ‘a’ for (a, 0) and i for (0, 1), then
(a, b) will be written as a + ib (where i = −1 )
We shall give below, some examples to illustrate the Algebraic operation on complex numbers, with
usual Notation.
Example 1. Simplify (4 + 5i) + (2 – 7i)
Solution. (4 + 5i) + (2 – 7i) = (4 + 2) + (5i – 7i)
= 6 + (5 – 7) i
= 6 – 2i
We note that to add two complex numbers, we add the real parts and the pure imaginary parts
separately.
With the new notation, the same example can be solved as follows :
(4 + 5i) + (2 + 7i) = (4, 5) + (2, –7)
= (4 + 2, 5 – 7)
= (6, –2)
= 6 – 2i
Example 2. Simplify (2 + 3i) + (5 – 7i)
Solution. (2 + 3i) + (5 – 7i) = 2(5 – 7i) + 3i (5 – 7i)
= 10 – 14i + 15i – 21i2
= 10 + i (–14 + 15) –21(–1)
= 10 + 21 + i (15 – 14)
= 31 + i
With the new notation, we have
(2 + 3i) (5 – 7i) = (2, 3) (5, –7)
= [2 × 5 – 3 (–7), 2(–7) + 3 × 5]
= (10 + 21, –14 + 15)
= (31, 1)
= (31 + i)
Note : By the above two examples, it is evident that any one of the two notations can be used.

1.6 Real and Imaginary Parts of a Complex Number

Let z = x + iy be any complex number. Then x is called the real part of z and y is called the
imaginary part of z. We write R(z) = x, I(z) = y. We say that the complex number z is purely real if
I(z) = 0, and that z is purely imaginary if R(z) = 0. Also, the complex number x – iy is called the complex-

2
conjugate of x + iy, the symbol z is usually used to denote the complex-conjugate of the complex
number z. That is, if z = x + iy, then z = x – iy.
If z is the conjugate of z, then z is the conjugate of z that the relation of conjugacy between
complex number is a symmetric one. We therefore, generally say that z and z are conjugate complex
numbers instead of saying that z the complex-conjugate of z.
The following results follow directly from the definition.
(i) The sum of two conjugate complex numbers is purely real.
(ii) The difference to two conjugate complex numbers is purely imaginary.
(iii) The product of two conjugate complex numbers is a non-negative real number.
In order to see the truth of the above three statements, we have z = x + iy so that
z = x – i y i.e. z = (x, y) so that z = (x, – y), then
z + z = 2x which is purely real [ z + z = ( x, y ) + ( x, − y ) = (2 x, 0)]
z − z = 2iy, which is purely imaginary [ z − z = ( x, y ) − ( x, − y) = (0, 2 y)]
zz = x2 + y2, which is a non-negative real number.
[ zz = ( x, y) ( x, − y) = ( x2 + y 2 , 0]
We give below some solved examples with the usual notation which would help us at a later state.
a + bi
Example 3. Simplify
c + di
To divide two complex numbers, multiply both numerator and denominator of the fraction by the
conjugate of the denominator.
a + bi (a + bi) (c − di)
=
c + di (c + di) (c − di)
(ac + bd ) + (bc − ad )i
=
c2 + d 2
 ac + bd   bc − ad 
=  2  +  2
2  2i
c +d  c +d 
Example 4. Find the value of
(2 + i) (3 – 2i) (8 + i)
Solution. Here (2 + i) (3 – 2i) = 6 + 3i – 4i – 2i2
= (6 + 2) + (3i – 4i) [ i2 = –1]
= (8 + i)
 (8 – i) (8 + i) = 64 – i2 = 64 + 1 = 65
Example 5. Find the value of (1 + i)3.
Solution. (1 + i)3 = 1 + 3i + 3i2 + i3
= 1 + 3i – 3 – i
= – 2 + 2i

3
 Example 6. Divide (1 + 3i) by (2 + i).
1 + 3i (1 + 3i) (2 − i)
Solution. =
2+i (2 + i) (2 − i)
(2 + 3) + (–1 + 6)i
=
4 +1
5 + 5i
= =1+ i
5
5 − 2i
Example 7. Express in the form x + iy
3 − 4i
5 − 2i (5 − 2i) (3 + 4i)
Solution. Here =
3 − 4i (3 − 4i) (3 + 4i)
(15 + 8) (20 − 6)i
=
9 + 16
23 + 14i  3   4 
= =   + i 
25  25   25 
Example 8. Find the roots of the quadratic equation
x2 – 6x + 10 = 0.
Solution. This can be written as (x2 – 6x + 9) + 1 = 0
or (x – 3)2 = –1 = i2
 x–3= ±i
or x= 3±i
Thus 3 + i and 3 – i are the roots of the quadratic equation
x2 – 6x + 10 = 0
It is seen that the roots are conjugate. Verify that these values satisfy the equation. The same
examples can also be solved as
6  36 − 40 6  −4
x= =
2 2
6  2i
= =3i
2

1.7 Graphical Representation of a Complex Number

The complex number z = x + iy may be representated graphically by a point P whose rectangular co-
ordinates are (x, y) i.e., we associate with each complex number z = x + iy the point of the plane which
has, with reference to a fixed rectangular system, the coordinates x and y and conversely with each point
having the co-ordinates (x, y) we associate the complex number z = x + iy.
From each theorem involving complex numbers we can deduce a definite relationship between the
geometrical points of the Cartesian plane and conversely, the diagram showing points which represent
complex numbers geometrically is called Argand Diagram.

4
 It may be noted that all points on the x-axis are of the form (x, 0) and are x + oi = x and so
corresponds to real numbers x. Similarly points on the y-axis correspond to pure imaginary numbers yi.
In addition, the complex number may be repesented by the directed line segment or vector OP.

1.8 Trigonometric form of Complex Numbers

The complex number z = x + iy is represented by the vector OP. This vector and hence the complex
number is described in terms of the length r of the vector and the angle  which is vector makes with the
positive direction of the x-axis (real axis) measured positively.
Thus OP = r = x2 + y 2 and tan q = y/x
Thus a complex quantity can always be put in the form r (cos  + i sin ) where r and  are both
real.

Let x + iy = r (cos  + i sin )

Equating real and imaginary parts, we get
x = r cos , and y = r sin .
By squaring and adding, we get
x2 + y2 = r2
r= x2 + y 2

5
 Here, we take only positive sign of the square root.
y r sin 
Also by dividing = = tan 
x r cos 
y
or  = tan −1 ( x  0)
x
y
r = x2 + y 2 is called the modulus or absolute value of the complex number x + iy and  = tan −1
x
is called amplitude (or argument or phase).
There are many values of  satisfying the equation tan  = y/x, the value of , such that – <    is
called the principal value of the amplitude. We shall generally take the principal value of .
In symbols the modulus of a complex number z = x + iy is also denoted by | z | or | x + iy |
i.e., | z | = | x + iy |
= x2 + y 2
That amplitude is denoted by the symbols amp z or arg z.
amp (x + iy) = amp z = 
here the value of  is so choosen that it is the principal value i.e.,
– <   
The form r(cos  + i sing ) is called the standard or the polar form of the complex number
x + iy,  R, y  R.
Example 1. Express 1 + i in the standard form.
Solution. Let 1 + i = r (cos  + i sin )
Equating real and imaginary parts, we get
1 = r cos ; 1 = r sin 
Then r2 = 2  r= 2

Also tan  = 1 i.e., =
4
Here cos  is positive and sin  is positive, and hence  lies in the first quadrant.

 =
4
  
Hence 1+i= 2  cos + i sin 
 4 4

The modulus of the complex number is 2 and the amplitude is .
4
Example 2. Express 1 − i 3 in the polar form.
Solution. Let 1 − i 3 = r (cos  + i sin )
so that 1 = r cos ,
− 3 =  sin 
 r = 2, tan  = − 3

6
Here cos  is +ve and sin  is –ve.
  lies in the fourth quadrant
Also since – <   

= −
3
     
Hence 1 − i 3 = 2 cos  −  + i sin  −  
  3  3 
Example 3. Express sin  + i cos  in the standard form.
Solution. Let sin  + i cos  = r (cos  + i sin )
sin  = r cos   r2 =1
 
cos  = r sin  
so that
 tan  = cot 
 r = 1, tan  = cot 
 
= tan  −  
2 

 = −
2
   
Hence sin  + cos  = cos  −   + i sin  −   .
2  2 

EXERCISE–1
1. Simplify the following :
(i) (2 – 5i) + (–3 + 4i) + (8 – 3i)
5+i
(ii)
3 − 2i

2. Evaluate the following :

(i) (–7 – 2i) (–1 – 5i)
(2 + 4i ) 2
(ii)
(3 − i ) 2

3. Solve the following equations :

(i) [(x + 2y), (2x –y – 6) = (3, 2)
−2 + 3i
(ii) (x – y) + i (x + y) =
1+ i

4. Express the following complex number in the polar form

(i) −1 + i 3
(ii) −2 − i 12

7
5. Find the trigonometric representation of
(i) sin  – i cos 
(ii) 1 + cos  + i sin 

6. Show that | cos  + i sin  | = 1

5+i
7. Express the complex number in the form (x + iy) where x and y are real. Find its modulus
2+i
and amplitude.

8
 LESSON 2

DE MOIVRE’S THEOREM

In this lesson, we shall discuss an important theorem which is used in finding the roots of a complex
number, solution of equations, expansion of trigonometric function etc. This theorem is known as De
Moivre’s Theorem.

2.1 De Moivre’s Theorem

Statement :
(i) If n is an integer, positive, or negative, then (cos  + i sin ) = cos n + i sin n.
(ii) If n is a rational number, then one of the values of (cos  + i sing )n is
Proof : Part (i) case (a) : Let n be a positive integer
By simple multiplication
(cos  + i sin ) (cos  + i sin ) = (cos  cos  – sin  sin )
+ i (sin  cos  + cos  sin )
= cos ( + ) + i sin ( + )
Again multiplying the above results by (cos  + i sin ), we have
(cos  + i sin ) (cos  + i sin ) (cos  + i sin )
= [cos ( + ) + i sin ( + )] (cos  + i sin )
= cos ( +  + ) + i sin ( +  + )
This process can be continued to any number of factors so that
(coa  + i sin ) (cos  + i sin ) (cos  + i sin ) ...... to n factors
= cos ( +  +  + ....... to n terms)
+ i sin ( +  +  + ...... to n terms)
In this expression, put  =  =  = ........ = .
So that we have (cos  + i sin )n = cos n  + i sin n .
Case (b) : Let n be a negative integer. Suppose n = – m where is a positive integer.
Then (cos  i sin )n = (cos  + i sin )–m
1
= (by the law of indices)
(cos  + i sin )m
1
= (by case (a))
(cos m  + i sin m )
(cos m  − i sin m )
=
(cos m  + i sin m )(cos m  − sin m )
(cos m  − i sin m )
= = cos m  − i sin m 
(cos 2 m  + sin 2 m )
= cos(−m) + i sin (−m)
= con n  + i sin n 

9
 Thus (cos  + i sin )n = cos n  + i sin n  for integral values of n, whether positive or negative
p
Proof of (ii) : If n is (a rotational number, positive or negative), let n = (q  0)
q

We shall take q as a positive integer and p as a positive or a negative integer.

q
      
We haver  cos + i sin  = cos  q  + i sin  q 
 q q  q  q
= cos  + i sin 
  
i.e., qth power of  cos + i sin  is cos  + i sin 
 q q
  
i.e.,  cos + i sin  is one of the qth roots of (cos  + i sin )
 q q
1
 
i.e., cos + i sin is one of the values of (cos  + i sin ) q
q q
Raising each of these quantities of pth power, we have the results that one of the values of
p
  
(cos  + i sin ) p/q
is  cos + i sin 
 q q
p p
i.e., cos  = i sin 
q q
Hence cos n  + i sin  is one of the values of (cos  + i sin )n if n is a rational number positive or
negative.
Corollary 1. cos n  – i sin n  is one of the values of (cos  – i sin )n for all rational values of n.
Illustratives : With the help of the De Moivre’s Theorems, we see that
1. (cos  + i sin )4 = cos 4  + i sin 4 
7
   7 7
2.  cos + i sin  = cos + i sin
 2 2 2 2
3. (cos  + i sin )–7 = cos (–7) + i sin (–7)
= cos 7 – i sin 7
4. Since (cos  + i sin ) (cos  – i sin )
= (cos2  + sin2 ) + i (sin  cos  – sin  cos ) = 1 + i0 = 1
1
 cos  − i sin  =
cos  + i sin 

   
5. sin  + i cos  = cos  −   + i sin  −  
2  2 
   
6. (sin  + i cos )n = cos n  −   + i sin n  −  
2  2 

10
 3
    
e.g., (sin  + i cos )3 = cos  −   + i sin  −   
 2  2 
   
= cos 3 −   + i sin 3 − 0 
 2   2 
 3   3 
= cos  − 3  + i sin  − 3 
 2   2 
= – sin 3 – i cos 3
Remarks : (i) As any complex number x + iy can be put in the form r (cos  + i sin ), therefore, De
Moivre’s Theorem may be written as
( x + iy)n = r n (cos  + i sin )n
= r n (cos n  + i sin n )
where r is modulus of the complex number and  is the amplitude.
(ii) The product and quotient of any two complex numbers can be obtained by using this method.
Let any two complex number be given by
r1 (cos 1 + i sin 1) and r2 (cos 2 + i sin 2),
then their product is
r1r2[cos (1 + 2) + i sin (1 + 2)]
i.e., the product of two complex numbers is a complex number whose modulus is the product of the
modulli and whose amplitude is the sum of the amplitudes two complex numbers.
The quotient of these two complex numbers, r1(cos 1 + i sin 1) and r2(cos 2 + i sin 2) is
r1 (cos 1 + i sin 1 ) r
= 1 (cos 1 + i sin 1 ) (cos 2 − i sin 3 )
r2 (cos 2 + i sin 2 ) r2
r1
= [cos(1 − 2 ) + i sin (1 − 2 )]
r2
Solved Examples
(cos 2 + i sin 2) 4 (cos  + i sin )5
Example 1. Simplify .
(cos 6 + i sin 6)3 (cos 4 − i sin 4) 7
Solution. By De Moivre’s Theorem, we have
cos 2 + i sin 2 = (cos  + i sin )2
cos  – i sin  = (cos  + i sin )–1
cos 6 + i sin 6 = (cos  + i sin )–1
cos 4 – i sin 4 = (cos  + i sin )–4
(cos 2 + i sin 2) 4 (cos  − i sin )5
Now,
(cos 6 + i sin 6)3 (cos 4 − i sin 4)7
(cos  + i sin )8 (cos  + sin ) −5
=
(cos  + i sin )18 (cos  + i sin ) –28
= (cos  + i sin )8 – 5 –18 + 28

11
 = (cos  + i sin )36 – 23
= cos  + i sin )13 = cos 13 + i sin 13.
(cos 2 + i sin 2)3 (cos  − i sin )5
Example 2. Simplify .
(cos 3 − i sin 3)8 (cos 5 + i sin 5) 2
Solution. By De Moiver’s Theorem, we have
(cos 2 + i sin 2)3 (cos  − i sin )5 (cos  + i sin )6 (cos  + sin ) −5
=
(cos 3 − i sin 3)8 (cos 5 + i sin 5) 2 (cos  + i sin ) −24 (cos  + i sin ) –10
= (cos  + i sin )6 – 5 + 24 – 10
= (cos  + i sin )15
= cos 15 + i sin 15
−1 + 3i
Example 3. Obtain the quotient of by using De Moivre’s Theorem.
1+ i
 1 3i   2 2 
2 − +  2  cos + i sin 
−1 + 3i  2 2 
= 
3 3 
Solution. =
1+ i  1 1    
2 + i 2  cos + i sin 
 2 2   4 4
  2    2   
= 2 cos  −  + i sin  − 
  3 4  3 4 
 5 5 
= 2 cos + i sin  .
 12 12 
Example 4. Obtain ( 3 + i) .


  3 1     
4
Solution. Here ( 3 + i) =  2  +   = 2  cos + i sin  
  2 2    6 6 

 4 4 
= 16 cos + i sin 
 16 6
 2 2 
= 16 cos + i sin 
 3 3
 1 3i 
= 16  − + 
 2 2 

= −8 + 8 3  i
Example 5. Prove that
n
 1 + sin  + i cos    n   n 
  = cos  − n  + i sin  − n 
 1 + sin  − i cos    2   2 
Solution. Let
1 + sin  + i cos  = r (cos  + i sin ).

12
Then equating real and imaginary part, we have
1 + sin  = r cos , cos  = r sin 
 
sin  −  
cos  2 
Hence, tan  = =
1 + sin   
1 + cos  −  
2 
sin  
= , where  = − 
1 + cos  2
 
2 sin cos
= 2 2 = tan 
 2
2 cos 2
2
  
 = = −
2 4 2
Also 1 + sin  – i cos  = r(cos  – sin )
 The given expression
r n (cos  + 1sin ) n (cos  + i sin ) n
= =
r n (cos  − i sin ) n (cos  + i sin ) − n
= (cos  + i sin )n (cos  + i sin )n
= (cos  + i sin )2n
= (cos 2n + i sin 2n)
     
= cos 2n  −  + i sin 2n  − 
 4 2  4 2
 n   n 
= cos  − n  i sin  − n 
 2   2 
Example 6. Prove that :
m b
(a + bi)2m/n + (a – bi)m/n = 2(a + b )
2 2 m/2 n
cos  tan −1 
n a
Solution. Here put a + bi = r (cos  + i sin )
Then a = r cos  and b = r sin 
 a2 + b2 = r2(cos2  + sin2 )  r2 = a2 + b2
b
and tan  =
a
b
Hence r= a 2 + b  and  = tan −1
a
Also a – bi = r cos  – i sin ).
Thus the given expression can be written as
(a + bi)m/n + (a – bi)m/n = [r (cos  + i sin )]m/n + [r (cos  − i sin )]m/n
 m m  m/n  m m 
= r m /n cos + i sin  + r cos − i sin
 n n   n n 

13
  m 
= 2r m /n  cos 
 n 
m b
= 2(a 2 + b2 )m/2n cos  tan −1 
n a
Example 7. If x = cos  + i sin , then show that
1
x m + m = 2 cos m
x
1
and xm − = 2 i sin m.
xm
Solution. Here xm = (cos  + i sin ) = cos m + i sin m,
1
and = (cos  + i sin )–m
xm
= cos m – i sin m.
1
Hence xm + = 2 cos m
xm
1
Also, xm − = 2 i sin m
xm
1 1
Example 8. If x + = 2 cos , show that x n + n = 2 cos n .
x x
Solution. Here, it is given that
1
x+ = 2 cos 
x
or x2 – 2x cos  + 1 = 0
or x2 – 2x cos  + (cos2  + sin2 ) = 0
or (x2 – 2x cos  + cos2 ) = –sin2 
or (x – cos )2 = ( i sin )2
 x – cos  = ± i sin .
i.e., x= cos  ± i sin 
Let x= cos  + i sin  (taking the + ve sign).
1 1
Then, = = cos  − i sin 
x cos  + i sin 
1
 xn + = (cos  + i sin )n + (cos  – sin )n
xn
= (cos n + i sin n) + (cos n – i sin n)
= 2 cos n
1
If x cos  – i sin , then = cos  + i sin . Then also the result can be proved.
x
Example 9. If sin  + sin  + sin  = 0 and cos  + cos  + cos  = 0. Prove that
cos 3 + cos 3 + cos 3 = 3 cos ( +  + )

14
 sin 3 + sin 3 + sin 3 = 3 sin ( +  + )
Solution. Let x = cos  + i sin 
y = cos  + i sin 
z = cos  + i sin 
Then x + y + z = (cos  + cos  + cos ) + i (sin  + sin  + sin )
= 0 + i0 = 0
 x3 + y3 + z3 = 3xyz
(cos  + i sin )3 + (cos  + i sin )3 + (cos  + i sin )3
= 3(cos  + i sin ) (cos  + i sin ) (cos  + i sin )
 (cos 3 + i sin 3) + (cos 3 + i sin 3) + (cos 3 + i sin 3)
= 3[cos( +  + ) + i sin ( +  + )]
 (cos 3 + cos 3 + cos 3) + i (sin 3 + sin 3 + sin 3)
= 3 cos( +  + ) + 3 i sin ( +  + )
Equating real and imaginary parts, we get
cos 3 + cos 3 + cos 3 = 3 cos( +  + )
and sin 3 + sin 3 + sin 3 = 3 sin ( +  + )
Hence the results.
Example 10. Show that for an integer n
 (n)
(1 + cos  + i sin ) n (1 + cos  − i sin ) n = 2n + 1 cos n cos
2 2
Solution. We have (1 + cos  + i sin )n + (1 + cos  – i sin )n
n n
       
=  2 cos 2 + 2i sin cos  +  2 cos 2 − 2i sin cos 
 2 2 2  2 2 2
n n
     
= 2 cos cos + i sin  + 2n cos n cos − i sin 
n n
2 2 2 2 2 2
n
  n n        
= 2n cosn  cos + i sin  + 2n cosn cos  −  + i sin  −  
2  2 2  2   2  2 
 n n  n   n  
= 2n cos n  cos + i sin + cos  −  + i sin  −  
2 2 2  2   2 
 n n n n 
= 2n cosn  cos + i sin + cos − i sin 
2 2 2 2 2
  n  
= 2n cos n  2 cos   
2  2 
  n 
= 2n + 1 cosn cos  
2  2 

15
Example 11. If sin  + sin  + sin  = 0 = cos  + cos  + cos . Prove that
(i) cos( + ) + cos( + ) + cos( + ) = 0
(ii) sin( + ) + sin( + ) + sin( + ) = 0
Solution. Let x = cos  + i sin , y = cos  + i sin , z = cos  + i sin 
so that x + y + z = (cos  + cos  + cos ) + i(sin  + sin  + sin )
= 0 + i0 = 0
1 1 1
Also = cos  − i sin , = cos  − i sin , = cos  − i sin 
x y z
1 1 1
 + + = (cos  + cos  + cos  ) − i (sin  + sin  + sin  )
x y z
= 0 – i0 = 0
 yz + zx + xy = 0
 xy + yz + zx = 0
 (cos  + i sin ) (cos  + i sin ) + (cos  + i sin ) (cos  + i sin )
+ (cos  + i sin ) (cos  + i sin ) = 0
 [cos ( + ) + i sin ( + )] + [cos ( + ) + i sin ( + )]
+ [cos ( + ) + i sin ( + )] = 0
 [cos ( + ) + cos ( + )] + cos ( + )] + i [sin ( + ) + sin ( + )]
+ sin ( + )] = 0
Equating real and imaginary parts on both sides, we get
cos ( + ) + cos ( + ) + cos ( + ) = 0
sin ( + ) + sin ( + ) + sin ( + ) = 0
Hence the result.
Example 12. If Z = cos  + i sin , show that
Z 2n − 1
= i tan n, n being an integer.
Z 2n + 1
Z 2n − 1 Z n − Z −n
Solution. = (Dividing numerator and denominator by Zn)
Z 2n + 1 Z n + Z −n
(cos  + i sin ) n − (cos  + i sin ) − n
=
(cos  + i sin ) n + (cos  + i sin ) − n
(cos n + i sin n) − (cos n − i sin n)
=
(cos n + i sin n) + (cos n − i sin n)
2 i sin n
= = i tan n proved
2 cos n
Example 13. If (cos  + i sin ) (cos 2 + i sin 2) ...... (cos n + i sin n) = 1 prove that
4k 
= , where k is an integer.
n(n + 1)

16
Solution. We have 1 = cos 0 + i sin 0
= cos 2k + i sin 2 k ......, where k is an integer. ...(1)
Also (cos  + i sin ) (cos 2 + i sin 2) ....... (cos n + i sin n) = 1
L.H.S. = cos ( + 2 + ...... + n) + i sin ( + 2 + ...... + n)
n(n + 1) n(n + 1)  n(n + 1) 
= cos  + i sin   1 + 2 + 3 + ...... + n =
2 2  2 
n(n + 1) n(n + 1)
Thus cos  + i sin  =1 = cos 2 k + i sin 2k by (1)
2 2
n( n + 1) 4k 
This gives us  = 2k    = Hence the result.
2 n(n + 1)
Example 14. If (1 + x)n = nC0 + nC1x + nC2 x2 + + nCx xn , n is a positive integer. Prove that
n−2 n
(i) n
C0 + nC4 + nC8 + = 2n − 2 + 2 cos
2 4
n
(ii) n
C0 − nC2 + nC4 − nC 6 + = 2n/2 cos
4
n
(iii) n
C1 − nC3 + nC5 − nC 7 − = 2n/2 sin
4
Solution. We are given that
(1 + x)n = n
C0 + nC1x + nC2 x2 + nC 3 x3 + nC4 x 4 +
Putting x = 1, –1, i, – i in succession, we get
2n = n
C0 + nC1 + nC2 + nC 3 + nC4 + n
Cn ...(1)
0= n
C0 – nC1 + nC2 – nC 3 + nC4 – ...(2)
(1 + i)n = n
C0 – i nC1 − nC2 – nC 3i + nC4 + ...(3)
= (nC0 – nC2 + nC4 ) + i(n C 1 − nC3 + nC5 ) ...(4)
(1 – i)n = n
C0 – i nC1 − nC2 + i nC 3 + nC4 ...(5)
Adding (1), (2), (4) and (5) we get
2n + (1 + i)n + (1 − i)n = 4(nC0 + nC4 + nC8 + .......) ...(6)
n
    
Now (1 + i)n =  2  cos + i sin  
  4 4 
 n n 
= 2n/2  cos + i sin  ...(7)
 4 4 
 n n 
Similarly, (1– i)n = 2n/2  cos − i sin  ...(8)
 4 4 
Adding (7) and (8) and substituting in (6), we get

4(nC0 + nC4 + nC8 + ) = 2n + 2.2n/2 cos n
4

17
 n 
 + 1 n
 
= 2 + n
2 2
cos
4
n−2 n
cos
n−2
 n
C0 + C4 + C8 +
n n
= 2 + 2 2 4 Proved
(ii) From (4) and (7)
(nC0 + nC4 + nC8 + ) + i (nC1 − nC3 + nC5 ......)
 n n 
= 2n/2  cos + i sin 
 4 4 
Equating real and imaginary parts, we get
n
C0 − nC2 + nC4 = 2n/2 cos n/4
n
C1 − nC3 + nC5 = 2n / 2 sin n / 4

EXERCISE 2.1
(cos  + i sin ) (cos 2 − i sin 2)9
1. Simplify
(cos 2 i sin 2)9 (cos  − i sin )8
10 10
     
 cos + i sin  +  cos − i sin 
2. Evaluate:    
15 15 15 15
6
  
 cos + i sin 
 3 3
(cos  + i sin ) 4
3. Use De’ Moivre’s Theorem to simplify
(sin  + i cos )3
   
4. Prove that (sin x – i cos x)n = cos n  − x  − i sin n  − x 
2  2 
5. Obtain the value of (3 + 4i)3 with the help of De Moivre’s Theorem.
6. If n is a positive integer, show that :
n
( ) ( )
n n
3 + i + 3 − i = 2n − 1 cos .
6
7. Prove that :
[cos  + cos ) + i (sin  + sin )]n + [cos  + cos ) – i sin  + sin )]n
− n( + )
= 2n + 1 cos n   cos
 2  2
 
 Hint : (cos  + cos  + i(sin  + sin ) 
 
 + −  +  −  
 = 2 cos cos + i  2 sin cos 
 2 2  2 2 
 −  +  +  
 = 2 cos  cos + i sin  
 2  2 2  

18
8. Show that
[(cos  – cos ) + i (sin  – sin )]n + [(cos  – cos ) – i(sin  – sin )]n
− n++
= 2n + 1 sin n   cos  
 2   2 
 Hint : (cos  − cos  + i (sin  − sin ) 
 + − − + 
 = 2 sin sin + 2i sin cos 
 2 2 2 2 
 −  +  +  
 = 2 sin − sin + i cos 
 2  2 2  
 
 −   +     +   
= 2 sin  +  +  + 
2  2   
cos i sin
 2 2  2

EXERCISE 2.2
1 1
1. If 2cos  = x + , 2 cos  = y +
x y
Show that
1
xm y n + m n
= 2 cos (m + n)
x y
xm y n
+ = 2 cos (m – n)
y n xm
where m and n are integers. What happens if m and n are rational numbers?
2. If x = cos  + i sin , y = cos  + i sin  and z = cos  + i sin  such athat x + y + z = 0 then
prove that :
1 1 1
+ + = 0
x y z
1
and xyz + = 2 cos ( +  + )
xyz
3. Let the complex numbers x, y, z be given respectively by cos a + i sin a, cos b + i sin b and cos c
+ i sin c, then prove that for any integers p, q, r.
1
x p y q z r + p q r = 2 cos (pa + qb + rc)
x y z
4. Let x1 + iy1, x2 + iy2 ......, xn + iyn be any n complex number and A + iB be some other complex
number, such that
(x1 + iy1) (x2 + iy2) ...... ( xn + iyn ) = A + iB
Then show thats
y y y B
(i) tan −1 1 tan −1 2 + ...... + tan −1 n = tan −1
x1 x2 xn A
(ii) ( x12 + y12 ) ( x22 + y22 ) ( xn2 + yn2 ) = A2 + B2

19
5. Expand (1 + i)n in two different ways and find the sum of the two series,
(1 − nC2 + nC4 − nC6 + ) and nC1 − nC3 + nC5 )
[Hint : First expand (1 + i)nby Boinomial theorem and then expand it by help of De Moivre’s
Theorem.
6. If sin  + sin  + sin  = cos  + cos  + cos  = 0. Prove that
(i) sin 2 + sin 2 + sin 2 = 0
(ii) cos 2 + cos 2 + cos 2 = 0
(iii) cos2  + cos2  + cos2  = 3/2

20
 LESSON 3

APPLICATIONS OF DE MOIVRE’S THEOREM

In this lesson, we shall discuss certain applications and uses of De Moivre’s Theorem. As mentioned
earlier, it helps in finding the roots of a number, solution of some equations and expressions of
trigonometric functions, etc.
We have seen that De Moivre’s Theorem can be stated as
q
    
cos q + i sin q =  cos + i sin  , where q is a positive integer.
q q  q q
q
  
i.e., cos  + i sin  =  cos + i sin 
 q q
 
 cos + i sin is a qth root of cos  + i sin 
q q
 
i.e., cos + i sin is one of the values of (cos  + i sin )1/q
q q
Now the questions arises :
‘What about the other values of (cos  + i sin )n for rational values of n ?
We shall now determine all the values of (cos  + i sin )1/q, where  is a positive integer.
We know that
(cos  + i sin ) = cos (2r + ) + i sin (2r + )
 (cos  + i sin )1/q = [cos (2r + ) + i sin (2r + )]1/q where q is any integer.
Now by De Moiver’s Theorem one of the values of (cos  + i sin )1/q is
2r +  2r + 
cos + i sin
q q
By giving ‘r’ various values, we get different values of (cos  + i sin )1/q.
Let us give r in succession the values 0, 1, 2, ...... q – 1 and we see that each of the following
quantities
 
cos + i sin
q q
2 +  2 + 
cos + i sin
q q
4 +  4 + 
cos + i sin
q q
....... ....... ....... ........ ......

2(q − 1)  +  2(q − 1)  + 
cos + i sin
q q

21
is equal to one of the values of (cos  + i sin )1/q. We note here that if we give to ‘r’ integral values
greater than (q – 1) viz., the values q, (q + 1), (q + 2), ...... etc. then we do not obtain any new values of
(cos  + i sin )1/q.
For example, for r = q we obtain the value
2q +  2q + 
cos + i sin
q q
   
i.e., cos  2 +  + i sin  2 + 
 q  q
 
i.e., cos + i sin .
q q
which is the same values as obtained by putting r = 0.
Thus the greatest values we need to assign to r is q – 1 for the values q, q + 1, q + 2 ...... will be
found to give the same result as the values, 0, 1, 2, ...... q – 1, for ‘r’.
Also no two of the quantities obtain by giving ‘r’, the values 0, 1, 2, ...... q – 1 will be the same for
all the angles involved there in differ from one another by less than 2 and no two angles differing by less
than 2 have their consines the same and also sines the same.
(2r + ) (2r + )
Thus the expression cos + i sin , where r = 0, 1, 2, ...... (q – 1) gives q and only
q q
q different values of (cos  + i sin )1/q. We may extend the above result to (cos  +
i sin )1/q where p and q are integers, q being taken as positive.
Thus (cos  + i sin )1/q has q and only q different values and the these are given by
(2r + ) p (2r + ) p
cos + i ain , where r = 0,1, 2, ...... (q − 1).
q q
Then, the  values are
p p
cos + i sin , when r = 0
q q
(2 + ) p (2 + ) p
cos + i sin , where r = 1
q q
(4 + ) p (4 + ) p
cos + i sin , where r = 2
q q
.......................................................................
[2(q − 1) + ] p [2(q − 1) + ] p
cos + i sin , when r = q − 1.
q q
Note. We may note here that the ‘q’ distinct values of (cos  + i sin )p/q can be arranged in a
geometrical progression, for if
2p 2p p p
 = cos + i sin and a = cos + i sin .
q q q q
then the q values of (cos  + i sin )p/q can be arranged as a, a, a2, ...... aq – 1 which is a G.P. The
student should note that

22
 r
p(2r + ) p(2r + )  2p 2p   p p 
cos + i sin =  cos + i sin   cos + i sin 
q q  q q   q q 

3.1 Extraction of Roots of Complex Quantity

We know that
x + iy = r (cos  + i sin ),
y
where r= x 2 + y 2 and  = tan −1 .
x
Hence (x + iy) = r [cos  + i sin ]
1/q 1/q 1/q

= r1/q [cos(2n + ) + i sin (2n + )]1 /q

 (2n + ) (2n + ) 
= r1/q cos + i sin 
 q q 
when ‘n’ is given in succession the values, 0, 1, 2, ......, q – 1.
This gives the required ‘q’th roots of x + iy.
It may be noted that even if we write x + iy = r [cos (2k  + ) + i sin (2k  + )], k being an integer,
the same values are obtained.
Example 1. Find the values of (1 + i)1/6.
 1 
Solution. We know that cos = sin .
4 2 4
 1 1 
 We can write (1 + i) = 2 +i 
 2 2
1
  
= 2 2 cos + i sin 
 4 4
1
     
= 2 2 cos  2n +  + i sin  2n +  
  4  4 
1
  2n + /4   2n + /4  
 + i sin 
12
(1 + i)1/6 = 2 cos  
  6   6 
when n is given the six values 0, 1, 2, 3, 4, and 5.
The six values of (1 + i)1/6, are therefore
1 1
    9 9 
212 cos + i sin  , 212 cos + i sin  ,
 24 24   24 24 
1 1
 17  17   12  25 25 
cos 24 + i sin 24  , 2 cos 24 + i sin 24  ,
2 12
   
1 1
 33 33  12  41 41 
212 cos + i sin  , 2 cos + i sin
 24 24   24 24 

23
 It may be observed that the 4th value, viz.,
1
 25   25 
212 cos   + i sin  
 24   24 
1
     
cos   + 24  + i sin   + 24  
= 212
    
1
  
= −212  cos + i sin ,
 24 24 
since cos ( + ) = – cos  and sin ( + ) = – sin . Similarly the 5th and 6th values can simplified. The
six value of (1 + i)1/6 can then be written as
1
   
 212  cos + i sin ,
 24 24 
1
 9 9 
 2 12
 cos + i sin ,
 24 24 
1
 17  17  
 212  cos + i sin ,
 24 24 
or simply as
1
 r r 
 2 12
 cos + i sin ,
 24 24 
where r = 1, 9 and 17.
3/4
1 3
Example 2. Find all the values of  + i  and show that the continued product of all the
2 2 
values is 1.
1 3
Solution. Let +i = r (cos  + i sin ),
2 2
1
then = r cos 
2
3
and = r sin 
2
1 3
so that r2 = + =1
4 4

and tan  = 3 = tan
3

or r = 1 and  =
3
1
3/4
1 3      4
3
Now  + i =  cos + i sin  
2 2   3 3  

24
 = (cos  + i sin )1/4
= [cos (2k + ) + i sin (2k + )]1/4
 
= cos (2k + 1) + i sin (2k + 1) , k = 0,1, 2, 3
4 4
Hence the required four values are
 
when k= 0 cos + i sin ,
4 4
3 3
k= 1 + i sin ,
cos
4 4
5 5
k = 2 cos + i sin ,
4 4
7 7
k = 3 cos + i sin ,
4 4
And the continued product of four values
   3 3   5    7 7 
=  cos + i sin   cos + i sin   cos + i sin   cos + i sin 
 4 4 4 4  4 4  4 4 
  3 5 7    3 5 7 
= cos  + + +  + i sin  + + + 
4 4 4 4  4 4 4 4 
= cos 4 + i sin 4
= 1 + i.0 [ cos 4 = 1, sin 4 = 0]
= 1.
Example 3. Find the seven 7th roots of unity and prove that the sum of their nth powers always
vanishes unless n be a multiple of 7, n being integer and that the sum is 7 when n is multiple of 7.
Solution. We have to find the 7 values of (1)1/7,
We know that cos 0 = 1 and sin 0 = 0
We can write 1 = cos 0 + i sin 0
= cos 2n + i sin 2n
2n 2n
 (1)1/7 = cos + i sin
7 7
where n = 0, 1, 2, 3, 4, 5, 6.
The 7 roots are, therefore
(cos 0 + i sin 0) i.e., 1,
2 2
cos + i sin =  (say)
7 7
2
 4 4   2 2 
 cos + i sin  i.e.,  cos + i sin  =
2
 7 7   7 7 
and similarly for the other roots.
Thus the seven 7th roots of unity are
1, , 2, 3, 4, 5, 6

25
  2 2 
where  =  cos + i sin .
 7 7 
The sum of the nth power of the roots
= (1)n +s ()n + (2)n + (3)n + (4)n + (5)n + (6)n
= 1 + n + 2n + 3n + 4n + 5n +6n
(which is geometric progression)
1 − ( )
n 7
1 − ( )
7 n
1 − (1 )
n
1−1
= = = = =0
1−  n
1−  n
1−  n
1 − n
Since 7 = 1
Also n  1, since n is not a multiple of 7.
If ‘n’ is a multiple of 7, i.e., if n = 7 m say where m is an integer, then
n 7m
 2 2   2 2 
n =  cos + i sin  =  cos + i sin 
 7 7   7 7 
= cos 2m + i sin 2m
= 1.
Hence when n is a multiple of 7, the sum of the nth power of the 7th roots of unity is
= 1 + 1 + 1 +1 +1 +1 + 1 = 7
Example 4. Prove that n x + iy + n x − iy has ‘n’ real values.

Solution. Let x = r cos  and y = r sin . Then

x + iy = r[cos (2k + ) + i sin (2k + )],
where k is an integer.
 (x + iy)1/n + (x + iy)1/n
= r1/n {[cos(2k  + ) + i sin(2k  + )]1/n + [cos(2k  + ) − i sin(2k  + )]1/n }
2k  + 
= 2r1/n cos , where k = 0, 1, 2, ...... (n – 1)
n
y
Here r2 = x 2 + y 2 , and tan  = .
x
y
2k  + tan −1
Hence (x + iy)1/n + (x – iy)1/n = 2( x 2 + y 2 )1/ 2n cos x
n
where k = 0, 1, 2, ...... (n – 1).
This is real and has n distinct values.

3.2 Solution of Equations

De Moivre’s Theorem can be used to obtain solutions of certain types of equations. This will be
illustrated by a few solved examples. Conversely De Moivre’s Theorem can be used to find the equation
whose roots are given as trigonometical functions.

26
 Example 5. Solve the equation zn = 1, where n is a positive integer.
Solution. Here we write
zn = 1 = cos (2k + 0) + i sin (2k + 0)
 2k  2k  
 z = (1)1/n =  cos + i sin 
 n n 
where k = 0, 1, 2, ..... (n – 1)
Hence the nth roots of unity are
 2 2 
(cos 0 + i sin 0), i.e., 1,  cos + i sin 
 n n 
2
 4 4   2 2 
 cos + i sin  =  cos n + i sin n 
 n n   
...............................................................
...............................................................
n −1
 2(n − 1) 2(n − 1)   2 2 
 cos + i sin  =  cos + i sin 
 n n   n n 
2 2
It may be noted that if we write  = cos + i sin ,
n n
then the nth root of ‘1’ can be written as 1, , 2, ...... n – 1. Thus we get an interesting result that the ‘n’
nth roots of unity form a G.P. Also it is seen that each of these roots can be expressed as a power of
another.
Exmple 6. Solve the equation z9 – z5 + z4 – 1 = 0.
Solution. Here (z9 – z5 + z4 – 1) = 0
 z5 (z4 – 1) + (z4 – 1) = 0
i.e.,  (z5 – 1) (z4 – 1) = 0
 Either (i) z4 = 1 or (ii) z5 = – 1
since 1 = cos 2 n + i sin 2 n
and – 1 = cos  + i sin 
= cos (2n + ) + i sin (2n + )
= cos (2n + 1)  + i sin (2n + 1) .
From (i), we have
2n 2n
(cos 2n + i sin 2n)1/4 = cos + i sin
4 4
n n
z = cos + i sin , where n = 0,1, 2, 3,
2 2
and form (ii), we have
z = (–1)1/5 = [cos (2n + 1) p + i sin (2n + 1) ]1/5
 (2n + 1) (2n + 1) 
= cos + i sin  , when n = 0,1, 2, 3, 4.
 5 5 
These (4 + 5) = 9 values give the complete solution of the given equations.

27
Example 7. Solve the expression z5 + 1 = 0.
Solution. z5 = –1 = [cos (2r + 1)  + i sin (2r + 1)]
where r is an integer.
 (2r +1) (2r + 1) 
 z = cos + i sin  , where r = 0,1, 2, 3, 4.
 5 5 
 
When r = 0, z = cos + i sin .
5 5
3 3
When r = 1, z = cos + i sin
5 5
And for r = 2, z = cos  + i sin  = –1.
For r = 3, the value of z is
7 7
cos + i sin
5 5
 3   3 
= cos 2 −  + i sin 2 − 
 5  5
 3   3 
= cos  −  + i sin  − 
 5  5
3 3
= cos − i sin
5 5
For r = 4, the value of z is
9 9
= cos + i sin
5 5
   
= cos  2 −  + i sin  2 − 
 5  5
   
= cos  −  + i sin  − 
 5  5
 
= cos − i sin
5 5
5
Thus the 5 roots of the equation z + 1 = 0 are
  3 3
–1, cos  i sin , cos  i sin .
5 5 5 5
6 3
Example 8. Find the roots of the equation z + z + 1 = 0
Solution. On solving the equation a quadratic in z3, we find
−1  3 1 3
z3 = =−  i
2 2 2
 2
= cos  i sin
3 3
 2   2 
= cos  2r  +   i sin  2r  + 
 3   3 

28
  2  
= cos (3r + 1)  i sin (3r + 1) 
 3 3 
1
 2(3r + 1) (3r + 1)  3
 z = cos  i sin 
 3 3 
 2(3r + 1) (3r + 1) 
= cos  i sin  , r = 0,1, 2.
 9 9 
On giving r the values, 0, 1, 2, the six roots of the given equation can be written as
2 2 8 8 14 14
cos  i sin , cos + i sin , cos  i sin
9 9 9 9 9 9
14 14  4   4 
Now cos  i sin = cos  2 −   i sin   − 
9 9  9   9 
 4   4  4 4
=  cos −  + i sin  −  = cos  i sin
 9   9  p 9
2 2 4 4 8 8
 The roots are  i sin
cos , cos  i sin , cos  i sin
9 9 9 9 9 9
Example 9. Solve the equation z + z + z + z + z + 1 = 0 multiply through by z – 1 we get z6 – 1 =
5 4 3 2

0.
Solution. z5 + z4 + z3 + z2 + z + 1 = 0
Now z6 – 1 = 0
 z6 = 1 = cos 2k + i sin 2k, where k is an integer
cos 2k  2k 
 z= + i sin
6 6
k k
= cos + i sin , where k = 0,1, 2, 3, 4, 5.
3 3
  2 2 2
 z = cos 0 + i sin 0, cos + i sin ; cos + i sin + i sin ; cos  + i sin 
3 3 3 3 3
4 4 5 5
cos + i sin ; cos + i sin ;
3 3 3 3
1 3 1 3 1 3 1 3
= 1, +i ,− +i , − 1, − − i , −i
2 2 2 2 2 2 2 2
1 3
= 1,  +i
2 2
The root 1 corresponds to the factor (z – 1)
1 3
The remaining roots –1,  i are the roots of the given equation
2 2

29
 EXERCISE
1. Find all the values of
(i) (1 + i)1/3
(ii) (–1 + i 3)1 4
(iii) (–1)1/6
(iv) (1 + i)2/3
2. Solve the equation:
(i) z7 = 1
(ii) z7 + z = 0
3. Find the nth roots of unity and show that from a series in G.P. whose sum is zero.
Also prove that the sum of their pth power always vanishes unless p be a multiple of n, p being
an integer and that then the sum is n.
4. Solve the equation : z7 + z4 + z3 + 1 = 0.
[Hint : z7 + z4 + z3 + 1 = (z4 + 1) (z3 + 1) = 0]
5. Use De Moivre’s Theorem to solve : z10 + 11z5 + 1 = 0.
[Hint : Put z5 = y]
6. Solve the equation : z4 – z3 + z2 – z + 1 = 0.
[Hint : Multiply the equation by (z + 1)]

3.3 Expansion of sin n  and cos n 

De Moivre’s Theorem may be applied to express trigonometric functions of multiple angles in terms
of the trigonometric function of the angles. The theorem can be employed to express powers of sines and
cosines of angle in terms of trigonometric functions of multiple angles.
First we shall learn to express cos n, sin n, tan n in terms of cos , sin , tan , etc.
For example, by De Moivre’s Theorem
cos 2 + i sin 2 = (cos  + i sin )2 = cos2  + 2 i sin  cos  + i2 sin2 
= (cos2  – sin2 ) + 2i cos  sin )
Equating the real and imaginary parts, we get
cos 2 = cos2  – sin2 ,
sin 2 = 2 cos  sin 
Generally
(cos n + i sin n) = (cos  + i sin )n, n being a positive integer.
By the Binomial Theorem for a positive integral index, we have
n(n − 1)
(cos  + i sin )n = cosn  + n cosn − 1  (i sin ) + cosn − 2  (i sin )2
2
(n + 1) (n − 2)
+ cosn − 3  (i sin )3 ...... + (i sin ) n
3
But i2 = –1, i4 = 1, i3 = – 1 etc.
Hence we can write

30
  n(n − 1) n−2
cos n + i sin n = cos n  − cos  sin  
 2
n(n − 1) (n − 2) ( n − 3) 
+ cos n − 4  sin 4
 + ......
4 
 n(n − 1)(n − 2) 
+ i  n cos n − 1  sin  − cos n −3  sin 3  + 
 3 
By equating the real and imaginary parts, we get
n(n − 1)
cos n = cosn  = cosn − 2  sin 2 
2
n(n – 1) (n − 2) (n − 3)
+ cosn − 4  sin 4  +
4
n(n − 1)(n − 2)
sin n = n cosn − 1  sin  − cosn − 3  sin 3  +
3
By using the binomial coefficients, we can also express these results as follows
cos n = cosn  − nC2 cosn − 2  sin 2  + nC4 cosn − 4  sin 4  + ......
= cosn [1 − nC2 tan 2  + nC4 tan 4  + ......] ...(i)
sin n = n
C1 cosn − 1  sin  + nC3 cosn − 3  sin3  + ......
= cos n[n C1 tan  − nC3 tan3  + ......] ...(ii)
Divinding (ii) by (i), we have
n
C1 tan  − n C3 tan 3  + n C5 tan 5  −
tan n =
1 − n C2 tan 2  + n C4 tan 4 q −
We have thus obtained the values of cos n, sin n, tan n in terms of cos , sin , tan .
Example 10. Obtain the values of cos 5, sin 5, and tan 5.
Solution. We can either apply the formulae or proceed as below:
cos 5 + i sin 5 = (cos  + i sin )5
= cos5  + 5 cos4  (i sin ) + 10 cos3 (i2 sin2 )
+ 10 cos2  (i3 sin3 ) + 5 cos  (i4 sin4 )
+ i5 sin5 
Thus equating real and imaginary parts,
cos 5 = cos5  – 10 cos3  sin2  + 5 cos  sin4 
= cos5  – 10 cos3  (1 cos2 ) + 5 cos  (1 – cos2 )2
= 16 cos5  – 20 cos3  + 5 cos 
(Note that cos 5 has been expressed in powers of cos  only).
sin 5 5 cos 4  sin  − 10 cos 2  sin 3  + sin 5 
Also tan 5 = =
cos 5 cos 5 − 10 cos3  sin 2  + 5 cos  + sin 4 

31
 5 tan  − 10 tan 3  + tan 5 
=
1 − 10 tan 2  + 5 tan 5 
(Note that tan 5 has been expressed in powers of cos  only.
Note. We can apply a similar method to obtain
sin ( +  +  + ......), cos ( +  + ) + ......), tan ( +  +  + .......)
and thus obtain expression for the sine, cosine and tangent of the sum of any numbers of angles in terms
of trigonometric ratio of individual angles.
We have
cos ( +  +  + ......) + i sin ( +  +  + ......)
= (cos  + i sin ) (cos  + i sin ) (cos  + i sin ) .......
Since
cos  + i sin  = cos (1 + tan )
cos  + i sin  = cos (1 + i tan )
cos  + i sin  = cos (1 + i tan ) etc.
We have
cos ( +  +  + ......) + i sin ( +  +  + ......)
= cos  cos  cos cos  ...... [(1 + i tan ) (1 + i tan )
(1 + i tan ) ......
= cos  cos  cos  ...... [1 + i(tan  + tan  + ......)
+ i2(tan  tan  + tan  tan  + ......]
+ i3 (tan  tan  tan  + ...... ) + ......]
= cos  cos  cos  ...... [1 + iS1 – S2 – iS3 + ......]
where S1 = tan  + tan  cos  ......
=  tan ,
S2 = tan  tan  + tan  tan  + ......
=  tan  tan 
S3 = tan  tan  tan  + .....
=  tan  tan  tan ,
i.e., S1 indicates the sum of the tangents of the angles, S2 indicates the sum of the products of the
tangents of the angles taken two at a time and S3 indicates the sum of the products taken three at a time
and so on.
Equating real and imaginary parts, we have
cos ( +  +  ......) = cos  cos  cos  ...... [1 – S2 + S4 – S6 + ......]
sin ( +  +  ......) = cos  cos  cos  ...... [S1 – S3 + S5 ......]
Also, by division, we get
S1 − S3 + S5 ......
tan ( +  + ) =
1 − S2 + S4 − S6 ......

32
3.4 Expansion of sin  and cos  in Powers of 
We know that for a positive integer n and an angle ,
n(n − 1)
cos n = cosn  − − cosn − 2  sin 2 
2
n(n − 1) (n − 2) (n − 3)
+ cosn − 4  sin 4  sin 4  ......
4

Now put n = , then n = and

2
( − )  sin  
cos  = cos n  − cos n − 2   
2   
4
( − ) ( − 2) ( − 3)  sin  
+ cos 2   
4   
Now if  is mode indefinitely small,  remaining constant and consequently n becoming indefinitely
sin 
large, cos  and both tend to unity, and hence in the limit when a → 0 we have

 2  4 7
cos  = 1 − + − + ......
2 4 6
Similarly by writing the value of sin n and proceeding as before, we can obtain.
 2  4 7
sin  =  − + − + ......
3 5 7
The method we have applied here is only an indication to get the values of cos  and sin  and does
not give a rigorous proof.
We shall now obtain a series for tan  in terms of  by using the series for cos  and sin .
sin 
Thus tan  =
cos 
3 5
− + −
3 5
=
2 4
1− + −
2 4
−1 
 3 5     2 4 
=   − + −   1 −  − −  
 6 120     2 24  

Expanding the second bracket by the Binomial theorem, we have

 3 5    2 4   2 2  
tan  =  − + −  1 +  − + −  + 
 6 120    2 24   2 24  

 3 5   2 4 q 4 
=   − + −  1 + − + 
 6 120  2 24 4 

33
  3 5   2 5 4 
=  − + −  1 + + q 
 6 120 2 24
  
3 3 5 5 5 5
= + − +  − +
2 6 24 12 120
3 2 5
= + +  +
3 15
3 2 5
Thus tan  =  + +  +
3 15
If powers of  more than five are neglected, i.e., for small values of , we have
3
sin  =  − ,
6
2 4
cos  = 1 − + ,
2 24
3
tan  =  + ,
3
and if power more than 2 are to be neglected then
2
sin  = , cos  = 1 − , tan  = 
2
It is understood that the angle  is expressed in radians.
Example 11. Find the value of  when
9
cos  =
100
Solution. Here cos  is nearly 1 and  is small, Hence we can take

cos  = 1 −
2
 99 1
i.e., = 1 − cos  = 1 − =
2 100 100
2 1.41
 = = = .141 radians
10 10
 180 
or  = 0.141   degrees
  
= 0.8 degrees approx.
Example 12. Find the value of sin 3º correct to theree places of decimals.
3
Solution. We have 3º = radians
180
3
3  13  1  3 
Now sin =  −   + ...
180  180  3  180 

34
 3 
= nearly =
180 60
 3.14159
 sin 3º = = = 0.0523 approx.
60 60

3.5 Expression of the Products of the Form cosm  sinn  Terms of sines or cosines of
Multiples of .
De Moivre’s Theorem can be employed to express a product of the form cosm  sinm  where m and
n are positive integers in terms of sines or consines of multiples or .
Let z = cos  + i sin 
1 1
then = = cos  − i sin 
z cos  + i sin 
1
Hence z+ = 2 cos 
z
1
z−
= 2 i sin 
z
Also from De Moivre’s Theorem, if r is an integer.
(i) zr = cos r  = i sin r 
1
(ii) = cos r  − i sin r 
zr
and, therefore, by adding and subtraction (i) and (ii), we get
1
2 cos r = z r +
zr
1
and 2i sin r = z r −
zr
We know that
m m
 1  1
(2 cos )m (2i sin )n = z +  z − 
 z  z
The right hand side of this equation is expanded of z and terms with indices equal but opposite in
1
sign are grouped together. The resulting sum is made up, if n is even, of term of the form z r + r and if n
z
1
is odd, of term of the form z r − r .
z
 1  1
Also we can express these terms  z r + r  and  z r − r  as 2 cos r and 2i sin r respectively.
 z   z 
The following examples will illustrate the method.
Example 13. Express sin6  in terms of cosines of multiples of .
Solution. Let z = cos  + i sin , so that

35
 1 1
z+ = 2 cos , z − = 2 i sin ,
z z
1 1
zr + = 2 cos r , z r − = 2 i sin r 
z r
zr
r being any positive integer.
6
 1
Now (2i sin )6 =  z −  = z 6 − C1 z 4 + 6C2 z 2 + 6C3 z 0 + 6C4 z −2 − 6C5 z −4 + 6C6 z −4
 z
 64 i6 sin6  = z 6 − 6 z 4 + 15z 2 − 20 + 15z −2 − 6z −4 + z −6
1  6 1   4 1   2 1  
 sin6  = −  z + 6  − 6  z + 4  + 15  z + 2  − 20 
64  z   z   z  
1
= − [2 cos 6 − 6.2 cos 4 + 15.2 cos 2 − 2]
64
1
= − [cos 6 − 6 cos 4 + 15 cos 2 − 20]
32
Example 14. Express sin4  cos2  in cosines of multiples of .
1
Solution. Let z = cos  + i sin  then = cos  − i sin .
z
1 1
so that z+ = 2 cos  and z − = 2 i sin 
z z
1 1
Also zr + = 2 cos r  and z r − = 2 i sin r 
z r
zr
4 2
 1  1
2 (i sin )4 (2 cos )2 =  z −   z + 
 z  z
2 2 2
 1  1  1
= z −  z −  z + 
 z  z  z
2 2
 1  1 
=  z −   z2 −  
 z   z 
 1  1
=  z2 − 2 + 2   z4 − 2 + 4 
 z  z 
2
 1   1   1 
=  z + 6  − 2  z4 + 4  +  z2 + 2  + 4
 z   z   z 
i.e., 64 sin4  cos2  = 2 cos 6 – 4 cos 4 + 2 cos 2 + 4.
1
Hence sin4  cos2  = (cos 6 − 2 cos 4 + cos 2 + 2)
32

36
 EXERCISE

1. Express sin 5 in terms of sin  only.

2. Express cos 6 in terms of cos  only.
3. Express tan 7 in terms of tan  only.
4. Find the value of sin 1º approximately.
sin  1013
5. Find the value of , where = .
 1014
6. Show that 16 sin5  = sin 5 – 5 sin 3 + 10 sin .
7. Express sin7  cos2  in terms of sines of multiples of .
8. Express cos6  in terms of cosines of multiples of .

37
 LESSON 4

APPLICATIONS OF DE MOIVRE’S THEOREM

In this lesson, we shall study the application of De Moivre’s theorem to find the sum of certain types
of trigonometric series. The series may be finite or infinite. The general method is known as the C + iS
method. If we are given as series of cosines such as C = cos  + cos 2  + ... cos n, we consider a similar
series of sines i.e. S = sin  + sin 2 + ... sin n.
The new series C + iS so formed can be summed up by different methods. Equating real and
imaginary parts we obtain the values of C and S.
The following formulae are mainly used in the summation of series.
1 − xn
(i) 1 + x + x 2 + x n − 1 = , x 1
1− x
(ii) (1 + x)n = 1 + nC1x + nC2 x2 + n
Cn xn
(iii) Euler’s Formula: If z is complex variable, then the complex exponential function ez is defined
z 2 z3
by e z = 1 + z + + + 
2! 3!
Replacing z by i  we have
(i) 2 (i)3 (i) 4 (i)5
ei = 1 + i + + + + +
2 3 4 5
2 3 4 5
= 1 + i − −i + + i + ...
2 3 4 5
 2 4   3 
= 1 − + +
 
i  − + ... 
 2 4   3 
= cos  + i sin 
Similarly e–i = cos  – i sin 
 ei + e–i = 2 cos 
and ei – e–i = 2 i sin 
Also (ei)n = eni = cos n + i sin n
where n is an integer.
We shall illustrate the method C + iS in the following solved examples.
Example 1. Find the sum to n terms of the series.
1 + x cos  + x2 cos 2 + ... + xn – 1 cos (n – 1)
Let C = 1 + x cos  + x2 cos 2 + ... + xn –1 sin (n – 1) 
S = x sin  + x2 sin 2 + ... + xn – 1 sin (n – 1)
 C + iS = 1 + x(cos  + i sin ) + x2(cos 2 + i sin 2)
+ ... + xn – 1 [cos(n – 1) + i sin (n – 1)] + i sin (n – 1) )]
= 1 + xei + x2e2i + ... + xn – 1 e(n – 1)i

38
 1 − x n e ni
= , summing the GP to n terms
1 − xei
(1 − x n e ni ) (1 − xe −i )
=
(1 − xei ) (i − xe −i )
[Multiply the numerator and denominator by 1 – xe–i]
1 − x n e ni − xe −i + x n + 1 e( n − 1)i
=
1 − x ( e i + e i ) + x 2
1 − x n [cos n + i sin n] − x (cos  − i sin ) + x n + 1 [cos (n − 1)  + i sin (n − 1)]
=
1 − 2 x cos  + x 2
(Using Euler’s Formulae)
Equating the real part we get
1 − x n cos n − x cos  + x n + 1 cos( n − 1) 
C=
1 − 2 x cos  + x 2
Example 2. Find the sum to n term of the series
sin 2 + sin 4 + sin 6 + ...
Let C = cos 2 + cos 4 + cos 6 + ... + cos 2n
S = sin 2 + sin 4 + sin 6 + ... + sin 2n
C + iS = e2i + e4i + ... e2ni
= e2i [1 + e2i + e2(n − 1)i ]
= x [1 + x + x2 + xn − 1 ] where x = e2i
x (1 − x n )
C + iS =
1− x
e 2i [1 − e 2 ni ]
=
1 − e 2 i
e 2i (1 − e 2 ni ) (1 − e −2i )
=
(1 − e 2i ) (1 − e −2i )
(Multiplying numerator and denominator by the conjugate of the denominator)
e 2i [1 − e 2 ni − e −2i + e 2( n − 1)i
=
1 − (e 2i + e −2i ) + 1
e 2i − e 2( n + 1)i − 1 + e 2 ni
=
1 − 2 cos 2 + 1
(cos 2 + i sin 2) − [cos 2(n + 1) + i sin 2 ( n + 1) ] − 1 + [cos 2 n + i sin 2 n]
=
2(1 − cos 2)
Equating the imaginary parts on both sides, we get

39
 sin 2 − sin 2(n + 1)  + sin 2n
S=
2(1 − cos 2)
2 sin  cos  − [sin 2(n + 1)  − sin 2 n)]
=
4 sin 2 
2 sin  cos  − 2 cos (2n + 1)  sin 
=
4 sin 2 
2 sin  [cos  − cos (2n + 1)  sin ]
=
4 sin 2 
2 sin ( n + 1)  sin n 
=
2 sin 
sin (n + 1)  sin n
= where sin   0 i.e.,   n
sin 
Example 3. Find the sum of the series
cos  nC1 cos ( + ) + nC2 cos ( + 2) + ... + nCn cos ( + n)
Let C = cos  nC1 cos ( + ) + nC2 cos ( + 2) + ... + nCn cos ( + n)
and S = sin  + nC1 sin ( + ) + ... + nCn sin ( + n)
Here C + iS = ei + nC1ei ( + ) + nC2ei ( + 2) + ... + nCnei ( + n)
= ei[1 + nC1ei + nC2e2i + ... + nCneni
= ei [1 + ei ]n
= ei[1 + cos  + i sin ]n
n
   
= ei  2 cos 2 + 2i sin cos 
 2 2 2
n n
    
= ei  2 cos  cos + i
 2  2 2 
n
   n n 
=  2 cos  [cos  + i sin ] cos + i sin 
 2  2 2
n
    n   n  
=  2 cos  cos   +  + i sin  q +  
 2   2   2 
Equating the real part on both sides, we get
n
   n 
C =  2 cos  cos   + 
 2  2 
Sometimes the sum to n terms of a trigonometric series can be obtained by using simple
trigonometric formulae viz.
2 sin A sin B = cos (A – B) – cos (A + B)
2 cos A sin B = sin (A – B) – sin (A + B)

40
 These formulae can be useful when the nth term of a series can be expressed at the difference of two
consines of sines. The following example will illustrate this method.
Example 4. Find the sum to n terms of the series.
sin  + sin 3 + sin 5 + ...
Let S = sin  + sin 3 + sin 5 + ... + sin (2n – 1)
Multiplying both sides by 2 sin  ( 0) we get,
2 sin  · S = 2 sin2  + 2 sin  · sin 3 + 2 sin  · sin 5 + ... + 2 sin  sin (2n – 1)
= [1 – cos 2] + [cos 2 – cos 4] + [cos 4 – cos 6] ...
+ [cos 2(n –1)  – cos 2 n)
= 1– cos2
n
2
= 2 sin n
2 sin  n
S= = sin 2 n cosec 
2 sin 
provided sin  0 i.e.,   n
Example 5. Find the sum of the infinite series.
sin 3 sin 5
sin  − + +
3 5
sin 3  sin 5
Let S = sin  − + +
3 5
cos 3 cos 5
C = cos  − + +
3 5
and we have
e3i e5i
C + iS = ei − + −
3 5
x3 x5
= x− + − where x = ei
3 5
= sin x
= sin [cos  + i sin ]
= sin (cos ) · cos (i sin ) + cos(cos ) · sin (i sin )
= sin (cos ) · cos h (sin ) + i cos(cos ) · sin h (sin )
Equating imaginary parts on both sides, we get
S = cos (cos ) · sin h (sin ) [ cos i = cos h, sin iq = i sin h]
from the definition of Hyperbolic functions)
Example 6. Sum of the series
1 1.3 1.3.5
1− cos  + cos 2 − cos 3 + 
2 2.4 2.4.6
1 1.3
Let C = 1 − cos  + cos 2 +
2 2.4

41
 1 1.3
S = − sin  + sin 2 +
2 2.4
1 i 1.3 2i 1.3.5 3i
Here C + iS = 1 − e + e − e + ...
2 2.4 2.4.6
1
−
= (1 + ei ) 2

1
−
= [1 + cos  + i sin ] 2

1
−
    2
=  2 cos 2 + 2 i sin  cos 
 2 2 2
1
−
     2
=  2 cos    cos + i sin  
 2 2 2 
1
−
  2   
=  2 cos  cos 4 − i sin 1 
 2  
Equating real parts on both sides we get,

cos
C= 4

2 cos
2
Example 7. Find the sum to infinitely of the series
1 1
sin  + sin 2 + 2 sin 3 + ...
2 2
1 1
Let C = cos  + cos 2 + 2 cos 3 + ...
2 2
1 1
S = sin  + sin 2 + 2 sin 3 + ...
2 2
1 1
 C + iS = ei + ei + 2 e 2i + ...
2 2
 1 1 
= ei 1 + ei + 2 e2i + ... 
 2 2 
1
= e i 
1
1 − e i
2
[Summing the G.P. to ]
 1 
ei 1 − e−i 
=  2 
 1 − i    1 − i 
1 − 2 e  1 − 2 e 

42
 1
ei −
= 2
1 i 1
1− (e + e−i ) +
2 4
1
cos  + i sin  −
= 2

− cos 
4
4(cos  + i sin ) − 2
=
5 − 4 cos 
Equating the imaginary parts on both sides, we get
4 sin 
S=
5 − 4 cos 
Example 8. Find the sum to  of the sides
cos   cos  + cos2   cos2 2 + cos3   cos 3 +
Let C = cos   cos  + cos2   cos2 2 + cos3   cos 3 +
S = cos   sin  + cos2  sin 2 + cos3   sin 3 +
Here C + iS = cos   ei + cos2   e2i + cos3   e3i + ...
= cos   ei [1 + cos   ei + (cos   ei )2 + ]
 1 
= cos   ei  i 
1 − cos   e 
(Summing the G.P. to )
cos  ei [1 − cos   ei ]
=
(1 − cos   e ) (1 − cos   e )
i − i

cos  ei − cos 2 

=
( )
1 − cos  ei + e −i + cos 2 

cos  [cos  + i sin ] − cos 2 

=
1 − 2 cos 2  + cos 2 
cos 2  + i sin   cos  − cos 2 
=
1 − cos 2 
i sin   cos 
=
sin 2 
It follows that C= 0
provided sin2  0
i.e.,   k where k is an integer
i.e., The sum of the given infinite series is zero, provided  is not a multiple for .

43
 EXERCISES

Find the sum to n terms of the following series :

1. cos  + cos 2 + cos 3 + ...
2. sin  · cos 2 + sin 2 · cos 3 + sin 3 · cos 4 + ...
3. sin  + sin( + ) + sin ( + 2) + ...
4. 2 cos 2 + 22 cos 4 + 23 cos 6 + ...
Find the sum to infinity of the following series :
1 1 .3
5. 1 + cos 2 − cos 4 + cos 6 + ...
2 2.4 2.4.6
1 1
6. sin  − sin 2 + sin 3 + ...
2 3
7. sin  · sin  + sin  · sin 2 + sin3  · sin 3 + ...
2

sin 2  sin 3 
8. 1 + sin   cos  cos 2 + cos 3 + ...
2 3
cos 2  cos3 
9. 1 − cos   cos  + cos 2 − cos 3 + ...
2 3
sin( + 2) sin ( + 4)
10. sin  − + + ...
3 4

44
 LESSON 5

THEORY OF EQUATIONS I

5.1 In this lesson we shall deal with the theory of equations. We presume that you are all familiar with
the theory of quadratic equations. In the present course of study of theory of equations, we shall confine
ourselves to cubic and biquadratic equations.

5.2 Polynomial and Equations

We know that a function p(x) of the form
p(x) = a0 xn + a1xn − 1 + a2 xn − 2 ... + an , a0  0,
where the coefficients a0, a1, a2, ... an are all complex numbers (a0  0) is called a polynomial of
degree n, n being a positive integer. For example, is a polynomial of degree 3. The expansion p(x) = 0 is
called an nth degree equation, that is,
p(x) = a0 xn + a1xn − 1 + a2 xn − 2 ... + an = 0, a0  0 ...(1)
The fundamental theorem of Algebra states that every equation of the form (1), has at least one
complex number as a root. The proof of this result is beyond the scope of the present course of study and
hence is omitted.
In the discussion thereafter, it will be presumed that the coefficients like a0a1 ... an are all complex
numbers. Now discuss the following theorem which is known as Factor theorem. We already know that
if  is a root of an equation, say, a0x2 + a1x + a2 = 0, a0  0 then x –  is a factor of the corresponding
polynomial a0x2 + a1x + a2. For example, 3 is a root of the equation x2 – x – 6 = 0, then it is easy to verify
that x – 3 is a factor of the polynomial x2 – x – 6. We generalize this idea into the following theorem :

5.3 Factor Theorem

A number  is a root of the equation
a0 xn a1xn − 1 + + an = 0 ...(1)
if and only if (x – ) is a factor of the polynomial
p(x) = a0 xn + a1xn − 1 + + an ...(2)
where a0  0, a1, a2 ..., an are complex numbers and n is a positive integer.
Proof. In the first part, it is given that  is a root of the equation (1) and we have to prove that
(x – ) is a factor of (2).
We divide the polynomial p(x) of degree n by a polynomial (x – ) of degree one. Then, the quotient
will be a polynomial of degree (n – 1). If the remainder is zero, then obviously (x – ) is a factor of p(x).
If the remainder is not zero, then the remainder is a constant and let it be R. Also let the quotient be
denoted by
a0 x n − 1 + b1 x n − 2 + b2 x n − 3 + + bn − 2 x + bn − 1
Then we have
a0 x n + a1 x n − 1 + + an − 1 x + an  ( x − ) [a0 x n − 1 b1x n − 2 + + bn − 2 x + bn − 1 ] + R

45
 Putting x =  on both sides, we get
a0  n + 1  n − 1 + + an − 1 + an  0 + R = R
In other words,
R = a0  n + 1  n − 1 + + an − 1 + an ...(3)
Now again, since  is a root of the equation (1), therefore by putting x =  in (1), we get
a0  n + a1 n − 1 + + n − 1  + n = 0 ...(4)

From (3) and (4), it follows that R = 0 so that (x – ) is a factor of p(x).

In the second part, it is given that (x – ) is a factor of the polynomial p(x) given by (2), we have to
show that  is a root of the equation (1).
Since (x – ) is a factor of the polynomial p(x), therefore when we divide p(x) by (x – ), we must
get R = 0, and consequently from (3), we get
a0  n + a1  n − 1 + + an − 1  + an = 0
which implies that a satisfies the equation
0 x n + a1 x n − 1 + + an − 1 x + an = 0
i.e.,  is a root of the equation (1).
This completes the proof of the theorem.
Now we discuss the theory of cubic equations.

5.4 Cubic Equation

The general cubic equation is of the form
a0x3 + a1x2 + a2x + a3 = 0, a0  0 ...(5)
where a0, a1, a2, a3 are given complex numbers.
By the fundamental theorem of algebra, this equation (5) has atleast one root. Let this root be
denoted by ‘’. Then by Factor Theorem (x – ) is a factor of the polynomial a0x3 + a1x2 + a2x + a3.
Divide the polynomial by x –  and let the quotient be denoted by a0x2 + b1x + b2.
Then we have
a0x3 + a1x2 + a2 x + a3 (x – ) (0x2 + b1x + b2) ...(6)
Now consider the quadratic equation
a0x2 + b1x + b2 = 0 ...(7)
Again by the fundamental theorem of algebra equation (6) must have at least one root, say,  and
accordingly (x – ) must be a factor of the polynomial a0 x2 + b1x + b2 = 0. Dividing a0x2 + b1x + b2 by (x
– ), we get the quotient, say, (a0x + c1) so that we have
a0 x 2 + b1x + b2 = ( x − ) ( a0 x + c1 ) ...(8)
 c 
Also a0 x + c1 = a0  x + 1  = a0 ( x −  ) ...(9)
 a0 

46
 c1
= − , a0  0
a0
Combining (6), (8) and (9), we get
a0 x3 + a1x2 + a2 x + a3  a0 (x – ) (x – ) (x – ) ...(10)
Substituting x = , ,  successively on both sides of (9), we observe that each of the values, , , 
satisfies the equation (5). Thus , ,  are the root of the equation (5). Also it may be noted that no
number different from , ,  can be roots of the equation (5).
For, if possible let , a number different from each of , ,  be a root of this equation (5). Then by
substituting x =  on both sides of (10), we find that while the left hand side is zero by virtue of the
supposition that  is a root but the right hand side is non-zero and thereby we arrive at a contradiction.
Hence, , ,  are the only roots of the equation (5). Thus we have proved that a cubic equation has three
and only three roots. These roots may be all distinct (unequal), may be repeated (equal) or may be that
only two are repeated.
From the above discussion, it is clear that if , ,  are the three roots of the a cubic equation, then
the corresponding factors are (x – ), (x – ), (x – ) and hence the equation is given by
(x – ) (x – ) (x – ) = 0
or x3 − ( +  + ) x2 + ( +  + ) x − a = 0.
We discuss the following examples :
Example 1. Find the equation whose roots are 1, 2, 3.
Solution. The required equation is given by
(x – 1) (x – 2) (x – 3) = 0.
or x3 – 6x2 + 11x – 6 = 0
[By putting  = 1,  = 2,  = 3 in equation (11) above]
Example 2. Find the equation whose roots are 3, 3, –2.
Solution. (x – 3) (x – 3) (x + 2) = 0.
or x3 – 4x2 –3x + 18 = 0

5.5 Reactions between the Roots and Co-efficient of a Cubic Equation

Let , ,  be the roots of a cubic equation
a0 x3 + a1x2 + a2 x + a3 = 0
Then we have the identify
a0 x3 + a1x2 + a2 x + a3  a0 ( x − ) ( x − ) ( x − )
which implies
a0 x3 + a1x2 + a2 x + a3  a0 [ x3 − ( +  + ) x2 + ( +  + ) x − ]
Equating the co-efficients powers of x, i.e., x3, x2, x and the constant terms, we get
a0 = a0
a1 = –a0 ( +  + )
a2 = a0 ( +  + )ss
3 = a0 (− ) = − a0 

47
 These give
a1
++= −
a0
a2
 +  +  =
a0
a3
 = −
a0
In other words, we can write these relations as
coefficient of x 2
Sum of the roots = –
coefficient of x 3
coefficient of x
Sum of the products of roots taken in pairs =
coefficient of x3
– constant term
=
coefficient of x3
Thus we note that though we are not able to find the values of , ,  separately, yet we have been
able to express the three conditions combinations namely  +  + ,  +  + ,  in terms of
coefficients 0  0, 1, 2, 3.

5.6 Symmetric Functions

Consider the following expressions :
(i) a2 + b2 + 2
(ii) ( + )2 + ( + )2 + ( + )2
(iii) 2 + 2 + 2 + 2 + 2 + 2
Each of the above is a function of , ,  with the property that if any two of , ,  are interchanged
the function remain unaltered. Such functions are called symmetric functions. More precisely a function f
(1, 2, ... n) of n variables is said to be symmetric function of 1, 2, ... n if it remains unaltered by
interchanging any two of
1, 2 ... n.
Thus, for example
 +  + ,  +  + ,  are symmetric functions of , , .
In the case of a cubic equation whose roots are , ,  the functions
 +  + ,  +  + ,  are called basic symmetric functions.
Sigma Notation for Symmetric Functions
It is often convenient to describe a symmetric function by indicating only one or more terms in such
a manner that the other terms are obtainable from the given term on replacing the roots therein by other
roots in all possible ways. Also then we put down the sign (sigma) before the given term.
Thus for a cubic equation
 = ( +  + )

48
  = ( +  + )
2 = 2 + 2 + 2 + 2g + 2 + 2
2 = 2 + 2 + 2
3 = 3 + 3 + 3 and so on.
It is interesting to know that every symmetric function of , ,  can be expressed in terms of the
symmetric function  +  + ,  +  + , and  whose values can be written down from the cubic
equation whose roots are , , . We illustrate this fact by the following examples.
Example 3. If , ,  be the roots of the equation
x3 + px2 + qx + r = 0
find the value of
(i) ( + )2
(ii) ( + ) ( + ) ( + )
Solution. From the given equation, we have
− co-efficient x 2 p
( +  + ) = 3
=− =− ps
co-efficient x 1
co-efficient of x q
( +  + ) = 3
= =p
co-efficient of x 1
constant term r
 = − 3
=− =−r
co-efficient of x 1
(i) ( − )2 =  (2 + 2 − 2 ) = 2 + 2 − 2 = 22 − 2
= 2[(a)2 − 2] − 2 = 2()2 − 6 = 2 p2 − 6q
(ii) We shall now express the symmetric function, ( + ) ( + ) ( + ) in terms of the basic
symmetric functions.
We have, by actual multiplication
( +  ) (  + ) ( + ) = ( + ) ( +  + a2 + ]
= 2 + (2 + 2 + 2 + 2 + 2 + 2)
Also we have ( +  + ) ( +  + )
= (2 + 2 + 2 + 2 + 2 + 2) + 3
that (2 + 2 + 2 +  + 2 + )
= ( +  + ) ( +  + ) –3
Thus we have ( + ) ( + ) ( + )
= 2 + {( +  + ) ( +  + ) –3}
= ( +  + ) ( +  + ) –
= – pq + r.
Another method: We have
( + ) ( + ) ( + )
= ( +  +  – ) ( +  +  – ) ( +  +  – )
= (–p – ) (–p – ) (–p – ), [ ( +  +  = – p)]

49
 = – (p + ) (p + ) (p + )
= – [p3 + p2 ( +  + ) + p) ( +  + ) + ]
= –[ p + p2 (− p) + ( pq − r )] = − pq + r.
Example 4. If , ,  be the roots of the equation x3 + px2 + qx + r = 0. Find the value of
(i)  22
 
(ii)   + 
 
Solution. We have ++= –p
 +  +  = q
 = – r
Now
(i) ( +  + )2 = 22 + 22 + 22 + 22 + 22 + 22
= (22 + 22 + 22) + 2 ( +  + )
 (22 + 22 + 22) = ( +  + )2 – 2 ( +  + )
= q2 – 2 (–r) (–p)
= 2 – 2rp.
   2 + 2
(ii)  +  = 
  
 2 + 2 2 +  2  2 +  2
= + +
  
 ( 2 + 2 ) +  (2 +  )2 +  ( 2 +  2 )
=

 2
=

 − 3
=

(− p) (q) − 3(− r )
=
−r
pq − 3r
=
r
Example 5. If , ,  be the root of the cubic equation x3 + px2 qx + r = 0. Find the value of
(3 + 3 + 3 ).
Solution. We have
( +  + ) (2 + 2 + 2)
= (3 + 3 + 3) + (2 + 2 + 2 + 2 + 2 + 2)
  2 = 3 + 2 

50
 3 =  2 – 2  ...(1)
But as shown in Example 3.
( + 2 + 2 +  + 2 + 2)
= ( +  + ) ( +  + ) – 3
= – pq + 3r
Again
2 + 2 + 2 = ( +  + )2 – 2( +  + ) = p2 – 2q.
Also
++= –p
Hence from (1), we get
3 = (3 + 3 + 3) = – p (p2 – 2q) – (–pq + 3r)
= –p3 + 3pq – 3r.
Another method :
As , ,  are the roots of the equation x3 + px2 + qx + r = 0, we have
3 + p2 + q + r = 0.
3 + p2 + q + r = 0.
3 + p2 + q + r = 0.
Adding these three equations we obtain
(3 + 3 + 3) + p (2 + 2 + 2) + ( +  + ) + 3r = 0 ...(1)
Now ++= –p
2 + 2 + 2 = ( +  + )2 – ( +  + )
= p2 – 2q
Substituting in (1), we obtain
(3 + 3 + 3) + p(p2 – 2q) – qp + 3r = 0
 3 + 3 + 3 = – p3 + 3pq – 3r.
Example 6. , ,  be the roots of the equation
x3 + qx + r = 0,
then prove that
5 + 5 + 5 3 + b3 + 3 2 + b2 +  2
= 
5 3 2
Solution. Since , ,  are the roots of equation x + qx + r = 0
3

 3 + q + r = 0 ...(i)
3 + qb + r = 0 ...(ii)
3 + q + r = 0 ...(iii)
Adding (i), (ii), (iii), we get
3 + q + 3r = 0
Also
 = 0,  = q,  = –r

51
 3 = –q  – 3r = – 3r ...(iv)
2 = ()2 – 2 = 0 – 2q = – 2q ...(v)
Multiply (i) by 2, (ii) by 2 and (iii) by 2 and adding we get
5 + q 3 +sss r 2 = 0
 5 = – q 3 – r 2 = –q(–3r) – r (–2q)
= 3qr + 2qr = 5qr by (iv) and (v)
1 5 1 1 
  = qr =  3   2 
5 3 2 
Hence the result.
Example 7. Find the equation whose roots , ,  satisfy the relations :
++= 3
2 + 2 + 2 = 5
3 + 3 + 3 = 7
Also find the value of 4 + 4 + 4.
Solution. Let the required equation be
ax3 + bx2 + cx + d = 0
b
so that ++= −
a
c
 +  +  =
a
d
 = −
a
b
We, therefore have 3= ++= − ...(ii)
a
and 5 = a2 + b2 +  2 = ( +  + ) − 2 ( +  + )
2c
= 9−
a
b c
Hence, from (ii), − 3 and from (iii ) = 2.
a a
Since , , , satisfy (i) we have
a3 + b2 + c + d = 0

a3 + b2 + c + d = 0

a3 + b 2 + c + d = 0
Adding these, we obtain
a(3 + 3 + 3) + b(2 + 2 + 2) + c ( +  + ) + 3d = 0
i.e., 7a + 5b + 3c + 3d = 0

52
 Dividing by ‘a’ we obtain
b c d
7+5 +3 +3 =0
a a a
b c
Substituting the values of and obtained above, we obtain
a a
d
7 − 15 + 6 + 3 =0
a
d 2
i.e., =
a 3
Thus the required equation (i) on the diving by a is
b c d
x3 + x 2 + x + = 0
a a a
2
i.e., x3 − 3x 2 + 2 x + =0
3
i.e., 3 x3 − 9 x 2 + 6 x + 2 = 0
Since , ,  are the roots of this equation, they must satisfy the equation.
 33 – 92 + 6 + 2 = 0
33 – 92 + 6 + 2 = 0
33 – 92 + 6 + 2 = 0
Multiplying the above equations by , ,  respectively and adding we get,
34 – 93 + 62 + 2 = 0
 34 = 93 – 62 – 2
1
34 = [9.7 − 6.5 − 6]
3
= 21 – 10 – 2
=9

5.7 Biquadratic Equations

We consider the general biquadratic equation, i.e., the equation of degree four viz.,
a0 x4 + a1x3 + a2 x2 + a3 x + a4 = 0
where a0  0, a1, a2, a3, a4 are complex coefficients.
Let ‘’ be a root of this equation so that (x – ) is a factor of the polynomial.
a0 x4 + a1x3 + a2 x2 + a3 x + a4
Then we have a certain
a0 x4 + a1x3 + a2 x2 + a3 x + a4  ( x − a) (a0 x3 + b1x 2 + b2 x + b3 ) ...(1)
where a0 x3 + b1x2 + b2 x + b3 is the quoedient.

53
 As proved previously a0 x3 + b1x2 + b2 x + b3 = 0 being a cubic equation, has 3 roots. Let these roots
be ,  and .
Then, we have
a0 x3 + b1x2 + b2 x + b3  a0 ( x − ) ( x −  ) ( x − ) ...(2)
From (1) and (2) we obtain
a0 x4 + a1x3 + a2 x2 + a3 x + a4  a0 ( x − ) ( x − ) ( x −  ) ( x − )ax ...(3)
Substituting x = , , ,  successively on the two dies of (3), we see that each of the values , , ,
 satisfies the equation a0 x4 + a1x3 + a2 x2 + a3 x + a4 = 0.
Again, if substitute for x, a value different from each of , , ,  on both sides of (3), we see that
since R.H.S. is not zero, L.H.S. is also not zero. Thus no value other than , , ,  is a root of the
equation.
We have thus shown that a biquadratic equation has four and four roots.
The roots , , ,  may not be all different and we have the following different possibilities
(i)  =      Two roots equal.
(ii)  =      Roots equal in pairs.
(iii)  =  =    Three roots equal.
(iv)  =  =  =  All the four roots equal.
(v)        No two roots equal.
The factors of a a0 x4 + a1x3 + a2 x2 + a3 x + a4 are
(a) (x – )2 (x – ) (x – ) for the case (i)
(b) (x – )2 (x – )2 for the case (ii)
(c) (x – )3 (x – ) for the case (iii)
(d) (x – )4 for the case (iv)
(e) (x – ) (x – ) (x – ) (x – ) for the case (v)

5.8 Relation between the Roots and Coefficients of a Biquadratic Equation

a0 x4 + a1x3 + a2 x2 + a3 x + a4 = 0
We have
a0 x4 + a1x3 + a2 x2 + a3 x + a4  a0 ( x − ) ( x − ) ( x −  ) ( x − )ax
 a0 [ x4 − ( +  +  + ) x3
+ ( +  +  +  +  + )x2
− ( +  +  + ) x +  + ]
Equating the coefficients of x3, x2, x and the constant term, we have
a1 = – a0 ( +  +  + )
a2 = a0 ( +  +  +  +  + )
a3 = – a0 ( +  +  + )
a4 = a0  

54
 These give
a1 co-efficient of x3
( +  +  + ) = − =−
a0 co-efficient of x 4

a2 co-efficient of x 2
( +  +  +  +  + ) = =
a0 co-efficient x 4
a2
 ( + ) ( + ) +  +  =
a0
a3 co-efficient of x
( +  +  +  = − =−
a0 co-efficient x 4
a3
 (  + ) +  ( + ) = −
a0
a4 constant term
 = =
a0 co-efficient of x 4
Any function of , , ,  which does not change its value of inter-changing any two of them is a
symmetric function.
It may be easily seen that
 +  +  +  ( )
 +  +  +  +  +  ( )
 +  +  +  ( )

are four symmetric functions known as basic symmetric functions of the roots of the biquadratic.
Every symmetric function of , , ,  can be expressed in terms of the above symmetric functions.

EXERCISE

1. If , , , are the roots of the equation

x3 – 5x2 + x + 12 = 0
Calculate the value of 2.
2. If , , , be the roots of the equation x3 + qx + r = 0, prove that
3(2 ) (5 ) = 5(3 ) ( ).
[Hint : Since, , , , are the roots of the given cubic equation]
3 + q + r = 0
3 + q + r = 0
3 + q + r = 0
To find 4 multiply equations (1) , , , respectively and add.
To find 5 multiply (1) 2, 2, 3, respectively and add.
3. Find the sum of the cubes of the roots of the equation
x3 – 6x2 + 11x – 6 = 0

55
4. If , , , be the roots of the equation
x3 + 5x2 – 6x + 3 = 0
Find the values of  −3
 −3 1 1 1 3  3 
 Hint :   + + = 
 3 3  3 33  3 

Again 33 − 3()2 =  [22 − ()]

5. If , , , be the roots of the cubic x3 + qx + r = 0 find the values of
1
(i) 
b+c
b2 + c2
(ii) 
b+c
(iii) (b + c)2
6. If , , , are the roots of the equation
x3 – px2 + qx – r = 0.
Find the value of
1
(i) 
2
(ii) 2
(iii) ( + ) ( +  ) (  + )

56
 LESSON 6

THEORY OF EQUATIONS II

In this lesson, we shall discuss the methods of transforming a given equation into another equation
or to solve the given equation under certain given conditions.

6.1 Transformation of Equations

We shall discuss methods of finding an equation whose roots are related to those of a given
equation. The example below, will indicate and illustrate and methods which may be employed for the
purpose.
We know that the cubic equation, with roots ,  and  is given by
(x – ) (x – ) (x – ) = 0
i.e., x – x ( +  + ) + x ( +  + ) –  = 0
3 2

i.e., x3 – x2 + x  –  = 0

Also the biquadratic equation with roots, , , ,  is
(x – ) (x – ) (x – ) (x – ) = 0
i.e., x4 − x3  + x2 − x +  = 0

Examples
Example 1. Find the equation whose roots are the negatives of the roots of the equation
5x3 – 3x2 + 7x + 2 = 0 ...(1)
Solution. If , , , are the roots of the given equation, then the roots of the new equation are –, –
, –.
We write y = –x so that when x takes values, , ,  then y takes the values –, –, – respectively,
Write y = –x  x = – y ...(2)
Eliminating x between (1) and (2)
We get, 5(–y)3 – 3 (–y)2 + 7(–y) + 2 = 0
or –5y3 – 3y2 – 7y + 2 = 0
If 5y3 + 3y2 + 7y – 2 = 0 is the required equation.
Example 2. Find the equation whose roots are the roots of
x4 – 5x3 + 7x2 – 17x + 11 = 0 ...(1)
each diminished by 5.
Solution. If , , ,  are the roots of the given equation we have to find the equation whose roots
are  – 4,  – 4,  – 4, and  – 4.
Write y =x – 4  x = y + 4
Eliminating x between (1) and (2) we get
(y + 4)4 – 5 (y + 4)3 + 7 (y + 4)2 – 17 (y + 4) + 11 = 0
i.e., y4 + 16y3 + 96y2 + 256y + 256 – 5(y3 + 12y2 + 48y + 64)
+7(y2 + 8y + 16) – 17 (y + 4) + 11 = 0

57
 Simplifying we get,
y4 + 11y3 + 43y2 + 55y – 9 = 0
as the required equation.
Example 3. Find an equation whose roots are three times those of the equation
2x3 – 5x2 + 7 = 0
Solution. If , ,  are the roots of given equation, then we write equation whose roots are 3, 3,
3.
We write y = 3x
y
 x= .
3
y
Substituting x= . in the given equation, we gets
3
3 2
 y  y
2  − 5   + 7 = 0
3 3
 2y3 – 15y2 + 189 = 0
which is the required equation.
Example 4. Find the equation whose roots are the squares of the roots of the equation.
x3 + qx + r = 0
Solution. We shall give two method for the solution.
First Method:
Let , , , be the roots of the given equation so that we require the equation with roots 2, 2. 2 we
write
y = x2 ...(1)
so that when x takes up the vales , , , then y takes up the values 2, 2, 2.
Aslo if x3 +qx + r = 0 ...(2)
then x necessarily takes up the values, , ,  and these only.
Thus the required equations will be obtained by eliminating x between (1) and (2) with y indicating
the unknown.
Thus we have
y y +q y +r = 0

or y ( y + 9) = – r
Squaring, we get
y (y + q)2 = r2
i.e., y3 + 2y2 + q + yq2 + r2 = 0
which is the required equation.
Second Method:
Indicating the unknown by y, the equation with roots 2, 22, is
(y – 2) (y – )2 (y – )2 = 0

58
 i.e., y3 – y2 (2 + 2 + 2) + y (22 + 22 + 22) – 222 = 0 ...(3)
We have now to compute the three symmetric functions.
viz., (2 + 2 + 2), (22 + 22 + 22) and 222
From the given equation, we have
( +  + ) = 0, ( +  + ) = ,  = – r
Hence (2 + 2 + 2) = ( +  + )2 – 2( +  + ) = – 2q
and (22 + 2 + 2) = [( +  + )2 – 2 ( +  + )] = q2.
Also 222 = r2
Making substitutions in (3), we obtain the required equation.
Example 5. If , ,  are the roots of the cubics x3 + px2 + qx + r = 0, find the equation whose roots
1 1 1
are  − ,  − and  − where , ,   0.
  
1  −1   −1   −r − 1   r + 1
Solution. − = and   = =   [  = –r]
      −r   r 
1  r + 1
Similarly − = 
  r 
1  r + 1
− =  
  r 
 r + 1 1 1 1
We write y =   x so that as x takes , ,  then y takes up the values  − , − , −
 r    
 r + 1
respectively. Eliminating x between the given equation and y =  x
 r 
3 2
 ry   ry   ry 
we get   + p  + q +r =0
 r + 1  r + 1  r + 1
i.e., r 3 y3 + pr 2 (r + 1) y 2 + qry (r + 1)2 + r (r + 1)3 = 0
r 2 y3 + pr (r + 1) y 2 + q(r + 1)2 y + (r + 1)3 = 0
Example 6. If , ,  are the roots of x3 – px2 + r = 0, form the equation whose roots are
+ +  +
, , and .
  
Solution. , ,  are the roots of
x3 – px2 + r = 0
  +  +  = p,  +  +  = 0,  = – r ...(1)
+ ( +  +  ) −  p − 
We have = =
  
+ p −  +  p −
Similarly = , =
   

59
 p−x
We write y= ...(2)
x
p − p − p−
so that when x takes the values , , ; y takes the values , and respectively.
  
The required equation is obtained by eliminating x between (1) and (2).
p
From (2) xy = p – x  x(y + 1) = p  x =
1+ y
 Putting the value of x in (1) we have
p p2
−p +r = 0
(1 + y 3 ) (1 + y ) 2
or p3 – p3 (1 + y) + r (1 + y)3 = 0
i.e., ry3 + 3ry2 + (3r – p3) + y + r = 0
is the required equation.
Example 7. Find the equation whose roots are less by 2 less than the roots of the equation
x4 + 2x3 – 3x2 – 2x + 2 = 0 ...(i)
Solution. If a, b, c, d be the roots of the given equation
We require the equation with roots a – 2, b – 2, c – 2, d – 2,
We write y = x – 2 i.e., x = y + 2
and eliminate x between (ii) and the given equation (i).
The eliminant is
(y + 2)4 + 2(y + 2)3 – 3(y + 2)2 – 2(y + 2) + 2 = 0
i.e., (y4 + 8y3 + 24y2 + 32y + 16) + 2(y3 + 6y2 + 12y + 8)
– 3 (y2 + 4y + 4) – 2 (y + 2) + 2 = 0
i.e., y4 + 10y3 + 33y2 + 42y + 18 = 0
Example 8. Transform the equation
ax3 + by2 + cx + d = 0 ...(i)
into one, from which the second term is missing.
Solution. Let , ,  be the roots of the given equation.
We shall find another equation whose roots are
 + h,  + h,  + h
where h is a number to be so chosen that the new equation is free of the second term
We write y =x + h  x = y – h ...(ii)
so that the new equation will be obtained on eliminating x between (i) and (iii)
The eliminant is
a(y – h)3 + (y – h)2 + c (y – h) + d = 0
i.e., ay3 + y2 (b – 3ah) + y(3ah2 – 2bh + c) + (–ah3 + bh2 – ch + d) = 0 ...(iii)

60
 We choose h so that  – 3ah = 0, i.e., b/3a
This choice of h will make the coefficient of y2 zero, i.e., the second term missing.
Substituting the value of h (iii), we obtain
 b2 2b2   −ab3 b3 bc 
ay 3 +  3a 2 − + c  y +  3
+ 2
− + d  = 0
 9a 3a   27a 9a 3a 
 3ac − b2  27a 2 d − 9abc + 2b3
i.e., ay 3 +   y + =0
 3a  27a 2

i.e., 273 y3 + 9a (3ac − b2 ) y + (27a 2d − 9abc + 2b3 ) = 0

which is the required equation.

EXERCISE
1. From an equation whose roots are negatives of the roots of the equation
x3 + 6x2 + 8x – 9 = 0.
2. Form an equation whose roots are twice the roots of
2x4 + 3x2 + 7x – 9 = 0.
3. Find the equation whose roots are the squares of the roots of the equation
x3 – x2 – 8x + 6 = 0.
4. Find the equation whose roots are the reciprocals of the roots of the equation
ax3 + bx2 + cx + d = 0
5. If a, b, c are the roots of x3 + qx + r = 0, for the equation whose roots are
b2c2, c2a2, a2b2.
6. Find the equation whose roots are the roots of the equation
x4 – 3x3 + 7x2 + 5x – 2 = 0
each diminished by 2.
7. Transform the equation
x3 – 6x2 + 4x – 7 = 0
into one, from which the second term is missing.

6.2 Roots Related by Some Conditions

Sometimes we are required to find conditions under which the roots of given equation may be
related in a given manner. We are also sometimes required to solve an equation it is given that the roots of
the equation are related in some given manner. The examples given below will illustrate the procedure to
be adopted in solving the two types of problems referred to here.
Example 1. Solve the equation
x3 – 12x2 + 39x – 28 = 0
whose roots are given to be in arithmetical progression.
Solution. We put down a general set of three number in A.P. We could take this set simply as
a, a + d, a + 2d
But it would, however, be found more convenient to take this set as
a – d, a, a + d.

61
 Let us suppose the roots of the given equation are
a – d, a, a + d.
So that we have,
(a – d) + a + (a + d) = sum of the roots = 12.
(a – d) a + a (a + d) + (a – d) (a + d) = sum of the product of the roots takes in pairs = 39.
(a – d) × a × (a + d) = product of roots = 28
Thus we
3a = 12 ...(i)
3a2 – d2 = 39 ...(ii)
a(a2 – d2) = 28 ...(iii)
From (i), we get a= 4
Substituting a = 4 in (ii), we get
d2 = 9; i.e., d = ±3
We may easily see that a = 4, d = 3 as also a = 4, d = –3 satisfy (iii).
Taking a = 4, d = + 3 and putting these values in a – d, d, a + d, we see that the required roots are 1,
4, 7.
Again taking a = 4, d = – 3, we shall see that the roots are 7, 4, 1 which are the same as those
obtained above except for the order.
Thus we have proved that the roots of the given cubic are 1, 4, 7.
Example 2. Solve
8x3 – 14x2 + 7x – 1 = 0
give that its roots are in geometrical progression.
Solution. We suppose that its roots are a/r, a, ar.
a 14 7
We have + a + ar = sum of the roots =
r 8 4
a a 7
  a + a + (ar ) + (ar ) = sum of the products of the roots taken in pairs = 8
r r
a 1
 r  a(ar ) = product of roots = 8
 
Thus we have
a(1 + r + r  ) 7
= ...(i)
r 4
a 2 (1 + r + r  ) 7
= ...(ii)
r 8
1
and a3 = ...(iii)
8

62
Dividing (i) by (ii), we get
1 1
= 2 so that a =
a 2
Substituting this value of a in (i), we get
1 + r + r2 7
=
r 2
or 2r 2 − 5r + 2 = 0
or (2r – 1) (r – 2) = 0
1
The value of a satisfied (iii).
2
1 a 1 1
Taking a = , r = 2, we see that the roots , a, ar are , ,1.
2 r 4 2
1 1
Taking a = , r = we obtain the same roots but in the reverse order.
2 2
1 1
Thus 1, , , are the required roots.
2 4
Examples 3. Solve the equation 3x3 + 11x2 + 12 x + 4 = 0, the roots being in H.P.
(Harmonoic Progression)
Solution. Let the roots of the given equation be , ,  then
11
, ,  = −
3
 +  +  = 4
4
 = − .
3
Since , ,  are in H.P.
2 1 1
 = +
  
2
= 
+
 ( + ) = 2
  +  +  = 3
From (ii) and (iv), we get
3 = 4
4
  =
3
4
Substituting  = in (iii), we get
3

63
 4 4
 = −
3 3
  = –1.
Thus –1 is the roots of the given equation and (x + 1) is the factor of
3x3 + 11x2 + 12x + 4.
The given equation may be written as
3x2 (x + 1) + 8x(x + 1) + 4(x + 1) = 0
 (x + 1) (3x2 + 8x + 4) = 0
Now the other two roots of the given equation are obtained by the solving the equation
3x2 + 8x + 4 = 0
−8  64 − 48 – 8  4 −2
 x= = = ,−2
6 6 3
Hence the required roots are –1, – 2, –2/3.
Note : If , , ,  are the roots of biquadratic equation
a0 x + a1x3 + a2 x  + a3 x + a4 = 0
a1
Then +++= − ,
a0
+a2
( + ) ( + ) +  +  = ,
a0
−a3 a
 ( + ) + ( + ) = and  = 4 .
a0 a0
Example 4. Solve the equation
x4 − 8x3 + 14x2 + 8x − 15 = 0
the roots being in A.P.
Solution. Let the required roots be a – d, a, a + d, a + 2d.
Then (a – d) + a + (a + d) + (a + 2d) = 8  4a + 2d = 8  d = 4 – 2a.
(a – d + a) (a + d + a + 2d) + a(a – d) + (a + d) (a + 2d) = 14
Substituting d = 4 – 2a in this equation, we get
(2a – 4 + 2a) (2a + 12 – 6a) + a (a – 4 + 2a) + (a + 4 – 2a) (a + 8 – 4a) = 14
 (4a – 4) (12 – 4a) + a(3a – 4) + (4 – a) (8 – 3a) = 14
 48a – 48 – 16a2 + 16a + 3a2 – 4a + 32 – 8a + 12a + 3a2 = 14
 – 10a2 + 40a – 30 = 0
 a2 – 4a + 3 = 0
 (a – 1) (a – 3) = 0  a = 1, 3
Now d = 4 – 2a = 2 when a = 1
And d = 4 – 6 = – 2 when a = 3.
 a = 1, d = 2 and a = 3, d = –2.

64
 Now when a = 1, d = 2 ; Roots are –1, 1, 3, 5.
And when a = 3, d = – 2 ; Roots are 5, 3, 1, –1. This is the same roots but in different order.
 Required roots are –1, 1, 3, 5.
Example 4. Solve the equation
x4 + 15x3 + 70x2 + 120x + 64 = 0
the roots being in G.P.
Solution. We know that if four number be in G.P. then the product of the first and fourth is equal to
the product of the second and third. Thus, if the roots are in G.P. then the product of the two of the roots
must be equal to the product of the other two.
Let the roots be , , ,  then
 =  ...(i)
 +  +  +  = –15 ...(ii)
( + ) ( + ) +  +  = 70 ...(iii)
( + ) + ( + ) = – 120 ...(iv)
 = 64 ...(iv)
From (i) and (v), we get
()2 = 64   = 8
 = 8 = 
Substituting these values in (iii), we get
( + ) ( + ) = 70 – 16
= 54 ...(vi)
From (ii) ( + ) ( + ) = –15 ...(vii)
From (vi) and (vii), we find that ( + ) and ( + ) are the roots of the equation
t2 + 15t + 54 = 0
 t2 + 9t + 6t + 54 = 0
 (t + 9) (t + 6) = 0
 t = – 6, –9
Thus ( + ) = – 6 and  +  = – 9.
Also,  = 8 and  = 8
Thus ,  are the roots of equation
y2 + 6y + 8 = 0
and ,  are the roots of equation
y2 + 9y + 8 = 0 ...(viii)
Solving (viii), we get
y = – 2, – 4 ...(ix)
and solving (ix), we get
y = –1, – 8

65
Thus roots of given equations are –1, – 2, – 4, – 8.
Example 5. The equation
x4 – 2x3 + 4x2 + 6x – 21 = 0
has two real roots whose sum is zero. Solve the equation.
Solution. Let the roots be , , , , where  +  = 0
We have,  +  +  +  = sum of the roots = 2
( +  +  +  +  + ) = sum of the products of the roots taken in pairs = 4.
( +  +  + ) = sum of the product of the roots taken in traids = –6.
 = product of the roots = –21.
We now make use of the fact that  +  = 0
Now  +  +  +  +  +  =  = ( + ) +  ( + ) + 
= ( + ) + ( + ) ( + )
= ( + ) + [ +  =0]
Also  +  +  +  = ( + ) +  ( + )
=  ( + ), [ +  = 0]
Thus we have
+= 2 ...(i)
 +  = 4 ...(ii)
( + ) = – 6 ...(iii)
 = – 21 ...(iv)
From (i) and (iii),  = – 3
Also  +  = 0 , i.e.,  = –
2 = 3 or  =  3
2 = , we have  = − 3
Putting  = –3 in (iii) and (iv), we obtain
+= 2
 = 7
These give, on elimination ,
 (2 – ) = 7
i.e., 2 – 2 + 7 = 0
2  4 − 28
 = =1 6 i
2
It will be seen that other choices of the values  and  give us the same of roots.
The roots are therefore, 3, − 3,1 + 6 i,1 − 6 i.

66
Example 6. Solve the equation x4 + x3 − 16x − 4x + 48 = 0
Give that the product of the two roots of the equation is 6.
Solution. Let , , ,  be the roots of the given equation.
Then  +  +  +  = – 1, ...(1)
( + ) + ( + ) +  +  = – 16 ...(2)
 ( + ) +  ( + ) = 4 ...(3)
 = 48 ...(4)
It is given that  = 6. Then by (4),  = 8.
By (2), ( + ) ( + ) + 6 + 8 = – 16
 ( + ) ( + ) = – 30 ...(5)
From (1) and (5) if follows that  +  and  +  are the roots of
y2 + y – 30 = 0
 (y + 6) (y – 5) = 0
= – 6, 5
 += 5,
 +  = – 6.
( – )2 = ( + )2 –
4 = 25 – 24 = 1
  –  = ±1
Also ( – )2 = ( + )2 – 4 = 36 – 4 × 8 = 4
 –= ±2
Let us take  –  = 1 and  –  = 2
Also  +  = 5 and  +  = – 6
 2 = 6, 2 = –4   = 3,  = – 2
and  =  – 1= 3 – 1= 2,  =  – 2 = – 4
 Required roots are 3, 2, –2, – 4.
Example 7. Solve the equation
x4 − 5x3 + 10 x + 4 = 0
given that the product of two of the roots of the equation is equal to the product of the other two.
Solution. Let , , ,  the roots of the given equation,
Then  =  ...(1)
( + ) + ( + ) = 5 ...(2)
( + ) + ( + ) +  +  = 10 ...(3)
() ( + ) +  ( + ) = 0 ...(4)
 = 4 ...(5)
Substituting  =  in (5), we get
22 = 4
  = 2 = 

67
Substituting in (3), we get
( + ) + ( + ) + 2 + 2 = 10
 ( + ) ( + ) = 6
Also ( + ) ( + ) = 5 From (2)
  +  and  +  are the roots of equation
y2 – 5y + 6 = 0
 (y – 3) (y – 2) = 0
 y = 3, 2
Then a+= 3 and y +  = 2
 ( – )2 = ( + )2 – 4
= 9–8=1
 –= ±1
Also ( – )2 = ( + )2 – 4
= 4 – 8 = – 4 = 4i2
 –= ± 2i
Let –= 1 and  –  = 2i
Also += 3 and  +  = 2
  = 2,  = 1,  = 1 + i,  = 1 – i
Required roots are 2, 1, 1 ± i.
Examples 8. Find the condition that the roots of the given equation
px2 + qx2 + rx + s = 0
may be in arithmetical progression.
Solution. Let a – d, a, a + d be the roots of the given equation.
q
We have (a – d) + a + (a + d) =
p
q
i.e., 3a = − ...(i)
p
r
(a – d) a + a (a + d) + (a + d) (a – d) =
p
r
i.e., 3a2 – d2 = ...(ii)
p
s
a (a + d) (a – d) = −
p
s
a(a2 – d2) = − ...(iii)
p
r
From (i) and (ii), d2 = 3a 2 −
p
q2 r q 2 − 3 pr
= − =
3 p2 p 3 p2

68
Substituting the values of a and d2 in (iii), we obtain
q  q2 q 2 − 3 pr  s
−  2 − 2  = −
3p  9 p 3p  p
q s
 3
(q 2 − 3q 2 + 9 pr ) = −
27 p p
 q(9pr – 2q2) = 27p2s
which is the condition to be necessarily satisfied if the roots of the given equation are in A.P.
Note : You may verify that the equation in Ex. 1 does satisfy this condition.
Example 9. Find the condition for the equation to have a pair of equal roots.
Solution. Let the roots be , , .
We have 2 +  = 0 ...(i)
2 + 2 = 0 ...(ii)
2 = – q ...(iii)
Now (i) and (ii) give
p
2 + 2(–2) = p i.e., 2 = −
3
−p
 = –2 − 2
3
Substituting in (iii), the values of 2 and  we have
p −p 
−  − 2  = –q
3 3 
Squaring we obtain
4p
− = q2
27
i.e., 4 p3 + 27q2 = 0
which is the required condition.

EXERCISES
1. Solve the equation
x3 − 6 x2 + 3x + 10 = 0
whose roots are in A.P.
2. Solve the equation
3x3 − 26 x2 + 52 x − 24 = 0 the roots being in G.P.
3. Solve the equation
28x3 − 39 x2 + 12 x − 1 = 0
the roots being in Harmonical Progression.

69
 4. Solve the equation
4 x3 − 16 x2 − 9 x − 36 = 0
where the sum of the two of the roots being equal to zero.
5. Solve the equation
2 x3 − x2 − 22x − 24 = 0
two of its roots being in the ratio 3 : 4.
6. Solve the equation
2 x3 + x 2 − 7 x − 6 = 0
given that the difference of two roots is 3.
7. Solve the equation
x3 − 5x2 − 2x + 24 = 0
given that the product of two of its roots is 12.
8. Solve the equation
3x4 − 40 x3 + 130 x2 − 20 x + 27 = 0
given that product of two of its roots is equal to the product of the other two.
9. Prove that the necessary condition for the roots of the equation
a0 x3 2 + 3a1x2 + 3a2 x + a3 = 0
to be in G.P. is a0 a32 = a13a3
10. Prove that the necessary condition for the root of the equation
x3 − px2 + qx − r = 0
to be in H.P. is 27r 3 − 9r 2 pq + 2rq2 = 0
11. Prove that the necessary condition for the sum of two roots of the equation
x − px3 + qx 2 − rx + s = 0
to be equal to the sum of the other two is
p3 − 4 pq + 8r = 0
12. Solve the equation
2 x − 15x + 35 x2 − 30 + 8 = 0
given that the product of two roots is equal to the product of the other two.
13. Solve the equation
x4 − 8x3 + 21x2 − 20 x + 5 = 0
given that the sum of two roots is equal to the sum of the other two.
14. Solve the equation
2 x4 − 15x + 35x2 − 30 x + 8 = 0
given that the product of the two roots is equal to the product of the other two.

70
6.3 NOW WE DISCUSS CERTAIN IMPORTANT THEOREMS REGARDING THE COMPLEX
ROOTS AND SURD ROOTS OF AN EQUATION
Theorem 1. In an equation with real coefficient complex roots occur in conjugate pairs.
Proof. Let f(x) = 0 ... (1)
be an equation, with real co-efficients. Let a + ib be a root of (1). Then we have to prove that a – ib is also
a root of (1).
Since a + ib is a root of (1), therefore, the factor of f(x) corresponding to the root a + ib is
[x – (a + ib)]
Now consider the product
[x – (a + ib)] [x – (a – ib)]
or [(x – a) – ib] [(x – a) + ib]
or [(x – a)2 + b2]
We divide f(x) by the second degree polynomial [( x − a)2 + b2 ]. The remainder will be a
polynomial of degree one. Also since the dividend f(x) and the divisor (x – a)2 + b2 are polynomials with
real coefficient the quotient and remainder will also be polynomials with real coefficient. Let the quotient
be denoted by f(x) and the remainder by rx + rx where r and r are real numbers. We have numbers, thus
f(x) = [( x − a)2 + b2 ] ( x) + (rx + r ) ...(i)
We put x = a + ib on the two dies and obtain
0 = 0 + (a + ib) r + r
or 0 = (ra + r) + irb
Equating the real and imaginary party to zero, we obtain
ra + r = 0 ...(ii)
rb = 0 ...(iii)
Now b  0 for if b = 0, a + ib would be real so we can suppose that b  0.
Therefore from (iii), r = 0.
Finally from (ii), r = 0
Substituting r = 0, r = 0 in (i), we obtain
f(x)  [( x − a)2 + b2 ] ( x)
 [ x − (a + ib)][ x − (a − ib)( x)]
Putting x = a – ib, on the two sides, we see that the R.H.S is zero and hence the L.H.S. is zero.
Thus a – ib is a root of f(x) = 0
Hence the result.
Theorem 2. In an equation with rational coefficients, surd roots occur in pairs, i.e., if p + q ,
where p and q are rational number and q is not the square of a rational number, is a root of an equation
with rational coefficients then p – q is also a root of the same equation f(x) = 0 with rational numbers.

Proof. We have [ x − ( p q )][ x − ( p − q )]

= [( x − p) ] − q ][( x − p) + q )]
= [(x – p)2 – q]

71
 We divide f(x) by [(x – p)2 – q]. Since the coefficients in the divident f(x) and the divisor
[(x – p]2 – q)] are all the rational, the co-efficients in the quotient and the remainder will also be
ratational. Moreover the remainder will be of the first degree.
We write
f(x) = [(x – p)2 – q] (x) + rx + r. ...(i)
where r and r are rational numbers.
Putting x = p + q on the two dies of (i), we obtain

0 = r( p + q ) + r

= (rp + r ) + r q
Equating the surd part and the non-surd part secondary to zero, we have
r q = 0 and rp + r = 0
From these, we deduce r = 0, r = 0 ( q  0)
Thus f(x)  [(x – p)2 – q] (x)
 [ x − ( p + q ) ( x − ( p − q ) ( x)]

so that x − ( p − q ) is a factor of f(x) and as such ( p − q ) is a root of f(x) = 0.

Hence the result.
Example 1. From an integer with lowest degree whose roots are 2, 3 + 4i.
Solution. Since in an equation with real co-efficients surd roots occur in pairs and complete roots
occur in conjugate pairs.
 Roots of required equation are  2, 3  4i.
Hence required equation is
( x − 2) ( x + 2) [ x − (3 + 4i) [ x − (3 − 4i) = 0
 (x2 – 2) [(x – 3)2] – (4i)2 = 0
 (x2 – 2) (x2 – 6x + 9 + 16) = 0
 (x2 – 2) (x2 – 6x + 25) = 0
 x4 – 2x2 – 6x3 + 12x + 25x2 – 50 = 0
 x4 – 6x3 + 23x2 + 12x – 50 = 0
Example 2. Solve the equation
3x4 − 10 x3 + 4 x2 − x − 6 = 0
1 + −3
some root being
2
Solution. The coefficients of the given equation are all real.
1 + −3 1 3 1 3
Now = +i is a root and accordingly the conjugate complex of this root viz. − i
2 2 2 2 2
is also a root.

72
   1 i 3    1 i 3   1  i 3   1 i 3
Now  x −  +    x −  −   =  x −  −   x −  + 
  2 2     2 2    2 2   2 2 

1   3i  
2 2
 1 3
=  x −  −   =  x −  + = ( x − x + 1)
2
 2  2   2 4
Thus (x2 – x + 1) is a factor of
3x4 + 10x3 + 4x2 – x – 6
Dividing (i) by (x2 – x + 1), we may see that
3x4 + 10x3 + 4x2 – x – 6  (x2 – x + 1) (3x2 – 7x – 6)
7  49 + 72 7  11 
The roots of 3x2 – 7x – 6 = 0 are = , i.e., 3, −
6 6 3
2 1 i 3
Hence the required roots are 3, − ,   .
3  2 2 

73
 Exercise-1
1. Simplify the following:
(i) (2 – 5i) + (–3 + 4i) + (8 + 3i)
5+i
(ii)
3 − 2i
2. Evaluate the following:
(i) (–7 –2i) (–1 – 5i)
(2 + 4i ) 2
(ii)
(3 + i ) 2
3. Solve the following equations:
(i) [(x + 2y), (2x – y – 6)] = (3, 2)
−2 + 3i
(ii) ( x − y) + i( x + y) =
1+ i
4. Express the following complex numbers in the polar form.
(i) –1 + i 3
(ii) –2 − i 12
5. Find the trigonometric representation of
(i) sin  – i cos .
(ii) 1 + cos  + i sin .
6. Show that | cos  + i sin  | = 1.
5+i
7. Express the complex number in the form x + iy where and x and y are real.
2+i
Find its modulus and amplitude.

Exercise-2
(cos  + i sin )8 (cos 2 − i sin 2)9
1. Simplify
(cos 2 i + i sin )9 (cos  − i sin )8
10 10
     
 cos + i sin  +  cos − i sin 
2. 
15 15   15 15 
6
  
 cos + i sin 
 15 15 
(cos  + i sin ) 4
3. Use De Moivre’s theorem to simplify
(sin  + i cos )5

   
4. Prove that (sin x − i cos x)n = cos n  − x  − sin n  − x 
2  2 
5. Obtain the value of (3 + 4i)3 with the help of De Moivre’s theorem.
n
6. In n is a positive integer, show that ( 3 + 1) n + ( 3 − i) n = 2n + 1 cos .
6
7. Prove that

74
 [(cos  + cos ) + i (sin  + sin )]n + [(cos  + cos ) − i (sin  + sin )]n
− n(q + )
= 2n − 1 cosn   cos
 2  2
[Hint : (cos  + cos ) + i (sin  + sin )]
+ −  + −
= 2 cos cos + i  2 sin cos 
2 2  2 2 

Exercise-3
1 1
1. If 2 cos  = x + , 2 cos  = y + show that
x y
1
xm y n + m n
= 2 cos (m + n)
x y
xm y n
and + = 2 cos (m – n),
y n xm
where m and n are integer. What happens if m and n are rational numbers?
2. If x = cos  + i sin , y = cos  + i sin  and z = cos  + i sin , such that x + y + z = 0 then prove
1 1 1
that + + =0
x y z
1
and xyz + = 2cos (pa + qb + rc)
xyz
3. Let the complex numbers x, y, z be given respectively by cos a + i sin a, cos b + i sin b and cos c +
i sin c, then prove that for any integers p, q, r
1
x p yq zr + p q r
= 2 cos (pq + qb + rc)
x y z
4. Let x1 + iy1; x2 + iy2 , ...... xn + iyn be any n complex number and A + Bi be some other complex
number, such that
( x1 + iy1 ) ( x2 + iy2 ) ...... ( xn + iyn ) = A + iB
then show that
y1 y yn B
(i) tan −1 + tan −1 2 + + tan −1 = tan −1
x1 x2 xn A

(iii) ( x12 + y12 ) ( x22 + y22 ) + ( xn2 + yn2 ) = A2 + B2

5. Expand (1 + i)n in two different ways and find the sum of the two series.
(1 – nC2 + nC4 − nC6 + ) and (n C1 − nC3 + nC5 ))
both the series being finite.
[Hint : First expand (1 + i)n by Binomial theorem and then expand it
by the help of De Moivre’s Theorem.]

75
 Exercise-4
1. Find the values of
1
(i) (1 + i) 3
1
(ii) (–1 + i 3) 4
1
(iii) (−1) 6
(iv) (1 + i)2 / 3
2. Solve the equations (i) z7 = 1 (ii) z7 + z = 0.
3. Find the nth roots of unity and show that they form a series in G.P. whose sum is zero. Also prove
that the sum of their pth power always vanishes unless p be a multiple of n, p being an integer and
that then the sum is n.
4. Solve the equation z 7 + z  + z  + 1 = 0.
[Hint : z 7 + z 4 + z3 + 1 = ( z 4 + 1) − ( z  + 1) = 0]
5. Use De Moivre’s theorem to solve
z10 + 11z5 + 1 = 0
[Hint : Put z5 = y]
6. Solve the equation
z 4 − z3 + z 2 − z − 1 = 0
[Hint : Mutiply the equation by (z – 1)]
7. Solve the equation z12 – 1 = 0 and find which of its roots satisfy the equation
z 4 + z  + 1 = 0.

Exercise-5
1. Express sin 5 in terms of sin  only.
2. Express cos 6 in terms of cos  only.
3. Express tan 7 in terms of tan  only.
4. Find the value of sin 1º approximately.
sin  1013
5. Find the value of , where = .
 1014
6. Show that sin 5 = 16 sin 5 – 5 sin 3 + 10 sin .
7. Express sin7  cos2  in terms of sines of multiples of .
8. Express cos9  in terms of cosines of multiples of .

Exercise-6
1. cos  + cos 2 + cos 3 + ......
2. sin  · cos 2 + cos 2 · sin 3 · cos 4 + ......
3. sin  + sin ( + ) + sin ( + 2) + ......

76
 4. 2 cos 2 + 22 cos 4 + 23 cos 6 + ......
Find the sum to infinity of the following series.
1 1 1.3
5. 1 + cos 2 − cos 4 + cos 6 + ......
2 2.4 2.4.6
1 1
6. sin  − sin 2 + sin 3 + .....
2 3
7. sin  · sin  + sin2  · sin 2 + sin3  · sin 3 + ......
sin 2  sin 3 
8. 1 + sin  cos  + cos 2 + cos 3 + ......
2 3

cos 2  cos3 
9. 1 − cos  cos  + cos + cos 3 +
2 3

sin ( + 2) sin ( + 4)

10. sin  − + +
2 4

Exercise-7
1. , ,  are the roots of the equation
x3 − 5x2 + x + 12 = 0
Calculate the value of 2.
2. If , ,  are the roots of the equation x3 + qx + r = 0, prove that
3(2 )(5 ) = 5(3 )(4 ).
[Hint: Since , ,  are the roots] of the given cubic equation
3 + q + r = 0
3 + q + r = 0
3 + q + r = 0
To find 4 multiply equations (1) by , ,  respectively and add.
To find 5 multiply equation by (2, 2, 2 respectively and add.
3. Find the sum of the cubes of the roots of the equation
x3 – 6x2 – 11x + 6 = 0
4. If , ,  be the roots of the equation
x3 +5x2 – 6x + 3 = 0
Find the values of –3.
 −3 1 1 1 3  3 
 Hint :   + + = 
  3
3
 3
 
3 2 2

 Again 33 – 3( ) 2 =  [ 22 −  ]
 

77
5. If a, b, c, be the roots of the cubic x3 + qx + r = 0
1
(i)  ,
b+c
b2 + c2
(ii)  ,
b+c
(iii) (b + c)2
6. If , , ,  be the roots of the equation
x4 + px3 + qx 2 + rx + s = 0
Find the value of
(i) 2
1
(ii)  .

7. If , ,  are the roots of the equation
x3 − px2 + qx − r = 0
Find the value of
1
(i)  2 ,

(ii) 2
(iii) ( + ) ( + ) ( + )

Exercise-8
1. From an equation whose roots are negatives of the roots of the equation
x3 − 6 x 2 + 8 x − 9 = 0
2. From an equation whose roots are twice the roots of the equation
2 x 4 + 3x 2 + 7 x − 9 = 0
3. Find the equation whose roots are the squares of the roots of the equation
x3 – x2 – 8x – 6 = 0
4. Find the equation whose roots are the reciprocals of the roots of the equation
ax3 + bx2 + cx + d = 0.

Exercise-9
1. If , ,  are the roots of the equation
x3 + px2 + qx + r = 0
form the equation whose roots are
1  1
a− ,− ,  −
  

78
2. If a, b, c are the roots of x3 + qx + r = 0, from the equation whose roots are
b  c 2 , c 2 a 2 , a 2b  .
3. Find the equation whose roots are the roots of the equation
x4 – 3x2 + 7x2 + 5x – 2 = 0
each diminished by 2.
4. Transform the equation
x3 − 6 x 2 + 4 x − 7 = 0
into one, from which the second term is missing.
5. If , ,  are the roots of the equation
x3 − px2 + qx − r = 0, where r  0
1 1 1
Find the equation whose roots are  + ,  + ,  +
  

Exercise-10
1. Solve the equation
3x4 − 40 x3 + 130 x2 − 20 x + 27 = 0
given that product of two of its roots is equal to the product of the other two.
2. Prove that the necessary condition for the roots of the equation
a0 x3 + 3a1x2 + 3a2 x + a = 0
to be in G.P. is
a0 + a23 = a13a3
3. Prove that the necessary condition for the roots of the equation
x3 – px2 + qx – r = 0
to be in H.P. is
27r 2 − 9 pqr + 2q3 = 0.
4. Prove that the necessary condition for the sum of two roots of the equation
x4 – px3 + qx2 – rx + s = 0
to equal to the sum of the other two is,
p3 − 4 pq + 8r = 0.

Exercise-11
1. Solve the equation
x3 − 6 x2 + 3x + 10 = 0
whose roots are in A.P.
2. Solve the equation
3x3 − 26 x2 + 52 x − 24 = 0
the roots being in G.P.

79
3. Solve the equations
28x3 – 39x2 + 12x – 1 = 0
the roots being in Harmonical Progression.
4. Solve the equation
4x3 + 16x2 – 9x – 36 = 0
5. Solve the equation
2x3 – x2 – 22x – 24 = 0
two of its roots being in the ratio 3 : 4
6. Solve the equation
2x3 + x2 – 7x – 6 = 0
given that the difference of two roots is 3.
7. Solve the equation
x3 – 5x2 – 2x + 24 = 0
given that the product of two if roots is 12.

80
 UNIT 2

LESSON 1

MATRICES
(BASIC CONCEPTS)

1. INTRODUCTION
You are already familier with addition and multiplication of matrices. We shall now talk about some
important types of matrices such as symmetric and skew-symmetric matrices, hermitian and skew-
hermitian matrices etc., elementary operations on a matrix inverse of a matrix, rank of a matrix, and
characteristic equation of a matrix. In the end we shall apply some of these concepts to solutions of
systems of lineare equations. However, before we do so, we shall briefly recapitalate the main facts about
addition and multiplication of matrices.

2. DEFINTION OF A MATRIX
Let S be any set. A set of mn elements arranged in a rectangular array of m rows and n column as
 a11 a12 a1n 
 a21 a22 a2 n 
 
 
 am1 am 2 amn 
is called an m × n (“m by n”) matrix over S. A matrix may be represented by the symbols || aij ||, [aij], [aij]
or by a single letter such as A. The aij’s in a matrix are called the element of the matrix. The indices is and
j of an element indicate respectively the row and the column in which the elements aij is located.
Since we shall be dealing only with matrices over the set of complex number therefore, we shall use
the word “matrix” so as to mean “matrix over C” throughtout, unless we state to the contrary.
Thes 1 × n matrics are called row vectors and the m × l matrices are called column vectors. The m ×
n matrix whose elements are 0 is called the null matrix (or zero matrix) of the type m × n. It is usually
denoted by Om × n or simply by O if there is no possibility of confusion.
If the number of rows and the number of columns of a matrix are equal (say each equal to n) the
matrix is said to be a square matrix of order n or an n-row square matrix. The elements a11 a22, ... amn of a
square matrix A are said to constitute the main doagonal of A. A square matrix in which all the off
diagonal elements are zero is called a diagonal matrix. Thus an n-rowed square matrix [aij] is a diagonal
matrix if aij = 0 whenever i + j. An n-rowed diagonal matrix is often written as
dia. [a11, a22, ..., amn]
A diagonal matrix in whic all the diagronal elements are equal is called a scalar matrix. In other
words, an n-rowed square matrix [aij] is a scalar matrix if for some number k.

aij = 
k , when i = j,
0, when i  j.
A scalar matrix in which each diagonal element is unity, is called a unit matrix. Thus, an n-rowed
square matrix [aij] is called a unit matrix if

81
 aij = 
1, whenever i = j,
0, whenever i  j.
The n-rowed unit matrix is usually denoted by In (or simply by I if there is no possibility of
confustion).
The matrix of elements which remain after deleting any number of rows and columns of a matrix A
is called a sub matrix of A.
Illustrations :
0 0 0 0
1.  0 0 0 0  is the 3 × 4 null matrix.
0 0 0 0
 
 3 1 2
2.  5 4 7  is a 3-rowed square matrix. 3, 4, 8 constitute the main diagonal of this matrix.
 −1 2 8 
 
1 0 0 
3.  0 7 0  is a 3-rowed diagonal matrix.
 0 0 −2 
 
1 0 0 
4.  0 7 0  is a 3-rowed scalar matrix.
 0 0 −2 
 
1 0 0
5.  0 1 0  is the 3-rowed unit matrix. We denote it by I .
0 0 1 3
 
1 8 7
6. The matrix 
 3 4  is submatrix of  −2 3 4  because it can be obtained from the latter by

 6 −5   1 6 −5 
 
deleting the first row and the first columne.

3. EQUALITY OF MATRICES
Two matrices A = [aij] and B = [bij] are said to be equal if (i) they are comparable, i.e., the number
of rows in B is the same as the number or rows A, and the number of columns in B is the same as the
number of columns in A; (ii) aij = bij = for every pair of subscripts i and j. Thus, for example, the matrices
 3 7  and  1 5 4  are not comparable; the matrices  −1 2 3  and  4 3 6  are comparable but
8 9 3 6 2  3 1 0 1 8 9
       
 2 4 7  and  4 4 7 
not equal; the matrices     are equal.
 6 3 −1  2.3 9 −1
From the definition of equality of matrices, it can be easily verified that if A, B, and C be any
matrices, then
(i) A = A (reflexivity)
(ii) A = B  B = A (symmetry)
(iii) if A = B and B = C, the A = (transitivity)

82
 The above statements (i)—(iii) can be summed up by saying that the relation of equality in the set of
all matrices is an equivlance relation.

4. ADDITION OF MATRICES
If A = [aij], and B = [bij] be two matrices of the same type, say m × n, their sum is the m × n matrix C
= [cij], where cij = aij + bij for every pair of subscripts i and j. In other words,
 a11 a12 a1n   b11 b12 b1n 
a a22 a2 n  b b22 b2 n 
If A =  21  B =  21 
   
 a m1 am 2 amn   bm1 bm 2 bmn 
 
 a11 + b11 a12 + b12 a1n + a1n 
then A + B =  a21 + b21 a22 + b22 a2n + b2n 
 
 am1 + am1 am 2 + bm 2 amn + bmn 

 −2 1 3   4 2 5
Illustrations. If A= 
−  and B =  3 0 6 
 4 2 1   
 −2 + 4 1 + 2 3 + 5   2 3 8 
then A+B=  = 
 4 + 3 2 + 0 −1 + 6   7 2 5 
Properties of matrix addition
Addtion of matrices, has the following properties:
(i) Addition of matrices is associative. That is, if A, B, and C be matrices of the same type, then A
+ (B + C) = (A + B) + C
(ii) Addition of matrices is commutative. That is, if A and B be matrics of the same type, then A +
B = B + A.
(iii) Property of zero matrix. If A be an m × n matrix and 0 denotes the m × n zero matrix, then A +
0 = 0 + A = A.
(iv) Negative of a matrix. If a be an m × n matrix, there exists an m × n matrix B, called the
negative of the matrix A, such that A + B = B + A = 0.

5. MULTPLICATION OF A MATRIX BY A SOLAR

If A = [aij] be an m × n matrix, and k be any complex number, then kA is defined to be the m × b
matrix whose (i, j)th the elements is k aij. The matrix kA is called the scalar multiple of A by k. The
following properties of scalar multiplication are worthnothing :
(i) If A and B are comparable matrices, and k is any complex number, then
k(A + B) = kA + kB.
(ii) If A be any matrix, and k and l be any two complex numbers, then
(k + 1) A = kA + lA.
(iii) If A be any matrix, and k and l be any two complex numbers, then
k(lA) = (kl) A.
(iv) For every matrix A,
1 A =A

83
6. MULTIPLICATION OF MATRICES
Definition 1. Let A = [aij] and B = [bij] be m × n and n × p matrices respectively. The m × p matrix
[cij], where
n
cij = aij, bij + ai2 + b2j + ain bnj =  aij bkj ,
k =1

is called the product of the matrices A and B and is denoted by AB.

The above definition expresses two facts:
(i) We can talk of the product AB if two matrices of and only if the number of columns of A is
equal to the number of rows of B. In the case this condition is satisfied, we say that A and B are
conformable to multiplication.
(ii) If A and B are conformable to multiplication, the (i, j)th element of the matrix AB is obtained by
multiplying the elements of the ith row of A by the corresponding elements of the jth column of
B and adding the products. The sum so obtained is the desired (ij)th elements of AB.
Properties of matrix multiplication
The following are same of the important properties of matrix mulplication :
(i) Matrix multiplication is associative. That is, if A, B, and C be of suitable sizes for the products
A(BC) and (A + B) C to exist, then A(BC) = (AB) C.
(ii) Matrix multiplication is not commutative. That is, given two matrices A and B, AB = BA is not
always true. It is important to note here that for pair of matrices A and B several different
possibilities arise.
(a) Neither of the products AB and BA exits;
(b) only one of the products AB and BA exist and the other one does not exist;
(c) both AB as well as BA exist but they are of different type;
(d) both AB as well as BA exist and are of the same type, but are not equal
(e) AB = BA.
All the above possibilities do exist for certain pairs of matrices.
The important thing to note is that the phase ‘matrix multiplication is not cummutative’ means
that AB is not always equal to BA. It does not exclude the possibility of AB and BA being
equal in some cases.
(iii) Multiplication of matrices is distributive with respect to addition, i.e.,
A(B + C) = AB + BC
and (B + C)D = BD + CD
where A, B, C and D are of the suitable sizes for the above relations to be meaningful.
(iv) Multiplication by the unit matrix. If A be any m × n matrix then Im A = A = A In

7. POSITIVE INTEGRAL POWERS OF SOURCES MATRIX

If A be an n-rowed square matrix, and n be a positive integer, then An is defined by setting A1 = A,
Ak + 1 = AkA. By the principle of finite induction this defines An for all positive integer An.
If A be an n-rowed square matrix and p and q be positive integers, it can be easily shown that
Ap · Aq = Ap + q, (Ap)q = Apq.

84
8. TRANSPOSE OF A MATRIX
Consider the matrices
 3 6
 3 1 −4 
A= , B =  1 0 .
6 0 7   −4 7 
 
The matrix A is a 2 × 3 matrix, and the matrix B is a 3 × 2 matrix. Also, the first column of B is the
same as the first row of A, and the second columne of B is the same as the second row of A. In other
words, B is the matrix obtained from A by writing the row of A as columns. We say that the matrix B is
the transpose of A.
Defintion 2. If A = [aij] be an m × n matrix, then the n × m matrix B = [bij], such that bij = aji is
called the transpose of A and is deonted by At.
From the above defintion we find that
(i) the transpose of an m × n matrix is an n × m matrix ;
(ii) the (i, j)th element of At is the (j, i)th element of A.
 3 −1 2   −1 3 5 
Example 1. Let A =   and B =  6 2 1  .
 0 4 7   
Compute At, (At)t, Bt, (A + B)t, At + Bt, (3A)t, 3At.
 3 0
At =  −1 4  (At)t =
 3 −1 2 
Solution.
 2 7 0 4 7
 
 
 −1 6 
Bt =  3 2  .
 5 1
 
 2 6  2 6
2 2 7    2 6 .
A+B=  , ( A + B ) t
= 2 6 , At
+ B t
=
6 6 8 7 8 7 8
   
9 0 9 0
 9 −3 6 
3A =   , (3 A)t =  −3 12  , 3 At =  −3 12  .
 0 12 21  6 21  6 21
   
Remark. In the above example we find that ( At )t = A, ( A + B)t = At + Bt , and (3A)t = 3At. These
results are only special cases of the following theorem :
Theorem 1. If At and Bt are transposes of A and B respectively, then
(i) (At)t = A
(ii) C A + Bt = At + Bt, if A and B are comparable.
(iii) kAt = kAt, k being any complex number
Proof. (i) Let A = [aij] be an m × n matrix. Then At is an n × m matrix and (At)t is an m × n matrix.
The matrices (At)t and A are, therefore, comparable.
Also, (i, j)th element of (At)t
= (j, i)th element of At
= (i, j)th element of A
Since the matrices (At)t and A are comparable and their (i, j)th elements are equal, threfore, (At)t = A.

85
 (ii) Let A = [aij] and B = [bij] be m × n matrices. Since A and B are both m × n matrices, therefore A
+ B exists and is an m × n matrix. Consequently (A + B)t is an n × m matrix, therefore
Again, At and Bt are both n × m matrices, so that At + Bt also exists and is an n × m matrix.
The matrices (A + B)t and At + Bt are both of the type n × m, and are therefore comparable.
Also, (i, j)th element (A + B)t = (j, i)th element of A + B
= aji + bji
= (j, i)th element of At + (i, j)th element Bt
= (j, i)th element of (A + B)t
Thus the matrices (A, B)t and At + Bt are comparable, and their corresponding elements are equal.
Hence (A + B)t + At + Bt .
(iii) Let A = [aij] be an m × n matrix. kA is an m × n matrix and therefore (kA)t is an n × m matrix.
Also At being an n × m matrix, kAt is an n × m matrix. The matrices (kA)t and kAt are both of type n × m,
and are, therefore, comparable. Also (i, j)th element of (kA)t = (j, i)th element of kA
= kaji
= k[(i, j)th elements of At]
Since the matrices (kA)t and kAt are comparable and their (i, j)th elements are equal, therefore
(kAt) = kAt.
Remark. If At = B, the Bt = (At)t = A, i.e., if B is the transpose of A, then A is the transpose of B.
1 3 4
 3 −1 2 
Example 2. If A =   and B =  −2 1 −1 ,
 1 0 −3   0 −4 2 
 
Compute (AB)t and BtAt.
1 3 4
 3 −1 2  
Solution. AB =   −2 1 −1
 1 0 −3   0 −4 2 
 
 5 0 17 
= ,
 1 15 −2 
5 1 
so that (AB)t =  0 15 
 17 −2 
 
 1 −2 0  3 1
Also, Bt =  3 1 −4  , At =  −1 0  .
 4 −1 2   2 −3 
   
5 1 
Therefore Bt At =  0 15  .
 17 −2 
 
Remark. In the above example (AB)t = BtAt. This is of course, only a particular case of the general
result which we state and prove in the following theorem.
Theorem 2. If A and B be matrices conformable to multiplication, then
(AB)t = BtAt

86
 Proof. Let A = [aij] and B[bij] be m × n and n × p matrices particularly.
The At = [cij], where cij = aji, is an n × m matrix.
Bt = [cij], where dij = bji, is an p × n matrix. The matrices (AB)t and BtAt are
both of type of p × m, and are therefore comparable.
Also (i, j)th elements of (AB)t
= (j, i)th element of (AB)t
n
=  a jk bki
k =1
n
=  a jk bki
k =1
n
=  dij ckj
k =1

= (i, j)th element of BtAt

Since the matrices (AB)t and BtAt are of the same type, and their (ij)th elements are equal, therefore,
(AB) = BtAt
t

9. SYMMETRIC AND SKEW-SYMMETRIC MATRICES

Consider the matrices
3 0 2  0 1 3
A = 0 4 −1 , B =  −1 0 −4 
 
 2 −1 5   −3 4 0 
   
In matrix A, (1, 2)th element is equal to (2, 1)th element, (1, 3)th element is equal to (3, 1)th element,
and (2, 3)th element is equal to (3, 2)th element. Because of these properties we say that the matrix A is
symmetric.
In matrix B, (2, 1)th element is the negative of (1, 2)th element, (3, 1)th element is the negative of the
(1, 3)th element, (3, 2)th element is the negative of the (2, 3)th element, and (1, 1)th element, (2, 2)th
element, and (3, 3)th element are own negatives, (i.e., they are all zero). Because of these properties we
say that the matrix B is skew-symmetric.
Symmetric and skew-symmetric matrices play useful (an important) roles in the theory of matrices.
Definition 3. A square matrix A = [aij] is said to be symmetric if aij = aji for all i and j.
Illustrations 1. The matrices
 4 1 − i 2  a h g 1 0 0
1 − i 3 7  h b f and  0 1 0  are all symmetric.

 2 7 i   g c  0 0 1
 f  
2. The matrices
 0 i 1 + i  0 3 6 − i
 −i 0 
−3 and  −3 0 −4  are both skew-symmetric
 −1 − i 3   0 
 0  −6 + i 4

87
 3. The matrices
 1 i 1 + i  0 1 −2 
 −i 2 6i  and  1 0 i  are neither symmetric nor skew-symmetric.
 −1 − i 6i   0 
 3  2 −i
In the following theorem we state and prove some basic facts about symmetric and skew-symmetric
matrices.
Theorem 3.
(i) A necessary and sufficient condition for a matrix A to be symmetric is that At = A.
(ii) A necessary and sufficient condition for a matrix A to be skew-symmetric is that At = –A.
(iii) The diagonal elements of a skew-symmetric matrix are all zero.
Proof. (i) Necessity. Let A = [aij] be a symmetric matrix. Since A is symmetric, it must be square
matrix, say of order n, At is then also of order n, so that At and A are comparable. Also, (i, j)th element of
At = aji = aij = (i, j)th elements of A.
Therefore At = A.
Sufficiently. Let A = [aij] be an m × n matrix such that At = A. Since A is an m × n matrix., therefore
At is an m × n matrix. Since At and A are equal matrices, they are comparable, so that n = m, and
consequently A is a square matrix. Also, as given, (i, j)th element of At = (i, j)th, which gives aji = aij.
Since A a square matrix such that aij = aij for all i and j, therefore A is symmetric.
(ii) Necessity. Let A = [aij] be a skew-symmetric matrix. Since A is skew-symmetric, it must be a
square matrix, say of order n. At is then also of order n, so that At and A are comparable. Also, (i, j)th
element of At = aji = –aij = (i, j)th element of –A threfore At = A.
Sufficiently. Let A = [aij] b an m × n matrix such that At = –A. Since A is m × n matrix, therefore At is
an n × m matrix and –A is an m × n matrix. Since the matries At and –A are equal, they are comparable, so
that n = m, and consequently A is a square matrix. Also (i, j)th element of At = –[(i, j)th element of A],
which gives aji = –Aij.
We shall now state and prove a theorem which assures us that every square matrix can be expressed
uniquely as a sum of a symmetric and skew-symmetric matrix.
Theorem 4. Every square matrix can be expressed uniquely as the sum of a symmetric and a skew-
symmetric matrix.
Proof. Let A be an n-rowed square matrix.
Let A = X + Y,
where X is an n-rowed symmetric, and Y is an n-rowed skew-symmetric matrix. Taking the
transpose of both sides of (i), we have
At = ( X + Y )t = X t = Y t = X − Y ... (2)
since X is symmetric and Y is skew-symmetric so that Xt = X and Yt = –Y
From (1) and (2), we get
1
X= ( A + At ), ... (3)
2
1
Y= ( A − At ) ... (4)
2

88
 We have shown that if A is expressible as the sum of a symmetric matrix X and a skew-symmetric
matrix Y, then X and Y must be given by (3) and (4). This establishes the uniqueness part. To demonstrate
the existence of a symmetric matrix, X and a skew-symmetric matrix Y such that A = X + Y, we have only
to see that if we write.
1 1
X= ( A + At ), Y = ( A − At ),
2 2
t
1 
then Xt =  ( A + At )  ,
2 
t
1 
=  ( A + At )  ,
2 
1 t
= ( A + A),
2
= X,
t
1 
Yt =  ( A − At ) 
2 
t
1 
=  ( A − At ) 
 2 
1 t
= ( A − A)
2
= –Y,
so that X is an n × n symmetric matrix and Y is an n × n skew-symmetric matrix. Furthermore, X + Y
= A, which completes the proof.
Example 3. Express the matrix
2 6 5
A =  3 1 4
 9 −1 7 
 
as the sume of symmetric and a skew-symmetric matrix.
Solution. Let
2 6 5
 3 1 4  = X + Y,
 9 −1 7 
 
where X is a 3-rowed symmetric matrix and Y is a 3-rowed skew-symmetric matrix.
Taking transposes of both sides (1), and using the facts that (X + Y)t, = , we have
 2 3 6
 6 1 −1  = X – Y.
5 4 7
 
From (1) and (3) we find that
 2 3 6  0 3 −4 
2X =  6 1 −1  2Y =  −3 0 5  ,
5 4 7  4 −5 0 
   

89
 so that
 9   3 
2 7  0 −2 
2 2
9 3  3 5
X= 1 , Y =  − 0 
2 2  2 2
 3   5 
7 7  2 − 0
 2   2 
Verification. Since Xt = X, Yt = Y, therefore that matrix X is symmetric and the matrix Y is skew-
symmetric. Also, by actual addition we find that X + Y = A.

10. HERMITIAN AND SKEW-HERMITIAN MATRICE

Let A = [aij] be an m × n matrix. The m × n matrix B = [bij] such that bij = is called the conjugate of
A and is denoted by A. The transpose of A is called the tranposed conjugate or transjugate of A and is
enoted byA.
For example
A=
 3 + 2i 4i 
4i 
If ,
 − 7 3 −

=
 3 − 2i − 4i 
then ,
 − 7 3 + 4i 
 3 − 2i
A = 
− 7

 − 4i 3 + 4i 
It can be easily the seen that :
(i) if A be any matrix, then (A) = A.
(ii) if and B be two matrices conformable to addition, then
(A + B) = A + B
(iii) if A be any matrix and k be any complex number, then
(kA) = kA .
(iv) if A and B be two matrix conformable to multiplication, then
(AB) = BA.
We have the proofs of (i), (ii) to the reader and prove only (iii).
Let A = [aij] be an m × n matrix and let B = [bij] be an n × p matrix.
The matrices (AB) B, A are respectively of types p × m, p × n, and n × m, so that BA is a p × m
matrix.
Since (AB) and BA are both the type p × m, therefore they are comparable. Also (ij)th element of
 n  n
(AB) = (j, i)th element of ( AB ) =   a jk bki  =  a jk bki
 k =1  k =1
 
n
Also, (i, j)th element of BA =  [(ik )th element of B  (kj )th element of A ]
k =1
n
=  [(ki)th element of B   ( jk )th element of A]
k =1

90
 n
=  bki a jk
k =1
n
=  akl bki ,
k =1

which is the same as the (i, j)th element of (AB).

Since (i) (AB) and BA are matrices of the same type, and (ii) (ij)th elements of (A, B) and BA
are equal for all i and j, therefore (AB) = BA.
Defintion 4. A square matrix A = [aij] is said to be Hermitian if for all i and j A square matrix A
[aij] is said to be skew-hermitina if for all i and j.
Illustration 1. The matrices
 2 −i   1 3 + 4i 
 i 3  ,  3 − 4i − 4 
are Hermitian.
  
2. The matrices
 2i 1 + 3i   1 5 − 7i 
 −1 + 3i  ,  −5 − 7i 3i 
are skew-hermitian.
 5 i  
 i 4 + i
3. The matrix 
1
is neither Hermitian nor skew-hermitian.
4 − i
In the following theorem we prove the important facts about hermitian and skew-hermitina matrices.
Theorem 5. Let A be an n-rowed square matrix.
(i) A necessary and sufficient condition for A to be hermitina is that A = A.
(ii) A necessary and sufficient condition for A to be skew-hermitian is that A = –A.
(iii) The diagonal elements of a hermitian matrix are all real, and the diagonal elements of a skew
hermitian matrix are either pure imagineries of zero.
Proof. (i) First let us assume that A is hermitian. Since A is an n-rowed square matrix, therefore A
(= ( A)t ) is also an n-rowed square matrix. Consequently A and A are matrices of the same type. Also, (i,
j)th element of A (j, i)th element of A = a ji = aij = (i, j ) th element of A. Hence A = A. Conversely, let us
suppose that A = A  (i, j )th element of A = a ji . Since A = A, therefore a ji = aij , i.e., aij = (a ji ) and
consequently A is hermitian.
(ii) First lets assume that A is skew-hermitian. Since A is an n-rowed square matrix, therefore, A
and –A are both n-rowed square matrices, and as such they are comparable. Also (i, j)th element of A = (j,
i)th element of A = a ji , and (i, j ) th element of –A = –aij. Since A is skew-hermitian, therefore a ji = − aij .
Hence (i, j)th element of A = (i, j)th element of –A.
Since A and –A are comparable, and their (i, j)th elements are equal for all i and j, therefore A = –
A.
Conversely, let us suppose that  = –A then (i, j)th elements of A and –A must be equal for all i and
j. This gives a ji = − aij . Taking conjugates of both sides, we have a ji = − aij , i.e., aij – a ji , so that A is
skew-hermitian.

91
 (iii) Let A be a hermitain matrix. then aij = a ji for all i and j. For a diagonal elements j = i, which
gives aij = a ji for all i, i.e., aij thus real. Thus the diagonal elements of A are all real.
If A is skew-hermitian, aij = − a ji for all i, j. For j = i this gives aij = − a ji . If aij = p + iq, then aji = p
– iq and so we have (p – iq) = –(p + iq) i.e., 2p = 0 or p = 0. The real part of aij is, therefore, zero and
consequently aji is either pure imaginary (if q  0).
We shall now state and prove a decomposition theorem which tells us that every square matrix can
be uniquely decomposed as the sum of the hermitian and a skew-hermitian matrix.
Theorem 6. Every square matrix can be express uniquely in the form X + iY, where X and Y are
hermitian matrices.
Proof. Let A be an n-rowed squrae matrix.
Let A = aij = − a ji .

= X  + (iY )
= X  − iY 
= X − iY ,
Since X = X and Y = Y.
From (1) and (2), we have
1 1
X=( A + A ), Y = ( A − A )
2 2i
We have thus shown that if A is expressivle as X + iY, where X and Y, are hermitian, the XZ and Y
must be given by (3). Thus establishes the uniqueness part.
To demonstrate the existence hermitain matrices X and Y such that A = X + iY, we have only to see
that if we write
1 1
X= ( A + A ), Y = ( A − A )
2 2i

1 
then X =  ( A + A ) 
 2 
1
= ( A + A )
2
1 
= ( A + A), since A = + A.
2
= X,

1 
Y =  ( A − A ) 
 2i 
1 
=− ( A − A )
2i
1 
=− ( A − A), since A = A.
2i
= Y,

92
 so that X and Y are hermitian.
Also, X + Y = A, which completes the proof.
Exampe 4. Express the matrix.
 1 7 1−i
A = 1 + i i 2
 − 3 1 4 + 2i 
 
in the form X + iY, where X and Y are hermitian matrices.
Solution. We know that every square matrix A can be uniquely expressed in the form X + iY, whre
X and Y are hermitian matrices given by
1 1
X= ( A + A ), Y = ( A − A )
2 2i
 1 1−i − 3
Here A =  7 − i 1
1 + i 2 4 − 2i 


 1 1 
 1 4− i −1 − i 
2 2
1  1 3
Therefore X = ( A + A ) =  4 + i 0 
2  2 2
 1 3 
 −1 + i 4
 2 2 
 1 1 
 0 − 3i − − 2i 
2 2
1  1 1 
Y= ( A − A ) =  + 3i 1 − i
2i  2 2 
 1 1 
 − + 2i i 2
 2 2 
Verification. Since X = X and Y = Y, the matrices X and Y are hermitian. Also X + iY = A.
Exercise
1. For each of the following matrices A, verify that At = A :
 1 2 −1   −1 6 −7   a h g
 2 0 −4  ,  6 3 8  ,  h b f
 −1 −4 3   −7 8 −5   g c 
     f

2. For each of the following matrices A, verify that At = –A.

0 i 1  0 3 1 + i   0 −3 2i 
 −i 0 −2  ,  − 3 0 − 4  ,  3 0 4 
 −1 2 0   −1 − i 4 0   −2i −4 0 
  
1 2 −2 
1 
3. If A = 2 1 2  , verify that A At = At = A = I3.
3  2 −2 −1 
 

93
4. If A be any square matrix, verify that the matrix A + At is symmetric and the matrix A —At is
skew-symmetric.
5. If A be any square matrix, prove that A + A, AA, A A are the hermitian and A – A is
skew-hermitian.
 3 1 −7 
6. Express the matrix  2 4 8  as X + Y where X is symmetric and Y is skew-symmetric.
 6 −1 2 
 
7. Find hermitian matrices A and B such that
 1 1+ i 3
A + iB =  4 2 1 − i
 −3 2 + 3i i 


94
 LESSON 2

ELEMENTARY OPERATIONS ON A MATRIX

AND INVERSE OF A MATRIX

1. INTRODUCTION
In this lesson we shall study two important concepts—namely elementary operations on a matrix,
and inverse of a matrix. Both are devices for producing new matrices out of a given matrix. However,
there is one difference. Elementary operations can be applied to a matrix of any type but the process of
finding the inverse can be applied only to some special types of matrices. Furthermore, while it is possible
to produce as many new matrices as we like from a given matrix by means of elementary operations, the
operation of finding the inverse, if applicable, can produce only one matrix from a given matrix.
Both the types of operations that we are going to study have one things in common. Both can be
used for solving a system of linear equations. There is one important connection between the two types of
operations. Elementary operations can be used to find the inverse of a matrix—and that is one reason for
studying them together in this lesson.

2. ELEMENTARY OPERATIONS
Consider the matrices
 −2 3 4   5 0 −6   −6 9 12   −2 3 4 
S =  , A =  −2 3 4  , B =  5 0 −6  , C =  1 6 2  ,
 5 0 −6       
 −2 −6 4   −2 3 16 
E =  , F =  5 0 −6  .
 5 0 −6   
The matrices A, B, C, D, E and F are related to S in as much as :
A can be obtained from S by interchanging the first and second rows.
B can be obtained from S by multiplying the first row of S by 3.
C can be obtained from S by adding 2 times the first row to the second row.
D can be obtained from S by interchanging the first and third columns.
E can be obtained from S by multiplying the second column of S by –2.
F can be obtained from S by adding 4 times the second column of S to the third column.
The operations described above are only examples of operations known as elementary operations.
Definition 1. An elementary operation is an operation of the following types:
Type I. Interchange of two rows of columns
Type II. Multiplication of a row or a column by a non-zero number
Type III. Addition of a multiple of one row or column to another row or column.
An elementary operation is called a row operation or a column operation according as it applies to
rows or columns.
It can be easily seen that if a matrix B is obatined from a matrix A by an E-operation then the matrix
A can be obtained form B by an E-operation of the same type.
It is convenient (and usual) to use the following notation of E-operations :

95
 We shall denote the interchange of ith and jth rows (columns) by R1  Rj (Ci  Cj); multiplication
of ith row (column) by a number k  0 by Ri → kRj (Ci → + kCi); and addition of k times the jth row
(column) to the ith row (column) by Ri → Ri + k Rj(Ci → Ci + kCj).
If a matrix B is obtained form a matrix A by a finite chain of E-operations, we say that A is
equivalent to B and write it is as A ~ B.
Elementary Matrices. A matrix obtained form a unit matrix by a single E-operation is called an
elemenatary matrix of an E-matrix. For example, the matrices.
0 1 0 1 0 0 1 0 0
 1 0 0,  0 3 0,  0 1 0 
0 0 1 0 0 1 3 0 1
     
are the E-matrices obtained from I3 by the E-operations R1  R2 (or C1  C2 ), R2 → 3R2 ) or (C2
→ 3C2) and R3 → R3 + 3R1 (or C1 → C1 + 3C3) respectively.
Remark. It can be easily see that the operations Ri  Rj and Ci  Cj have the same effect on In; Ri
→ kRi and Ci → kCi have the same effect on In; Ri → Ri + kRj and Cj → Cj + kCi have the same effect
on In. It an E-matrix is obtained from In by an E-operation , we say that it is the E-matrix corresponding
to the operation .
Effect of an elementary operation on the product of two matrices.
Before we consider the effect of an E-operation on the product of two matrices, let us consider the
following example.
 3 1 −2  1 2
Example 1. Let A =  1 0 −4  and B =  −1 4  .
 −3 5 6   3 −5 
   
Let C and D be the matrices obtained form A and AB respectively by the E-row operation R1  R2
Compute C, D and CB and show that D = CB.
Solution.
 3 1 −2   1 2 
AB =  1 0 −4   −1 4 
 −3 5 6   3 −5 
  
 −4 20 
=  −11 22  ,
 10 −16 

1 0 −4 
C= 3 1 −2  ,
 −3 5 6 

 −11 22 
D =  −4 20  ,
 10 −16 

1 0 −4   1 2 
CB =  3 1 −2   −1 4 
 −3 5 6   3 −5 


96
  −11 22 
=  −4 20 
 10 −16 
 
= D.
In the above example we find that the E-row operation R1  R2 on the product AB is equivalent to
the same E-operation on the pre-factor A. in other words.
Whether we apply the E-row operation R1  R2 to the matrix A and then post-multiply the resulting
matrix by B, or first multiply the matrices A and B, and then apply the E-row operation R1  R2 to the
product we get the same result.
The above example suggests the following theorem :
Theorem 1. An elementary row operation on the product of two matrices is equivalent to the same
elementary row operation on the pre-factor.
Proof. Let A [aij] and B = [bij] be m × n and n × p matrices respectively. We shall show that if A* be
the matrix obtained form A by an E-row operation, and (AB)* be the matrix obtained from AB by the same
E-row operations, then A * B = (AB)*.
We shall consider the three types of E-rows operations one by one and prove the result in each case.
Type I. Consider the E-row operation Ri  Rj
Let A * and (AB)* be the matrices obtained from A and B respectivelyss by the E-row operation Ri
 Rj since A is of type m × n, therefore A* is also of type m × n, and consequently A * B is of type m ×
p. Also since AB is of type m × p, (AB)* is also of type m × p. The matrices A * B and (AB)* are, therefore
of the same type.
The matrices A* and A differ form each other in the ith and jth rows only, therefore, it follows that A
* B and AB differ form each other in the ith and jth row only. Also (AB)* and AB differ from each other in
the ith and jth row only.
Consequently, the matrices A* B and (AB)* can differ at the most in ith and jth rows only. It follows
that in order to prove the equality of A * B and (AB)* it is enough to show that the ith (as well as jth) row of
the matrices A * B and (AB)* are identical.
Now (i, k)th element of A* B
n
=  a jr brk
r =1

= ( j, k )th element of AB
= (i, k )th element of ( AB)*;
showing that the ith rows of A * B and (AB)* are identical.
Similarly the jth rows of (AB)* and A * B are identical.
Hence A * B = (AB)*
Type II. Consider the E-row operations Ri  kRi (k  0)
Let A* and (AB)* be the matrices obtained from A and AB respectively by the E-row operation
Ri  kRi (k  0).

97
 As in the case of the Type I, (i) the matrices A * B and (AB)* are of the same type (ii) A * B and
(AB)* can differ in the ith row only at the most.
n
=  {(i, p)th element of A*  (p  l )th element of B}
p=1

n
=  kaipbpl
p=1

n
= k  aip bpl
p=1

= k  {(i, l )th element of AB}.

= (i, l )th element of ( AB)*,

so that the ith rows of A * B and (AB)* are identical.
Hence A * B = (AB)*
Type III. Let A* and (AB)* be the matrices obtained from A to AB respectively by the E-row
operation Ri → Ri + kRj. As in the case of Type I, (i) the matrices A * B and (AB)* are of the same type.
(ii) A * B and (AB)* can differ in the ith row at the most. Therefore in order to complete the proof it
is enough to show that the ith rows of A * B and (AB)* are identical.
Now (i, j)th element A * B
n
= {(i, r )th element of A*  (r  l )th element of B}
i =1

n
=  (air + ka jr ) brl
r =1

n n
=  air brl + k  a jr brl
r =1 r =1

= (i, l)th element of AB + k times = (j, l)th element of AB

= (i, l)ss element of (AB)*,
showing that the ith rows of A * B and (AB)* are identical.
Hence A * B = (AB)*
From the above we find that if A* and (AB)* be the matrices obtained from A and AB respectively
by the same E-row operation of any one of the three types, then A * B = (AB)*.
A theorem similar to the above holds for E-column operations.
Theorem 2. An elementary column operation on the product of two matrices is equivalent to the
same column operation on the post-factor.
Proof. Given a matrix M and an E-column operation c, let us denote by c(M) the matrix obtained for
M by the E-column operation c. Also, let us denote by c* (M) the matrix obtained from M by the
corresponding E-row operation.

98
 Let A and B be two matrices of types m × n and n × p respectively. Denoting by c(A) and c(AB) the
matrices obtained from A and AB respectively by the E-column operation c, we have
c(AB) = [c * ( AB)t ]t ,
= [c * ( Bt At ]t ,
= [c * ( Bt ) At ]t
= ( At )t [c * ( Bt )]t
= Ac(B).
This proves the theorem.
Remark. Instead of deducing the theroem 2 from theorem 1, we can prove it by imitating the proof
of theorem 1.
As direct consequences of the above theorems we can prove the following important theorem.
Theorem 3. Every elementary row (resp. column) operation on a matrix is equivalent to pre-
multiplication (resp. post-multiplication) by the elementary matrix corresponding to that theorem.
Proof. Let A = [aij] be an m × n matrix. With the same notation as in the proof of theorem 2, we
have
A* = (ImA)* = I*m A,
from which we find that A* can be obtained from A by pre-multiplying A by the elementary matrix
Again, writing
A = AIn, we have
c(A) = c(Aln)
= Ac(ln),
From which we find that c(A) can be obtained from A by post-multiplying A by the elementary
matrix c(In). Hence the theorem.
Example 2. Reduce the matrix.
 –1 2 1 8 
A =  3 1 −1 0 
 4 2 1 7
 
to triangular from by E-row operations.
Solution.
Step 1. By suitable E-row operations on A well shall reduce A to a matrix in which all elements in
the first column except the first element are zero. We can do this by the E-row operations, R2 → R2 + 3R1
and R3 → R3 + 4R1. These two operationa yield
 −1 2 1 8 
A ~  0 7 −2 24  = B, say
 0 10 5 39 
 
Step 2. By suitable E-row operations, we shall reduce B to a matrix in which all the elements in the
second column except the first two are zero. Since we would like that in this process the elements of the

99
 10
first column remain unaltered, therefore we apply the E-row operations R3 → R3 – R2 to B. This will
7
reduce the (3, 2)th element to zero and will not alter any element of the first column.
We thus get
 
 −1 2 1 8 
B ~  0 7 −2 24  = C , say.
 
 0 0 55 33 
 
 7 7 
The matrix C is a triangular matrix. We thus find that the E-row operations R2 → R2 + 3R1, R3 → R3
10
– 4R1, R3 → R3 – R2 reduce A to the triangular matrix C.
7
Remarks. 1. If a matrix has more than three rows, then we shall have several steps which would be
similar to step 2 of the above example.
2. If one of the elements in the first column is either 1 or — 1, it is convenient but not essential to
bring it ot the (1, 1)th place as the initial step in reducing a matrix to triangular form.
3. If every element in the first column of the given matrix happens to be zero then step 1 of the
above example is not required and we have to start with step 2.
4. If the (1, 1)th element is zero, but there is at least one element in the first column different from
zero, we apply a suitable E-row operation and bring this non-zero element to (1, 1)th place. For example,
in order to reduce the matrix
 0 3 −1 
 − 2 −4 −5 
 3 1 9
 
to triangular form we apply the E-row operation R1  R2 so as to obtain the matrix.
 − 2 −4 −5 
 0 3 −1 
 3 1 9
 
in which the (1, 1)th element is different from zero. We can now effect triangular reduction of this
matrix in the same way as the in the above example.
5. In the illustration in remark 4 above, we could as well have applied the E-row operation R1  R3
to the given matrix to obtain a matrix in which the (1, 1)th element is difficult from zero.
6. Triangular reduction of a matrix is not unique. In fact, if we apply the E-row operation R3 → 7R3
on the matrix C in the above example, then
We get the matrix
 −1 2 1 8 
 0 7 −2 24 
 0 0 55 33 
 
which is also a traingular matrix.
Theorem 4. Every matrix can be reduced to a triangular form by elementary row operations.
Solution. We shall prove the theorem by inductive on the number of rows.

100
 Let A = [aij] be an m × n matrix.
The theorem is trivially true when m = 1, for every matrix having only one row is a triangular
matrix.
Let us assume that the theorem holds for all matrices having (m – 1) rows, i.e., every matrix having
(m – 1) rows can be reduced to triangular form by E-row operations. We shall show that A can be reduced
to triangular form by E-row operations.
Three different cases arise according as (i) a11  0, (ii) a11 = 0 but a11  0 for some i, (iii) a11 = 0 for
all i. We shall consider these cases onle by one.
Case (i). Let a11  0. The E-row operation R1 → a11–1 R1 reduces A to an m × n matrix B = [bij] in
which b11 = 1. The E-row operations Rf → Rf – bf1 R1 (f = 1, 2, ..., m) reduce B to an m + n matrix C in
which Cf1 = 0 whenever f > 1. The matrix C is of the form
1 c12 c12 c1n 
0 c22 c23 c2 n 
 
 
0 cm 2 cm3 cmn 
By our hypothesis, the (m – 1) rowed matrix
 c22 c23 c2n 
 
c cmn 
 m2 cm3
can be reduced to triangular from by E-row operations. If we apply the corresponding E-row
operations to C, it will be reduced to triangular form.
Case (ii). Let a11 = 0, but ali  0 for some f such that 1  f  m. By applying the E-row operation R1
→ R1, to A we get a matrix D = [dij] in which d11 = a11  0.
We can now proceed as in case (i) and reduce D to triangular form by E-row operations.
Case (iii). If ail = 0 for all i such that l  i  m, (i.e., all the elements in the first column are zero),
then A is of the form
0 a12 a13 a1n 
0 a22 a23 a2 n 
 
 
0 am 2 am3 amn 
By hypothesis we can reduce the (m – 1) - rowed matrix
 a22 a23 a2 n 
 
a amn 
 m2 am 3

to triangular form by E-row operations. When we apply the corresponding E-row operations to A, it will
be reduced to triangular form.
From the above we find that in all the three cases the matrix A can be reduced to trinagular form by
E-row operations.
The proof is now complete by the principle of finite induction.

101
 Exercise 1
1. Apply the elementary operation R2  R3 to the matrix
 1 −1 3 
 4 2 −6 
5 8 9 

2. Apply the elementary operation C2 → 2C2 to the matrixss
 −1 3 7 6 
 5 −1 4 −2 
 
3. Write down the elementary matrix obtained by applying R3 → R3 – 4R1 to l3.
4. Reduce the matrix
 1 2 −3 4 
 3 −1 2 0 
 2 1 −1 5 
 
to triangular form by applying E-row operations
5. Reduce the matrix
 1 −1 −1
4 1 0 
8 1 1 
 
to I3 by E-row operations.
6. Verify that the E-row equation R1 → R1 – R3 on the matrix AB is equivalent to the same E-row
operation on A, where
 1 −1 2   2 3 −5 
A =  −3 1 4  , B =  1 −2 6  .
 0 1 3 3 1 1 
   

3. INVERSE OF A MATRIX
Consider the matrices
1 2 −1   4 −1 1 
A =  −4 −7 −4  , B =  4 1 0  .
 −4 −9 5  8 1 1
   
It can be easily seen that
AB = BA = I3.
Because of this relationship between A and B we say that the matrix B is an inverse of the matrix A.
In fact, we have the following definition.
Definition 2. If A = [aij] be an n-rowed square matrix, then a square matrix B = [bij] is said to be an
inverse of A if AB = BA = In.
Since the relation AB = BA remains unaltered when we interchange A and B, it follows that if a
matrix B is an inverse of a matrix, A then A must be the inverse of B. Furthermore, it is obvious from the
defintion of an inverse of a matrix that if a matrix A has an inverse, then it must be a square matrix. In
fact, if A be an m × n matrix and B be a p × q matrix such that AB and BA both exist, we must have p = n,
and g = m. If, AB and BA are to he comparable (which is a necessary condition for AB and BA to be
equal) we must also have m = n, i.e., A must be a square matrix.
A matrix a having an inverse is called an invertible matrix.

102
 It is quite natural to ask the following questions regarding inverses of matrices :
Question 1. Does every square matrix have an inverse?
Question 2. In case the answer to question 1 is ‘No’, how can we test as two whether a given matrix
in invertible?
Questions 3. Can a matrix have more than one inverse?
Questions 4. In case the answer to question 2 is ‘No’ how can we proceeed to determine the inverse
of invertible matrix?
We shall try to answer to above quetions one by one.
Answer to question 1. Let us consider the matrix.
 0 0
A=  .
 0 0
If be any 2-rowed square matrix, we find that
AB = BA = 0.
We thus find that there cannot be any matrix B for which AB and BA are both equal to I*. Therefore
A is not invertible. We, therefore, find that a given square matrix need not have an inverse.
Before taking up question 2, we shall try to answer question 3.
Answer to questions 3. As the following theorem shows, if a matrix possesses an inverse, it must
be unique.
Theorem 5. If B and VC be inverses of a square matrix, A then B = C.
Proof. Let B and C be inverses of a square matrix A. Since B is an inverse of A, therefore
AB = BA = I ... (1)
Again, since C is an inverse of A, therefore
AC = CA = I ... (2)
From (1) we find that
C(AB) = CI = I A... (3)
Also, from (2) we find that
(CA)B = IB = B ... (4)
Since C(AB) = (CA)B , it follows from (3) and (4), that
C= B
In view of the above theorem it is only proper to talk of the inverse of a square matrix rather than
taking of an inverse.
The inverse of an inversible matrix is denoted by A–1.
Answer to question 2. Suppose a square matrix A possesses an inverse B. Then
AB = BA = I.
Since the determinant of the product of two matrices is equal to the product of their determinantsm,
we find that
| AB | = | I |
i.e., | A | × | B |= 1

103
 From the above relation we find that | A |  0. Thus a necessary condition for a square matrix to
have an inverse is that | A |  0. We shall now show that this conidtion is sufficient as well. In order to do
1
so, let us consider the matrix C = adj. A.
| A|
By using the idendity
A(adj. A) = (adj. A) A = | A | I,
1 1
We find that AC = ( A adj. A) = | A | I = I,
| A| | A|

 1  1 1
and CA =  adj. A  A = (adj. A) A =  | A | I = I.
| A|  | A| | A|
Since AC = CA = I,
it follows that C is the inverse of A.
From the above discussion we find that a square matrix possess an inverse if and only if
| A |  0.
This answers questgion 2. In order to test whether a square matrix possesses an inverse, we have
simply to calculate | A |. If | A | = 0 then A does not possesses an inverse but if | A |  0 then A possesses in
inverse.
Answer to question 4. White trying to answer quetions 2, we saw that if | A |  0, then the matrix
1
adj. A is the inverse of A. This provides one possible answer to question 4.
| A|
If a square matrix A possesses an inverse A, then in order to find the inverse we compute adj A, and
1
multiply it by the scalar . The resulting matrix is the desired inverse of A.
| A|
Example 3. Find the inver of the matrix
 2 5 0
 0 1 1  by first computing its adjoint.
 −1 0 3 
 
Solution. The given matrix be denoted by A.
The co-factors of the elements of the first row of A are 3, – 1, and 1 respectively.
The co-factors of the elements of the second row of A are –15, 6, and –5 respectively.
The co-factors of the elements of the third row are 5, –2, and 2 respectively.
 3 −1 1 
Therefore adj. A =  −15 6 −5 
 5 −2 2 
 
Also | A | = 2.3 + 5(–1) + 0.1 = 1
 3 −1 1 
Therefore A–1 =  −15 6 −5 
 5 −2 2 
 

104
4. USE OF ELEMENTARY OPERATIONS TO COMPUTE THE INVERSE OF A MATRIX
We have already described one method for finding the inverse of an invertible matrix. We shall now
describe another method for the same, namely by applying elementary row operations.
Suppose we wish to find the inverse of an n × n matrix A. We consider the identity
IA = A.
We reduce the matrix A on the right hand side to triangular from by E-row operations, and apply the
same E-row operations to the pre-factor I on the left hand side. In this manner we get the identity
PA = Q,
where P and Q are some triangular matrices. As our next step, we apply E-row operations on Q and
reduce it to the unit matrix I. The same E-row operations are, of course, applied to P. We get the identity
BA = I.
where B is obtained from P by E-row operations. The matrix B is the desired inverse of A. We shall
illustrate the procedure by considering a few examples.
Example 4. Find the inver of the matrix.
 3 −1 1 
A =  −15 6 −5 
 5 −2 2 
 
Solution. Consider the identity
1 0 0   2 3 −5   2 3 −5 
 0 1 0   −3 −5 9  =  −3 −5 9 
 0 0 1   1 1 −2   1 1 −2 
    
3
By performing the E-row operation R2 → R2 + R1 on the matrix on right as well as on the pre-
2
factor on the left, we have
1 0 0   2 3 −5 
3   2 3 −5   1 3
 1 0   −3 −5 9  =  −3 − 
 2   1 1 −2   2 2
0 0 1   1 1 −2 
1
Performing R3 → R3 – R1 on the matrix on the right as well as on the pre-factor on the left, we
2
have
 1 0 0 2 3 −5 
 3   2 3 −5   1 3
 1 0   −3 −5 9  =  0 − 
 2   1 1 −2   2 2
 − 1 0 0 0 − 1 1 
   
 2   2 2
Performing R3 → R3 – R2 on the matrix on the right as well as on the pre-factor on the left, we have
 1 0 0 2 3 −5 
 3   2 3 −5   1 3
 1 0   −3 −5 9  =  0 − 
 2   1 1 −2   2 2
 −2 −1 1 0 0 −1
Performing R1 → R1 – 5R3, on the matrix on the right as well as on the pre-factor on the left, we
have

105
  11 5 −5  2 3 0
 3   2 3 −5   
1 3  1
− −  −3 −5 9  =  0 − 0
 2 2 2   1 1 −2   2 
  0 0 −1
 −2 −1 1 
Performing R1 → R1 + 6R2 on the matrix on the right as well as on the pre-factor on the left, we have
 2 2 4 2 0 0
3 −5  
1 3  
 3 2 
1
− −  −3 −5 9  =  0 − 0
 2 2 2   1 1 −2   2 
−2 −1 1   0 0 −1
  
1
Performing R1 , R2 → − 2 R2 , and R3 → − R3 on the matrix on the right as well as on the pre-factor
2
on the left, we have

 1 1 2  2 3 −5   1 0 0 
 3 1 −3   −3 −5 9  =  0 1 0 
 2 1 −1  1 1 −2   0 0 1
    
From the above identity we find that the desired inverse is
 1 1 2
B =  3 1 −3 
 2 1 −1 
 
Verification. By actual multiplication we find that
AB = BA = I,
The above working an often be arranged in a more convenient and compact from as in the following
example.
Example 5. Find the invers of the matrix
 4 −1 −4 
 3 0 −4  .
 3 −1 −3 
 
Solution. We shall first reduce the given matrix to I3 by E-row operations
 4 −1 −4 
 4 −1 −4   
 3 0 −4  ~  0 3 3 3 
−1  R2 → R2 − R1 , R3 → R3 − R1 
 3 −1 −3   4  4 4 
 
0 − 1 
 0
 4 
 4 −1 −4 
 3  1 
~ 0 −1  R3 → R3 + R2 
 4  3 
0 0 − 1
 
 3

106
  4 −1 0
 3 
~ 0 0   R1 → R1 − 12 R3 , R2 → R2 − 3R3 
 4 
0 0 −1
 
 3
4 0 0
 3  4 
~ 0 0   R1 → R1 + R2 
 4  3 
0 1
0 − 

 3
 1 0 0
 1 4 
~  0 1 0   R1 → R1 , R2 → R2 , R3 → − 3R3 
 0 0 1  4 3 
 
3 3
The given matrix has been reduced to I3 by the E-row operation R2 → R2 − R1 , R3 → R3 − R2 ,
4 4
1 4 1 4
R3 → R3 + R2 , R1 → R1 − 12 R3 , R2 → R2 − 3R3 , R1 → R1 + R2 , R1 → R1 , R2 → R2, and R3 → 3R3.
3 3 4 3
We shall now perform these operations on l3 in this order. We then have
 1 0 0
 1 0 0  
 0 1 0  ~  − 3 1 0  R → R − 3 R , R − 3 R 
 0 0 1  4  2 2 1 3
4 
1
   4
 − 3 0 1
 
 4 
 1 0 0
 3  1 
~ − 1 0   R3 → R3 + R2 
 4  3 
 −1 1 1
 
 3 
 13 −4 −12 
 9 
~ 0 −3  [ R1 → R1 − 12 R3 , R2 → R2 − 3R3 ]
 7 
 −1 1 1

 3 
 16 −4 −10 
 9  4 
~ 0 −3   R1 → R1 + R2 
 4  3 
 −1 1 
 1
 3 
 4 −1 −4 
 1 4 
~  3 0 −4   R1 → R1 , R2 → R2 , R3 → − 3R3 
 3 −1 −3   4 3 
 
The last matrix is the desired invers.
The following theorem is sometimes useful for conputing inverses:

107
 Theorem 6. If A and B are n-rowed invertible matrices, then At and AB are both invertible and
(a) (At)–1 = (A–1)t.
(b) (AB)–1 = B–1A–1.
Proof. (a). Since AA–1 = A–1A = In, therefore by the reversal law for transposes we have
(AA–1)t = (A–1A)t = Itn,
i.e., (A–1)t At = A(A–1)t = In.
By the definition of the inverse of a matrix it follows that the matrix At is invertible and its inverse is
the matrix (A–1)t.
(b) Since A and B are both n-rowed invertible matrices, therefore A–1 and B–1 exist and are both n-
rowed square matrices, and consequently B–1A–1 is also an n-rowed square matrix. Let us denote the
matrix B–1A–1 C. We wish to show that B–1A–1 is the inverse of the n-rowed square matrix AB. In order to
achieve our aim we have to verify that
(AB) C = C(AB) = In,
where C denotes the matrix B–1A–1.
Now (AB)C = ( AB) ( B−1 A−1 )
= A( BB –1 ) A−1
= AIn A–1
= AA–1
= In
Also, C(AB) = (B–1 A–1) AB
= B–1(A–1 A)B
= B–1In B
= B–1B
= In .
From the above relations we find that (AB) C = (AB) = In.
Hence it follows that the matrix AB is invertible and its inverse is the matrix, i.e.,
(AB)–1 = C = B–1A–1.
Remark. Result (b) of the above theorem is often referred to as the reversal law of inverses.
The following example illustrates how the reversal law for inverse can be used for computing the
inverses of the product of two invertible matrices.
Example 6. Obtain the invers of the matrices
 1 p 0  1 0 0
A =  0 1 p , B =  q 1 0
 0 0 1  0 q 1
   
and hence that of the matrix
 1 p 0  1 0 0
C =  0 1 p , B =  q 1 0
 0 0 1  0 q 1
   

108
 Solution. In the very begining let us observe the following two facts which will be very useful :
(i) B is simply the matrix At with p replaced by q every where in At.
(ii) C = AB.
Step 1. Compute A–1
To compute A–1 we reduce A to I3, by E-row operations.
 1 p 0  1 p 0
 0 1 p  ~  0 1 0  [ R2 → R2 − pR3 ]
 0 0 1   0 0 1
   
 1 0 0
~  0 1 0  [ R1 → R1 − pR2 ]
 0 0 1
 
We find that A can be reduced to I3 by applying the E-row operations R2 → R2 – pR3 and R1 → R1 –
pR2. We now apply these E-row operations to I3 in this very order. We than get the matrix.
1 0 0
I3 ~  0 1 − p  [ R2 → R2 − pR3 ]
0 0 1

 1 − p p2 
= 0 1 − p  [ R1 → R1 − pR2 ]
 
0 0 1

 1 − p p2 
Thus A–1 =  0 1 −p .... (1)
 
0 0 1

From (i) we find that
 1 0 0
t −1
(A ) 
−1 t
= (A ) = − p 1 0
 2 
 p − p 1
Replacing p and q in the above identity throughout and using (i), we have
 1 0 0
B–1 =  −q 1 0 ... (2)
 2 
 q − q 1
By applying the reversal law to (ii) we have
C–1 = (AB)–1 = B–1A–1
 1 0 0   1 − p p2 
=  −q 1 0 0 1 −p
 2   
 q − q 1  0 0 1

 1 −p p2 
 
=  −q 1 + pq − p q − p .
2

 q 2 − pq 2 − q p 2 q 2 + pq + 1


109
 Exercise 2
1. Find the invers of the matrix.
 3 −15 5
 −1 6 −2 
 1 −5 2 
 
by calculating its adjoint.
2. Show that the matrix.
 1 −2 3 
 0 −1 4 
 −2 2 1

possesses an inverse.
 1 1 0
3. Given that A =  −2 −1 0  , calculate A2 and A–1.
 0 0 2
 
 1 2 −1  2 1 2
4. If A =  −4 −7 4  , B =  2 2 1 , verify that (AB)–1 = B–1A–1
 −4 −9 5   1 2 2
   
5. By applying elementary row operations find the inverse of the matrix
 1 1 2
 3 1 −3  .
 2 1 −1
 

110
 LESSON 3

RANK OF A MATRIX

1. INTRODUCTION
In this lesson we shall introduce the important concept of the rank of a matrix. We shall show that
the rank of matrix remains unaltered by elementary operations. This will give us a neat method for finding
the rank of matrix.
In the next lesson we shall use the concept of rank to determine whether a given system of linear
equation possesses a solution, and if so, then how many linearly independent solutions are there.

2. RANK OF MATRICES
Consider the matrix
 2 −3 −1 4 
A =  7 0 6 −5  .
 1 4 −3 2 
 
A is a 3 × 4 matrix. By delecting any one column of A we can obtain a 3 × 3 matrix. Let us recall
that such a matrix is called a sub-matrix of A. The determinant of such a matrix is called a minor of A.
Thus for example,
2 −3 −1 2 −1 4 −3 −1 4
7 0 6 , 7 6 −5 , 0 6 −5
1 4 −3 1 −3 2 4 −3 2
are all minor of A. Since each of these determinants has 3 rows, these minors are called 3-rowed
minors.
By detecting any one row and any two columns of A we can obtain sub-matrices of the type 2 × 2.
The determinant of any such sub-matrices is called a 2-rowed minor of A. Thus
2 −3 −1 4 0 6
, ,
7 0 6 −5 4 −3
are all 2-rowed minors of A.
In general, we have the following definitions:
Definition 1. Let A be an m × n matrix. The determinant of any s × s submatrix of A, obtained by
deleting m-s rows and n-s columns of A is called and s-rowed minor of A.
Let us now consider the matrix
2 3 −5 
A =  1 −7 8  .
 9 −19 9 
 
The matrix A has only one 3-rowed minor, namely
2 3 −5
1 −7 8 ,
9 −16 9
and this can be seen to be zero.

111
 A has 9 two-rowed minors (why ?) These are not all zero. In fact,
2 3
1 −7
is a 2-rowed minor which has the value 2. (–7) – 3.1 = –17.
The fact that every 3-rowed minor of A is zero and there is at least one 2-rowed minor which is not
zero, is usally expressed bny saying that the rank of the matrix A is 2.
Let us now consider the matrix
 1 0 0
I =  0 1 0 .
 0 0 1
 
It has only one 3-rowed minor, namely
1 0 0
0 1 0 ,
0 0 1
and this minor has the value 1. Since I a 3-rowed non-zero minor, and there are not minors having more
than three rows, we say that I is of rank 3.
In general, we have the following definition :
Definition 2. (1) A number r is said to be the rank of mark of a matrix A if
(i) A possesses at least one r-rowed minor which is different from zero.
(ii) A does not possess any non-zero (r + 1)-rowed minor
(2) The rank of every zero matrix is, by definition zero.
The rank of a matrix A is usually denoted by p(A).
Illustrations 1. If A In, then p(A) = n.
 0 0 0 0
2. If A =   , then ( A) = 0.
 0 0 0 0
From the definition of the rank of a matrix we have the following useful criteria for determining the
rank of a matrix;
(i) If all the (r + 1)-rowed minors of a matrix vanish or if the matrix does not possess any (r + 1)-
rowed minor, then the rank of the matrix must be either less than or at the most equal to r.
(ii) It atleast one r-rowed minor of a matrix does not vanish, then the rank of the matrix must be
either grether than or at least equal to r.
It is important (and useful) to note that if every (r + 1)-rowed minor of a matrix is zero, then every
higher order minor is automatically zero (because we can expand every higher order minor ultimately in
terms of (r + 1)-rowed minors.
Example 1. Find the rank of the matrix.
 2 −4 5 6 
A= .
 3 8 7 −1
Solution. The given matrix has at least one non-zero minor of order 2, namely
2 −4
3 8

112
 (In fact this minor = 2.8 – (–4).3 = 28) Actually, A has 6 2-rowed minors all the of which are non-
zero.
Since A has a non-zero 2-rowed minor, of therefore (A)  2.
Again, since A is a 2 × 4 matrix, therefore it does not have any 3-rowed minors. Consequently (A)
 2.
From the relations (A)  2 and (A)  2 we have (A) = 2.
Example 2. Find the rank of the matrix
 1 1 1
A =  a b c ,
 2 2

2
a b c 
all of a, b, and c being real numbers.
Solution.
1 1 1
|A|= a b c
a2 b2 c2
= (b – c) (c – a) (a – b).
The following different cases arise:
Case I. a, b, and c are all different. If a, b, and c all the different from each other, then | A |  0.
Therefore A has a non-zero three-rowed minor, namely
1 1 1
a b c ,
a2 b2 c2
so that (A)  3.
Also, A has no higher order minors.
Therefore (A)  3.
Hence (A) = 3.
Case II. Two of the numbers a, b, and c are equal but a, b, and c are not all equal. For the sake of
definiteness assume that a = b  c.
In this case | A | = 0 since | A | is the only 3-rowed minor and this minor vanishes, therefore
| A |  2. Also, A has atleast one non-zero 2-rowd minor, namely
1 1
b c
Therefore (A)  2.
Hence (A) = 2.
In the same way as above we find that (A) = 2 in the cases b = c  a or c = a  b.
Case III. a = b = c,

113
 1 1 1
Since a = b = c, therefore | A | = a a a = 0 so that the only 3-rowed minor is zero. Also all
a2 a2 a2
the 2-rowed minors of A vanish. In fact all three columns of A are identical in this case and therefore all
the nine 2-rowed minors of A vanish. Therefore, (A)  1.
Also, A has a non-zero of 1-rowed monor. (In fact these are at least three 1-rowed non-zero minors,
namely those which are obtained by deleting the 2nd and 3rd rows and any tow columns of A). Therefore
(A)  1.
Hence (A) = 1.

3. RANK AND ELEMENTARY OPERATIONS

In the two examples discussed above, we has obtained the ranks of the matrices being considered by
a straight forward consideration of all the minors. This may not be so easy in all the cases. It we have to
find the rank of A 4 × 4 matrix, we may have to consider a fourth order determinants, 16 third order
determinants, and perhaps some second order determinants.
By using elementary operations we can conveniently handle the problem of determining the rank of
a matrix. The following theorem gives the relation between elementary operations on a matrix and its
rank.
Theorem 1. The rank of a matrix remains unaltered by elementary operations.
Proof. We shall prove the theorem for elementary row operations only. The proof for elementary
column operations is similar.
Let A = [aij] be an m × n matrix of rank r.
Case I. Rp  Rq.
Let B be the matrix obtained from A by the E-row operation Rp  Rq and let the rank of B be s.
In order to verify that (B) = r, we shall show that all (r + 1)-rowed minors of B vanish and atleast
one r-rowed minor does not vanish.
Let B* be any (r + 1) - rowed squae sub-matrix of B. Since B has been obtained from A by the
operation Rp  Rq, therefore every row of B is also a row of A (and every row of A is a row of B). This
implies that | B* | =  | A* | for some (r + 1) -rowed submatrix A* or A. But every (r + 1)-rowed minor of A
is zero (because (A) = r). Therefore | B* | = 0, and consequently every (r + 1)-rowed minor of B is zero.
This implies that (B)  r, i.e., s  r.
Since B has been obtained from A by interchaning the pth and qth rows, therefore we can get A from
B by applying Rp  Rq to B. By interchanging the roles of A and B in the above proof it follows that r  s.
The relations s  r and r  s together yield r = s.
Case II. Rp → cRp (C  0)
Let us denote by B the matrix obtained from A by the opeation Rp → cRp, and let the rank of B be s.
Let B* be any (r + 1) - rowed sub-matrix of B and let A* be the correspondingly placed sub-matrix of A.
Let us compare | B* | and | A* |. Two possiblities arise :
(i) pth row of B is one of the row struck off to obtain B* from B.
In this case B* is identical with A* and | B* | = | A* |.

114
 (ii) pth row of B is not among the rows struck off to obtain B* from B. In this case | B |* =
c | A* |, for B* and A* agree with each other in all rows except one row, and every element of this
paritcular row of B* is c times the corresponding element of A*.
Since A is of rank r, therefore every (r + 1) -rowed minor of A is zero. In particular | A* | = 0 and in
each of the two cases (i) and (ii) above, we have | B* | = 0.
Since every (r + 1) - rowed minor of B is zero, therefore P(B)  , i.e., s  r. Also, since A can be
obtained from B by the E- operation Rp → C–1Rp, therefore by interchanging the roots of A and B it
follows that r  s. The relations s  r and r  s together yield r = s.
Case III. Rp → Rp + kRq (k  0).
Let us denote by B the matrix obtained from A by the E-operation Rp → kRq, and the rank of B be s.
Let B* be any (r + 1) - rowed sub-matrix of B and let A* be the corresponding sub-matrix of A.
Three different possibities arise :
(i) Neither the pth row nor qth row of B has been struck off while obtaining B*. In this case A* and
B* differ in only one row (the rows which correspond to pth row of A) and have all the other rows
identical. We have | B *| = | A* |. Since every (r + 1) - rowed minor of A is zero, threfore is particular | A*
| = 0, and consequently | B* | = 0.
(ii) The pth row of B has not been struck off while obtaining B* from B but the qth row has been
struck off.
Let us denote by C* the (r + 1) - rowed matrix obtained from A* by replacing apj by aqj. By the
property of determinants we have
| B* | = | A* | + k| C* |.
Observe that C* can be obtained from A by first performing the E-operation Rp → Rq and then
striking off those very rows and columns of the new matrix thus obtained as are strucke off for obtaining
B* from B. This implies that | C* | = ±1 times some (r + 1) - rowed minor of A. Since (A) = r,s therefore
every (r + 1) - rowed minor of A is zero, and so | A* | = 0, and | C* | = 0. From (1) it follows that | B* | =
0.
(iii) The pth row of B is one of those rows which have been struck off while obtaining B*. Since A
and B differ only in the pth row, therefore in this case B* = A* an so | B* | = | A* |. Since A is of rank r,
therefore | A* | = 0 and consequently | B* | = 0. Since | B* | = 0, in all the cases (i) - (iii), therefore every
(r + 1) - rowed minor of B is zero and so (B)  r, i.e., s  r. Also since A can be obtained from B by the
E-operation Rp → Rp – kRq, therefore by interchanging the roles of A and B it follows that r  s. Hence r
= s.
We have thus shown that the rank of a matrix remains unaltered by E-row operations, and so
remarked in the beginning this complest the proof.
Example 3. Reduce the matrix
 3 −1 1 3 
 −1 −4 −2 −7 
 2 1 3 0 

 −1 −2 3 0 
to triangular form by E-row operations and hence determine its rank.
Solution. Let us denote the given matrix by A. We shall first of all interchange R1 and R4 so that we
get –1 as the (1, 1) the entry. This will ultimately simplify the entire working.

115
  3 −1 1 3 
 −1 −4 −2 −7 
A= 
 2 1 3 0 
 −1 −2 3 0 

 −1 −2 3 0
 −1 −4 −2 −7 
~  , by R1  R4
 2 1 3 0 

 3 −1 1 3 
 −1 −2 3 0 
 0 −2 −5 −7 
~  , by R2 → R2 − R1 , R3 → R3 + 2 R1 , R4 → R4 + 3R1
 0 −3 9 0 
−7 10 2 
 0
 −1 −2 3 0 
 0 −2 −5 −7 
 33 21  3 7
~ 0 0 , by R3 → R3 − R2 , R4 → R4 − R2
 2 2 2 2
 55 55 
0 0 
 2 2
 −1 −2 3 0
 0 −2 −5 −7 
~   , by R → R − 5 R
 0 0 33 21  4 4
3
3
 2 2
0 0 0 10 

If we denote the above matrix by B, then we see that B = 330  0, so that rank of B = 4.
But A and B have the same rank.
Therefore A is of rank 4.
 2 3 4 5 6
 3 4 5 6 7
Example 4. Find the rank of the matrix  4 5 6 7 8  .
 
 9 10 11 12 13 
14 15 16 17 18 
 
Solution. Let us denote the given matrix by A. Observe that if we subtract the lement of the first row
from the corresponding elements of the second row, then the second row will consist of ones only. In fact
1 1
by R2 → R2 − R1 , R3 → R − R1 , R4 → R4 − R1 , R5 → R5 − R1 , and then R3 → R3 , R4 → R4 ,
2 7
1
R5 → R12 , we find that
12
 2 3 4 5 6
 1 1 1 1 1
A ~  1 1 1 1 1
 
 1 1 1 1 1
 1 1 1 1 1
 

116
 2 1 2 4
3
1 0 0 0
0
~ 1 0 0 0  , by C2 → C2 − C1 , C3 → C3 − C1 , C4 → C4 − C1 , C5 → C5 − C1 ]
0

1 0 0 0
0
1 0 
 0 0 0
 2 1 0 0 0
 1 0 0 0 0
~  1 0 0 0 0  [by C3 → C3 − 2C2 , C4 − 3C2 , C5 → C5 − 4C2 ]
 
 1 0 0 0 0
 1 0 0 0 0
 
Denoting the last matrix by B, we find that every 3-rowed minor of B is zero, and B has atleast one
2 1
non-zero two-rowed minor, namely .
1 0
There (B) = 2.
But (A) = (B), so that (A) = 2.
Remark. Observe that in the above example we did not proceed in the normal way for triangular
reduction.
Example 5. If x, y, z are all unequal, show by using elementary operations or otherwise that the
matrix
 0 x − y x − z y + z
y − x 0 y − z x + z
A= 
 z − x z − y 0 x + y 
y + z z + z x+ y 0 
is of rank 2.
Solution.
 0 x−y x−z y+ z
y − x 0 y−z x+ z
A= 
 z − x z−y 0 x+ y 
y + z z+x x+y 0 
 x−y x− y x− y y − x
y− x 0 y− z x + z
~  [by R1 → R1 − R2 ]
 z − x z−y 0 x + y 
y + z z+x x+ y 0 

 1 1 1 −1
y − x 0 y − z x + z  1 
~  , by R1 →
0 x + y  
R1 
 z − x z − y x−y 
y + z x− y x+y 0 

 1 0 0 0
y − x x − y x−z y + z
~  , [by C 2 → C2 − C1 , C3 → C3 − C1 , C4 → C4 + C1 ]
 z − x x − y x−z y + z 
y + z x− y x−z y + z 

117
 1 0 0 0
0 x − y x−z y + z
~  , [by R2 → R2 − ( y − z ) R1 , R3 →
 0 x − y x−z y + z 
0 x − y x−z y + z 

Two different cases arise according as y + z  0 or y + z = 0.

Case I. y + z  0.
1 0 0 0
0 1 1 1  1 1 1 
A~  , by C2 → C2 , C3 → C3 , C4 →
1  
C4 
 0 1 1 x− y x−z y+ z 
0 1 1 1 

1 0 0 0
0 1 0 0
~ , [by C3 → C3 − C2 , C4 → C4 − C2 ]
 0 1 0 0 
0 1 0 0 
1 0 0 0
0 1 0 0
~ , [by R3 → R3 − R2 , R4 → R4 ~ R2 ]
 0 0 0 0 
0 0 0 0 

1 0 0 0
0 1 0 0
Since the matrix  is of rank 2, therefore it follows that (A) = 2
 0 0 0 0 
0 0 0 0 
Case II. y + z = 0
1 0 0 0
0 x − y x−z 0
A~ 
 0 x − y x−z 0 
0 x − y x−z 0 
1 0 0 0
0 1 1 0  1 1 
~  , by C2 → C2 , C3 →
0  
C3 
 0 1 1 x−y x−z 
0 1 1 0 

1 0 0 0
0 1 1 0
~  , [by R3 → R3 − R2 , R4 → R4 − R2 ]
 0 0 0 0 
0 0 0 0 
1 0 0 0
0 1 0 0
~  , [by C3 → C3 − C2 ]
 0 0 0 0 
0 0 0 0 
which is of rank 2.
Hence (A) = 2 in this case also.

118
Thus (A) = 2
Example 6. Find the rank of the matrix
 0 r −q x 
 −r 0 p y
 q −p 0 z 
,
 
 −x − y −z 0 
where px + qy + rz = 0, and all of p, q, and r are positive real numbers.s
Solution. Let the given matrix be denoted by A.
 0 r −q x 
 −r 0 p y
A= 
0 z 
,
 q − p 
 −x − y −z 0 
 0 r −q px 
 −r 0 p py 
~  , [by C4 → pC4 , since p  0]
 q − p 0 py 
 q − p −2 0 
 0 r −q px + qy 
 −r 0 p 0
~  , [by C4 → C4 − yC3 ]
 q − p 0 pz 
 − x − y − z yz 

 0 r −q px + qy + rz 
 −r 0 p 0 
~ 
q − p 0 0  , [by C4 → C4 + zC2 ]
 
 − x − y − z 0 
 0 r −q 0 
 −r 0 p 0
~  , since px + qy + rz = 0.
 q − p 0 0 

 −x − y −z 0 
 0 r −qr 0
 −r 0 pr 0
~  , [by C3 → rC3 ]
 q −p 0 0 

 −x − y −rz 0
 0 r 0 0
 −r 0 pr 0 
~  , [C → C2 + qC2 ]
 q −p − pq 0  3
 −x − y −qy − rz 0 

 0 r 0 0
 −r 0 0 0
~  , [by C3 → C3 + qC2 ]
 q − p 0 0 
 − x − y − ( px + qy + rz ) 0 

119
  0 r 0 0
 −r 0 0 0
~  , since px + qy + rz = 0.
 q − p 0 0 
 − x − y 0 0 
= B(say)
Now every 3-rowed minor of the matrix B vanishes, and there is atleast one non-zero 2-rowed minor
0 r
namely , therefore the rank of B is 2. Since A ~ B, therefore (A) = (B) = 2.
−r 0
Example 7. If the number a, b, and c are all different from each other, and the numbers, x, y,z are
also all different from each other slow that the matrix
1 a x ax 
A = 1 b y by 
1 c z cz 

is of rank 3.
Solution.
1 a x ax 
A = 1 b y by 
1 c z cz 

1 a x ax 
~ 0 b − a y − x by − ax  , [by R2 → R2 − R1 , R3 → R3 − R1 ]
0 c − a z − x cz − ax 

1 a 0 0
~ 0 b − a y − x by − ax  , [by C2 → C2 − aC1 , C3 → C3 − xC1 , C4 → C4 − axC1 ]
0 c − a z − x cz − ax 

1 0 0 0
~ 0 b − a y − x b( y − x )  , [by C4 → C4 − xC2 ]
0 c − a z − x c( z − x) 

1 0 0 0 

~ 0 b−a y−x 0  , [by C4 → C4 − bC3 ]
0 c − a z − x (c − b) ( z − x ) 

= B(say).
Let us first of all consider the rank of B. since B is a 3 × 4 matrix, (B)  3. Also, since B has a 3-
1 0 0 
rowed non-zero minor, namely  0 y − x 0  , therefore r(B)  3.
 0 z − x (c − b)( z − x) 
 
From the inequalities (B)  3 and (B)  3 it follows that (B) = 3.
Also, since the rank of a matrix remains unaltered by elementary operations, therefore (A) = (B) =
3.
Hence the rank of the given matrix is 3.

120
 Exercise
1. Find the rank of the matrix
 3 −1 2 −6 
 6 −2 4 −9  .
 
2. Find the rank of the matrix
 −1 2 3 
 0 3 4 .
 0 0 0
 
3. Find the rank of the matrix.
 1 1 1
 a b c .
 3 3

3
a b c 
4. Reduce the matrix
 1 −1 2 −3 
4 1 0 2
0 3 1 4 
.

0 1 0 2 
to triangular form and hence find its rank.
5. Find the rank of the matrix.
 12 22 32 42 
 2 
2 32 42 52 
.
 32 42 52 62 
 2 
4 52 62 7 2 

121
 LESSON 4

SYSTEMS OF LINEAR EQUATIONS

1. INTRODUCTION
In this lesson we shall apply our knowledge of matrices to obtain solutions of systems of linear
equations.
Consider a system of m linear equations.
a11 x1 + a12 x2 + + a1n xn = b1 , 
a21 x1 + a22 x2 + + a2 n xn = b2 
 ... (1)

am1 x1 + am 2 x2 + + amn xn = bm , 
in n unknows x1, x2, ..., xn.
We can write the above system of equation as
 x1   b1 
   
x2 b2
 a11 a12 a1n     
  .   . 
 a21 a22 a2 n     
. = .
   
... (2)

  .   . 

 am1 am 2 amn     
 .   . 
 x  b 
 n  n
If we write
 x1 
 
 a11 a12 a1n   x2 
   . 
a a22 a2 n   
A =  21 , X = . 
 
   . 
 am1 am 2 amn   
 . 
x 
 n
 b1 
 b2 
 . 
and B =   , so that A is an m × n
 . 
 . 
b 
 n
matrix, X is an n × 1 matrix, and B is an m × 1 matrix, then we can write (2) as
AX = B

122
 The matrix A is called the co-efficient matrix, and [A, B] is called the augmental matrix, where [A,
B] stands for the m × (n + 1) matrix obtained by adjoining one column, namely (b1, b2) ... bm)t to the m ×
n matrix A.
We shall confine out study to the following aspects :
I. Application of the inverse of a matrix to study a system of n equations AX = B in n-unknowns. We
shall show that if the matrix A possesses an inverse, then the system has the unique solution which is
given by X = A–1B. As a corollary of this result we shall show that if the matrix A is invertible, then the
system of equations AX = 0 does not possess any non-zero solution.
II. We shall use elementary row operations to put AX = B in a simplified form from which the
solution of the system can be written easily.
III. We shall apply the concept of rank of a matrix to discuss the existence and uniqueness of
solutions of AX = B. We shall show that if A be m × n matrix of rank r, then (i) the general solution of the
system of homogeneous equations AX = 0 contains n–r arbitrary constants, and (ii) the system of non-
homogeneous equations AX = B posses a solution if and only if the matrix [A, B] is also of rank r. In case
this condition is satisfied, the solution contains n–r arbitrary constant.

2. USE OF INVERSE OF MATRIX TO SOLVE A SYSTEM OF LINEAR EQUATIONS

Consider a system of n linear equations
a11 x1 + a12 x2 + + a1n xn = b1 , 
a21 x1 + a22 x2 + + a2 n xn = b2 
 ... (1)

an1 x1 + an 2 x2 + + ann xn = bn , 
in n unknowns x1, x2, ... xn.
We can write the above system of equation as
 x1   b1 
 x  b 
 a11 a12 a1n   2   2 
 a21 a22 . .
a2 n     
  . = .  ... (2)
  .   . 
 an1 an 2 ann     
. .
 x  b 
 n  n
If we write
 x1 
 x2 
 a11 a12 a1n   . 
a a22 a2 n   
A =  21 , X =  . ,
   . 
 an1 an 2 ann   . 
x 
 n
 b1 
 b2 
 . 
and B =  ,
 . 
 . 
b 
 n

123
 so that A is an n × n matrix, X is an n × 1 matrix, and B is an n × 1 matrix then we can write (2) as
AX = B ... (3)
If A is a non-singular matrix, so that A– exists, we pre-multiply (3) throughout by A–1 , as that
1

A–1 (AX) = A–1 B

 (A–1A)X = A–1B
 IX = A–1B
 = A–1B. ... (4)
From (4) we find that if the system of equations (3) has a solution, it is given by (4). By substituting
X = A–1B in (3) we find that
AX = A(A–1B) = (AA–1) B = IB = B, showing that X = A–1B is indeed a solution. Thus we have the
following result.
Corrollary. If A is an n-rowed non-singular matrix, the system of homogeneous linear equations.
AX = 0 has X = 0 as its only solution.
Example 1. Show that the system of equations
4x + 3y + 3z = 1
–x – z = 2
–4x – 4y – 3z = 3
has a unique solution, and obtain it by first computing the inverse of the co-efficient matrix.
Solution. The given system of equations can be written as
 4 3 3  x  1
 −1 0 −1  y  =  2  ,
 −4 −4 −3   x   3
    
or as AX = B.
 4 3 3  x 1
where A =  −1 0 −1 , X =  y  , B =  2  .
 −4 −4 −3  z  3
     
4 3 3
|A|= −1 0 −1
−4 −4 −3
0 −1 −1 −1 −1 0
= −3 +3
−4 −3 −4 −3 −4 −4
= 4(–4) – 3. (–1) + 3.4
= –1.
Since | A |  0, the matrix is invertible. Since the co-efficient matrix is invertible, therefore the given
system of equations possesses a unique solution.

We shall now find A–1.

 −4 −3 −3 
Now adj. A=  1 0 1
 4 4 3 


124
 1
A–1 = adj. A
| A|
 4 3 3
= − 1 0 −1 

 −4 −4 −3 
 
X = A–1B
 4 3 3  1 
= − 1 0 −1   2  ,

 −4 −4 −3   3 
  
 19 
=  −4  .
 −21 
 
Therefore x = 19, y = –4, z = –21 is the desired solution.
Verification. When x = 19, y = –4, z = –21,
4x + 3y + 3z = 4.19 + 3(–4) + 3 (–21) = 1,
–x –z = –19 + 21 = 2
–4x – 4y –3z = –4 (19) –4 (–4) –3 (–21) = 3.
Example 2. Show, by considering the co-efficient matrix, that the system of equations
x + 2y – z = 0,
–4x –7y + 4z = 0,
–4x – 9y + 5z = 0,
does not possess a non-zero soltion.
Solution. The co-efficient matrix of the given system of equation of given by
 1 2 −1 
A =  − 4 −7 4 
 − 4 −9 5 
 
−7 4 −4 4 −4 −7
|A|= −2 −
−9 5 −4 5 −4 −9
= 1 – 2(–4) –8
= 1.
Since | A |  0, therefore A is invertible. Since A–1 exists, the given system of equations does not
possess a non-zero solution.

3. USE OF ELEMENTARY ROW OPERATION TO SOLVE A SYSTEM OF LINEAR

EQUATONS
Consider the system of m linear equations

125
 a11 x1 + a12 x2 + + a1n xn = b1 , 
a21 x1 + a22 x2 + + a2 n xn = b2 
s ...(1)

am1 x1 + am 2 x2 + + amn xn = bm , 
By writing
 a11 a12 a1n 
 
a21 a22 a2 n 
A= 
 
 
 am1 am 2 amn 

 x1   b1 
 x2   b2 
 .   . 
X =   , and B =  
 .   . 
 .   . 
x  b 
 n  n
we can write (1) as
AX = B,
where A = [aij] is an m × n matrix, X = ( x1, x2 , xn )t is an n × 1 matrix, and B = (b1 , b2 , bn )t is an m ×
1 matrix.
The system of equations (1) remains unaltered by the following operations :
(i) Interchange of any two equations;
(ii) Multiplication of both sides of an equation by non-zero number k;
(iii) addition of k times an equation to another equation.
The above equations are equivalent to the following elementary row operations on the matrices A
and B.
(i) interchange of two rows.
(ii) multiplying a row of A and the corresponding row of B by a non-zero number k;
(iii) adding k times an equation to another equation.
Let us recall that :
(a) A matrix can be reduced to triangular form by E-row operations.
(b) An E-row operation on the product of two matrices is equivalent to the same E-row operation
on the pre-factor.
In view of the all the facts stated above we find that in order to solve a given system of linear
equation we may proceed as follows :
(1) Write the given system in matrix notation as AX = B.
(2) Reduce A to a triangular matrix say A* by E-row operations.
(3) Keep on applying the E-row operations in (2) above to the matrix B in the same order as in (2)
above.
(4) Suppose the result of (2) and (3) gives us the equation A * X = B* where A* is a triangular
matrix. By writing A * X = B* as a system of linear equations, we shall arrive at a system of

126
 equations which can be solved easily. Of course, it can be happen that the system does not
possess a solution. This will happen if one of the equations in A * B = B* is of the form
0 · x1 + 0 · x2 + ... + 0 · xn = b,
for some non-zero number b.
We shall illustrate the method with the help of examples.
Example 3. Solve the system of equations
2x – 5y + 7z = 6,
x – 3y + 4z = 3,
3x – 8y + 11z = 11.
 2 −5 7  6
Solution. Let A = 1 −3 4 , B =  3 
 
 3 −8 11 11
   
 x
and assume that there exists a matrix X =  y  , such that AX = B
z
 
 2 −5 7   x  6
Now  1 −3 4   y  =  3 
 3 −8 11  z   11 
    
 1 −3 4   x  3
  2 −5 7   y  =  6  , by R1  R2
 3 −8 11  z  11
    
 1 −3 4  x   3
 0 1 −1 y  =  0  , by R2 → R2 − 2 R1 , R3 → R3 − 3R1
0 1 −1   2
  z   
 1 −3 4   x   3
 0 1 −1  y  =  0  , by R3 → R3 − R2
0 0 0  z   2
    
 x – 3y + 4z = 3,
y – z = 0,
0 = 2.
Since the conclusion 0 = 2 is false, therefore our assumption that for some X, AX = B, is false.
Consequently there is no X for which AX = B, that is, the given system of equations has no solution.
Remark. A system of equations having no solution is said to be inconsistent.
Example 4. Solve the system of linear equations
x – 2y + 3z = 6,
3x + y – 4z = –7,
5x – 3y + 2z = 5.

127
  1 −2 3  6  x

Solution. Let A = 3   
1 −4 , B = −7 , and assume that there exists a matrix X =  y  , such
 5 −3 2   5 z
     
that AX = B.
 1 −2 3  x   6
Then 3 1 −4   y  =  − 7 
 5 −3 2   z   5
    
 1 −2 3  x   6
  0 7 −13   y  =  −25  , by R2 → R2 − 3R1 , R3 → R3 − 5 R1.
0 7 13   z   −25 
  
 1 −2 3  x   6
  0 7 −13   y  =  −25  , by R3 → R3 − R2
0 7 0   z   0
  

 x − 2 y + 3z = 6,
7 y − 13 z = − 25
13 25
Form the second of above equations, we have y = z − . Substituting this value of y in the first
7 7
 13 25  5 8
equation, we have x = 2y – 3z + 6 =  z −  − 3z + 6 = z − .
7 7  7 7
Therefore we have
5 8
x= z−
7 7
13 25
y= z− ,
7 7
z = z,
5 8 13 25
i.e., x = k − , y = k − , z = k , where k is any real number. By actual substitution we find
7 7 7 7
that these values satisfy the given equations.
Hence all the solutions of the given system of equations are
5 8 13 25
x= k − , y = k − , z = k,
7 7 7 7
where k is any real number
Example 5. Solve the following system of linear equations :
2x – 3y + z = 0,
x + 2y – 3z = 0,
4x – y –2z = 0.
 2 −3 1  x
 
Solution. Let 1 2 −3 , and assume that there exists X =  y  , such that AX = O, where O
 4 −1 −2  z
   
stand for the zero-matrix of type 3 × 1.

128
  2 −3 1  x 
Then  1 2 −3   y  = O
 4 −1 −2   z 
  
 1 2 −3   x 
  2 −3 1  y  = O, by R1  R2
 4 −1 −2   z 
  
 1 2 −3   x 
  0 −7 7   y  = O, by R  R – 2R , R → R – 4R
 0 −9 10   z  2 2 1 3 3 1
  
 1 2 −3   x 
  0 −7 7   y  = O, R →
 0 −9 10   z  2
  
 1 2 −3   x 
  0 −7 7   y  = O, R → R + 9R
 0 −9 10   z  3 3 2
  
 x + 2y – 3z = 0
y – z =0
z =0
 x = 0, y = 0, z = 0.
Thus we find that if the given system possesses a solution, it must be given by
x = 0, y = 0, z = 0.
Also, by actual substitution it follows that x = y = z = 0 is a solution.
Hence x = y = z = 0 = is the only solution of the given system.
Example 6. Solve the following system of equations :
x + 3y + 4y + 7w = 0,
2x + 4y + 5z + 8w = 0,
3x + y + 3z + 3w = 0,
x 
 1 3 4 7 y
Solution. Let A =  2 4 5 8  and assume that there exists a matrix X =   such that AX = O,
 3 1 2 3  z 
 
 w
where O stands for the zero-matrix of type 4 × 1.
x 
 1 3 4 7  
Then  2 4 5 8  y  = O
 3 1 2 3  z 
   w
 
x 
1 3 4 7  
  0 −2 − 3 − 6   y  = O, by R → R – 2R , R → R – 3R
 0 −8 −10 −18   z  2 2 1 3 3 1
   w
 

129
 x 
1 3 4 7  
  0 −2 −3 −6   y  = 0, R → R – 4R
0 0 2 6  z  3 3 2
   w
 
 x + 3y – + 4z + 7w = 0
–2y – 3z – 6w = 0
2z + 6w = 0
3 1
 z = −3w, y = w, x = w
2 2
1 3
 x = k , y = k , z = − 3k , w = k ,
2 2
where k is any real number. Thus we find that if the given system possesses a solution it must be
given by
1 3
x= k , y = k , z = − 3k , w = k , where k is any real number.
2 2
By actual substitution we find that (1) is indeed a solution of the given system.
1 3
Hence all the solutions of the given system of equations are given by x = k, y = k,
2 2
z = − 3k , w = k , where k is any real number.
Example 7. Solve the following system of equations :
x – 3y + 2z = 0,
7x – 21y + 14z = 0,
–3x + 9y – 6z = 0.
 1 −3 2  x
Solution. Let A =  7 −21 17  and assume that there exists a matrix X =  y  , such that AX = O,
 −3 9 −6  z
  
where O is the zero-metrix of type 3 × 1.
 1 −3 2   x 
Then  7 −21 14   y  = O
 −3
 9 −6   z 

 1 −3 2   x 
  0 0 0   y  = 0, by R → R – 7R , R → R + R
0 0 0  z  2 2 2 3 3 1
  
 x – 3y + 2z = 0
 x = 3y – 2z, y = y, z = z
 x = 3k1 – 2k2, y = k1, z = k2, ... (1)
where k1, k2 are any real numbers.
Thus we find that if there is a solution of the given system, it must be (1).
Again, by actually substituting x = 3k1 – 2k2, y = k1, z = k2 in the given system of equations, we find
that

130
 x – 3y + 2z = (3k1 – 2k2) – k1 + 2k2 = 0,
7x – 21y + 14z = 7(3k1 – 2k2) – 21k1 + 12 k2 = 0,
–3x + 9y – 6z = – 3(3k1 – 2k2) + 9k1 – 6k2 = 0,
so that (1) is indeed a solution. Therefore we conclude that
x = 3k1 – 2k2, y = k1, 3 = k2, where k1, k2 are real numbers, are all the solutions of the given system of
equations.

4. APPLICATION OF RANK OF A MATRIX TO A SYSTEM OF LINEAR EQUATIONS

Let us consider a system of m linear equations in n variables:
AX = B,
where A is an m × n matrix, B is an m × 1 matrix, and X is an n × 1 matrix. It can be shown that the above
system of equtions is consistent if and only if the matrices A and [A, B] have the same rank. Further more,
if the common rank be r, the general solution contains n–r arbitrary constants.
We shall not prove the about result, but will contnet ourselves with the followig illustrative
examples.
Example 8. Show, without actually solving, that the following system of equations is inconsistent.
x – y + z = – 1,
2x + y – 4z = –1,
6x – 7y + 8z = 7.
Solution. The given system of equations can be written as
AX = B
 1 −1 1 x  −1
where A = 2 1 −4  , X =  y  , B =  1 .
 6 −7 8  z  7
    
Let us find the ranks of the matrices A and [A, B]
 1 −1 1
A = 2 1 −4 
 6 −7 8 

 1 −1 1
~  0 3 −6  , by R 2 → R2 − 2 R1 , R3 → R3 − 6 R1
 0 −1 2 
 
 1 −1 1 1
~  0 1 −2  , by R2 → R2
 0 −1 −2  3
 
 1 −1 1
~  0 1 −2  , by R3 → R3 + R2
0 0 0
 
 1 −1 1
Since  0 1 −2  , is of rank 2,
0 0 0
 

131
 therefore (A) = 2
 1 −1 1 −1
Also, [A, B] =  2 1 −4 −1
 6 −7 8 7 

 1 −1 1 −1
~  0 3 −6 1 , by R 2 → R2 − 2 R1 , R3 → R3 − 6 R1
 0 −1 2 13 
 
 1 −1 1 −1 1
~  0 3 −6 1 , by R3 → R3 + R2
  3
 0 0 0 40 
 3 
 1 −1 1 −1 3
~  0 3 −6 1 , by R3 → R3
 0 0 0 1 40
 
 1 −1 1 0 
~  0 3 −6 1 , by R1 → R1 + R3 , R2 → R2 − R3
 0 0 0 1
 
 1 0 0 0
~  0 3 −6 1 , by C2 → C2 + C1 , C3 → C3 − C1
 0 0 0 1
 
 1 0 0 0 1 −1
~  0 1 1 0  , by C2 → C2 , C3 → C3
 0 0 0 1 3 6
 
 1 0 0 0
~  0 1 0 0  , by C3 → C3 − C2
 0 0 0 1
 
showing that the rank of arugmented matrix is 3.
Since the ranks of the co-efficient matrix and augmented matrix are not equal, the given system of
equations is inconsistent.
3x + 4y – 6 z + w = 7
x – 2y + 3z + 2w = –1
x – 3y + 4z – w = –2
5x – y + z – 2w = 4.
Solution.
A B
 3 4 −6 1 : 7
 1 −2 3 −2 : −1
?=  
 1 −3 4 −1 : −2 
 5 −1 1 −2 : 4 

132
  1 −2 3 −2 : −1
 3 4 −6 1 : 7
?~   , by R1  R2
 1 −3 4 −1 : −2 
 5 −1 1 −2 : 4 
 1 −2 3 −2 : −1
 0 10 −15 7 : 10 
?~ 
0 −1 1 1 : −1  , by R2 → R2 − 3R1 , R3 → R3 − R1 , R4 → R4 − 5R1
 
 0 9 −14 8 : 9 
 1 −2 3 −2 : −1
 0 10 −15 7 : 10 
 1 17  1 9
? ~ 0 0 − : 0  , R3 → R3 + R2 , R4 → R4 − R 2
 2 10  10 10
 1 17 
0 0 − : 0
 2 10 
 1 −2 3 −2 : −1
 0 10 −15 7 : 10 
?~   , by R → R − R .
 0 0 − 1 17 : 0 4 4 3
 2 10 
0 0 : 0 
 0 0
From the above we find that the ranks of A and [A, B] are both equal to 3. Since the co-efficient
matrix and augment both have the same rank, therefore the given system of equations is consistent.
Remark. Observe that in the above example we have computed the ranks of A and [A, B]
simultaneously. This has meant considerable simplification in workings.
Exericise
1. Does the following system of equations possess a unique common solution?
x + 2y + 3z = 6,
2x + 4y + z = 7,
3x + 2y + 9z = 10.
It so, find it.
2. Show that x = y = z = 0 is the only common solution of the following system of equations:
2x + 3y + 4z = 0,
x – 2y – 3z = 0,
5x + y – 8z = 0.
3. Solve the system of equations :
x + y + z = 3,
3x – 5y + 2z = 8,
5x – 3y + 4z = 14.
4. Solve the followng system of equations :
x – 3y + 2z = 0,
7x – 21y + 14z = 0,
–3x + 9y – 6z = 0.

133
5. Which of the following system of equations are consistent ?
(a) x – 4y + 7z = 8,
3x + 8y – 2z = 6,
7x – 8y + 26z = 31.
(b) x – y + 2z = 4,
3x + y + 4z = 6,
x + y + z = 1.

134
 LESSON 5

THE CHARACTERISTIC EQUATION OF A MATRIX

1. INTRODUCTION
In this lesson we shall study the important concept of characteristic equation of a square matrix.
The characterstic equation of a square matrix has many applications. It can be used to compute
power of a square matrix, and inverse of a non-singular matrix. It is also useful in simplifying matrix
polynomials.
We shall begin with some definitions.
Definition. Let A = [Aij] b an n × n matrix. The matrix A – I, where  is a scalar and I is the n-
rowed unit matrix, ims called the characteristic matrix of A. The determinant |A – | is called the
characteristic polynomial of A. The equation |A – I| = 0 is called the characteristic equation of A and its
roots are called the characteristic roots (or latent roots or eigen values) of A.
Remark. It can be easily seen that the diagonad elements of A are of the first degree in  and the
off-diagonal elements are independent of . |A – I| is, therefore a polynomial of degree in , and the co-
efficient of n is (–1)n. It follows that |A – I| = 0 is of degreen n is . By the fundamental theorem of
algebra, every polynomial equation of degree n has exactly n roots. Therefore it follows that every square
matrix of order n has exactly n characteristic roots.
Example 1. Find the characteristic roots of the matrix.
 1 −1 4 
 
A = 0 3 7
0 0 5
 
Solution.
1−  −1 4
| A − I| = 0 3− 7 = (1 −  )(3 −  )(5 −  )
0 0 5−
The roots of |A – l| = 0 are, therefore, 1, 3, and 5.
Hence the characteristic roots of A are 1, 3, and 5.
Example 2. Show that the matrices A and A have the same characteristic roots.
Solution. Le A = [aij] be an n  n matrix. The characteristic equation of A is
|A – I| = 0 (i)
Also the characteristic equation of At is
|At – I| = 0,
which is the same as
|(A – I)|t = 0 (ii)
Since the value of a determinant remains unaltered by taking its transpose, therefore
|A – I| = |(A – I)t|
Consequently
|A – I| = 0,  |(A – I)|t = 0 (iii)

135
 i.e., |A – I| = 0 and |(A – I)t| = 0 are the same equation. Since |A – I| = 0 is the characteristic
equation of A, and |(A – I)t| = 0 is the characteristic equation of At, it follows that A and At have the same
characteristic roots.
Example 3. Show that the characteristic roots of a triangular matrix are just the diagonal elements
of the matrix.
 a11 a12   a1n 
 
 0 a22   a2 n 
 0 0 a33  a3n 
 
Solution. Let A =        be an n × n triangular matrix.
     
 
     
 0 0  ann 
 0
The characteristic equation of A is
a11 −  a12 L a1n
0 a22 −  L a2 n
. .
=0
. .
. .
0 0 ann − 1
i.e., (a11 – ) (a22 – ) … (ann – ) = 0. (i)
The roots of (i) are given by
 = a11, a22, a33, …, and ann which are just the diagonal elements of A.
Hence the charactristic roots of A are just the diagonal elements of A.
Example 4. Show that if 1 2, … n be the charactertistic roots. of A. then k1, k2, …, kn are the
characteristic roots of kA.
Solution. Let A = [aij] be an n-rowed square matrix. The characteristic equation of A is
 a11 −  a12 L a1n 
 
 a21 a22 −  L a2 n 
 . . 
 =0
 . . 
 . . 
 
 an1 an 2 ann −  
Let the roots of (i) be 1, 2, …, n. to obtain the equation whose roots are k1, k2, … kn, we have
to substitute * = k in (i). The resulting equation in * will have roots k1, k2, …, kn,
By the transformation * = k, equation (i) is transformed to

136
 *
a11 − a12 K a1n
k
*
a21 a22 − K a2 n
k
. . .
=0
. . .
. . .
. . .
*
an1 an2 ann −
k
ka11 −  * ka12 K ka1n
n . . .
1
   . . . =0
k
. . .
kan1 kan 2 K kamn −  *

ka11 −  * ka12 K kann −  *

. .
 . ka22 −  L . =0
. .
kan1 kan 2 K kann −  *

 kA − * I = 0 ...(ii)
Since (ii) is the characteristic equation of kA, therefore it follows that the characteristics roots of (ii)
are k times characteristics roots of (i), i.e., the characteristics roots of kA are k1, k2, ... kn.
Example 5. If A and P be square matrices of the same type and if P be invertible, show that the
matrices A and P–1 AP have the same characteristic roots.
Solution. The characteristics matrix of P–1 AP is P–1 AP – I.
Now P–1 AP – I = P−1 ( A − P(I ) P−1 ) P,
= P–1 (A – I) P, ... (i)
since P (I) P–1 = (I) P P–1 = (I) I = I.
The characteristic equation of P–1 AP is | P–1 AP – I) | = 0. Using (i) we find that this equation is
the same is
| P–1 (A – I) P | = 0,
 | P–1 | | A – I | | P | = 0, since the determination of the product of two or more matrices
is equal to the product of the determines of the matrices.
 | P–1 | | P | | A – I | = 0, since | P–1 |, | A – I |, | P | commute
 | P–1 P | | A – I | = 0, since the determinant of a product is equal to the product of
determinants

137
  | I | | A – I |= 0, since P–1 P = 1
 | A – I | = 0, since | I | = 1
Since A P –  I | = 0  |A –  I | = 0, if follows that the characteristic roots of P–1AP are the
|P–1
same as the characteristic roots of A.
Exercise
1. Find the characteristics roots of each of the following matrices:
 1 2 3 1 −1 −1
(i)  0 −4 2  (ii) 1 −1 0  .
0 0 7 1 0 −1
   
2. Show that the matrices.
o g f o f h  o h g
g o h ,  f o g ,  h o f
f o   h o   g o 
 h g f
have the same characteristic equation.
3. Show that the matrices
a b c b c a  c a b
 b c a ,  c a b ,  a b c 
 c a b a b c  b c a
     
4. Show that the characteristics roots of A* are the conjugates of the characteristics roots of A.
5. Show that 0 is a characteristics root of a matrix if and only if the matrix is singular.
6. If A and B are n-rowed square matrices and if A be invertible, show that the matrices A–1B and B
A–1 have the same characteristic roots.

2. CAYLEY-HAMILTON THEOREM
We shall now establish the relation between a square matrix and its characteristics equation. The
relation, which is known after it discovers Arthur Cayley and W.R. Hamilton is known as Cayley-
Hamilton Theorem. It was established by Hamilton in 1953 for a special class of matrices and was stated
(without proof) by Caylay in 1858.
Theorem. Let | A – I | = (–1)n = {n + a1n − 1 + a2n − 2 + + an } be the characteristics polynomial
n −1
of an n + n matrix A. Then the matrix equation X + a1 X n
+ a2 X n − 2 + + an I = O is satisfied by X =
A.
Proof. The elements of A – I are at most of the first degree in . (In fact if we write the matrix in
full, we find that the diagonal elements are of the first degree in , and all the off-diagonal elements are
independent of ). The elements of adj (A – I) are (n – 1) × (n – 1) determinants, whose elements are at
most of the first degree in . Therefore adj. (A – I) is an n × n matrix whose elements are atmost of
degree n – 1 in . The matrix adj (A – I) can, therefore, be written as a matrix polynomial of degree n – 1
in .
Let adj (A – I) = B0n–1 + B1n–2 + ... + Bn–1, where B0, B1, ... Bn–1 are n × n matrices whose
elements are functions of aij’s (the elements of A).
Since (A – I) adj (A – I) = | A – I | I,.
Therefore(A – I) (B0n–1 + B1n–2 + ... + Bn–1) = (–1)n

138
 Comparing the co-efficients of like powre of  on both sides, we hve
– IB0 = (–1)n I,
AB0 – IB1 = (–1)n a1I,
AB1 – IB2 = (–1)n a2I,
....................................
ABn – 1 = (–1)n an I.
Premultiplying the above equations by An, An–1 ... I, respectively and adding the corresponding sides
of the resulting equation, we have
O = (–1)n [An + a1An–1 + ... + anI].
Thus An + a1An–1 + ... + a2An –2 + ... + anI = O, and this proves the theorem.
Remark. Calyey-Hamilton theorem is often expressed by saying that every square matrix satisfies
its characteristics equation.
We shall now illustrate the Cayley-Hamilton therorem with the help of an example.
 1 0 2
Example 6. Verify that the matrix A =  6 1 2  satisfies its characteristics equation.
 1 −2 0 
 
Soluiton. For the given matrix A,
1 −  0 2
| A – I | =  6 1−  2
 −2 − 
 1
= (1 – ) [{–  (1 – ) + 4} + 2 {– 12 – (1 – )}
= (1 – ) (2 –  + 4) + 2 ( – 13)
= – 3 + 22 – 3 – 22
The characteristics equation of A is therefore
– 3 + 22 – 3 – 22 = 0
or – 3 + 22 + 3 + 22 = 0 ...(i)
6 0 2  6 0 2 
Now A2 =  6 1 2  6 1 2 
 1 −2 0  1 − 0 
  
 3 −4 2
=  14 −3 14  ,
 −11 −2 −2 
 
 3 −4 2   1 0 2 
A3 =  14 −3 14   6 1 2
 −11 −2 −2   1 −2 0 
  
 −19 −8 −2 
=  10 −31 22 
 −25 2 −26 

Substituting A for  in the left-hand side of (i) we have

139
 A3 – 2A2 + 3A + 22 I
 −19 −8 −2   3 −4 2   1 0 2  1 0 0
=  10 −31 22  − 2  14 −3 14  + 3  6 1 2  + 22  0 1 0 
 −25 2 −26   −11 −2 −2   1 −2 0   0 0 1
      
0 0 0
=  0 0 0  = 0.
0 0 0
 
Thus we have
A3 – A2 + 3A + 22I = 0, showing that the matrix A satisfies its characteristic equation.

3. APPLICATION OF CAYLEY-HAMILTON THEOREM TO COMPUTE POWERS AND

INVERSE OF A GIVEN SQUARE MATRIX
Cayley-Hamilton theorem can be used to compute powers of a square matrix and invers of a non-
singular square matrix.
Let the characteristic equation of an n-rowed square matrix A be
n + a1n – 1 + a2n – 2 + ... + an = 0. ...(i)
By Cayley-Hamilton theorem the matrix A satisfies (i).
Therefore we have form (i),
Am + a1An – 1 + a2An – 2 + ... + anAm – n = 0. ...(ii)
Substituting m = n + 1, n + 2, ... n + m in (iii) we have the relations
An + 1 + a1An + a2An – 1 + ... + anA = 0. ...(iii)
An + 2 + a1An + 1 + a2An + ... + anA2 = 0.
.....................................................................................
An + k + a1An + k – 1 + a2An + k – 2 + ... + anAk = 0.
By substituting the values of A, A2, ... An in the first of the above equations, we get the value of An +
1. Similarly by substituting the values A2, A3, ... An + 1 in the second of the above equations we can get the

value of An + 2, and so on.

If the matrix A is non-singular, then an  0.
1 −1
But multiplying (ii) throughout by A we have
an

 1  n − 1  a1  n−2
a A + A + + A−1 = 0
 n   an 

 1   a   an − 1 
or A–1 =  −  An − 1 +  − 1  An − 2 + − I ...(iv)
 an   an   an 
By substituting the values of I, A, A2, ... An – 1 in the right-hand side of (iv) and simplifying we can
get the value of A–1.

140
Example 7. Find the characteristic equation of the matrix
 1 2 3
A =  2 3 4,
 1 0 −1 
 
and hence compute its cube.
Solution. The characteristic equation of the matrix A is
1−l 2 3
? 2 3− 4 = 0.
1 0 −1 − 
 (1 – ) (3 – ) (–1 – ) – {2 (–1 – ) – 4} + 3 (–3 + ) = 0
 3 – 32 + 8 = 0 ...(i)
By Cayley-Hamilton theorem by substituting A for  in (i), we have
? A3 –3A2 – 8A = 0,
so that A3 = 3A2 + 8A. ...(ii)
 1 2 3  1 2 3
Now A2 =  2 3 4   2 3 4 
 1 0 −1  1 0 −1
  
 8 8 8
? =  12 13 14  ...(iii)
 0 2 4
 
Substituting the values of A2 and A in (ii) we have
 8 8 8  1 2 3
A3 = 3 12 13 14  + 8  2 3 4 
 0 2 4  1 0 −1
   
 32 40 48 
=  52 63 74  .
 8 6 4
 
Example 8. Find the characteristics equation of the matrix
 1 0 2
A = 0 1 2
 1 2 0
 
and hence compute is inverse.
Solution. The characteristic equation of the matrix A is
1−  0 2
0 1−  2 =0
1 2 −
 (1 – ) {–(1 – ) – 4} + 2 (–1 + ) = 0
 3 – 22 – 5 + 6 = 0. ...(i)
By Cayley-Hamilton theorem, A satisfies (i) so that we must have
A3 – 2A2 – 5A + 6I = 0. ...(ii)

141
 Multiplying (ii) throughout by A–1 we have
A2 – 2A – 5I + 6A–1 = 0,
so that
6A–1 = –A2 + 2A + 5I
2
 1 0 2 1 0 2  1 0 0
= −  0 1 2 + 2 0 1 2 + 5 0 1 0
 1 2 0 1 2 0   0 0 1
    
 3 4 2 1 0 2  1 0 0
= −  2 5 2 + 2 0 1 2  + 5 0 1 0 
 1 2 6 1 2 0   0 0 1
    
 4 −4 2 
=  −2 2 2 
 1 2 −1 
 
4 −4 2 
1 
whence A–1 = −2 2 2  ,
6  1 2 −1
 
 2 2 1
 3 −3 3
 1 1 1 
= −
 3 3 3
 1 1 1
 − 
 6 3 6
Verification. By actual multiplication it can be easily seen that AA–1 = I.

4. APPLICATION OF CAYLEY-HAMILTON THEOREM : SIMPLIFICATION OF MATRIX

POLYNOMIALS
By using Cayley-Hamilton theorem of an n-rowed square matrix. A, a polynomial in A of degree
greater than n can be reduced to a polynomial in A of degree less than n, i.e., a polynomial of degree
greater than two in a 2 × 2 matrix A can be reduced to a linear polynomial in A, a polynomial of degree
greater than three in a 3 × 3 matrix A can be reduced to a quadratic polynomial in A, and so on. The
following example will illustrate the method.
 3 1
Example 9. If A =   , express 2A – 3A + A – 4I as a linear polunomial in A.
5 4 2
 −1 2 
Solution. The characteristic equation of A is
3− 1
= 0,
−1 2 − 
i.e., (3 – ) (2 – ) + 1 = 0,
i.e., 2 – 5 + 7 = 0.
The given polynomial is f (A), where
f () = 25 – 34 + 2 – 4.
By division alogrithm for polynomials we can write
f() = (2 – 5 + 7) (23 + 72 + 21 + 57) + 138 – 403

142
  f(A) = (A2 – 5A + 7I) (2A3 + 7A2 + 21A + 57) + (138A – 403 I) ...(i)
Since every square matrix satisfies its characteristic equation, therefore
A2 – 5A + 7I = 0.
Substituting from (ii) in (i) it follows that f(A) = 138A – 403 I.
Exercise
1. Verify that the matrix
 1 2 1
A =  −1 0 3  satisfies its characteristic equation.
 2 −1 1 
 
2. Obtain the characteristic equation of the matrix
 1 0 2
A =  0 2 1 and hence calculate its inverse.
 2 0 6
 
 1 2
3. If A =   , express
 −1 3 
A6 – 4A5 + 8A4 – 12A3 + 14A2 as a linear polynomial in A.
4. Evaluate the matrix polynomial A5 – 27A3 + 65A2 as a 3 × 3 matrix, where
 0 0 1
A =  3 1 0
 −2 1 4 
 
 1 0 0
5. If A =  1 0 1 
0 1 0
 
show that A3 – A = A2 – I.
Deduce that for every integer n  3, An – An – 2 = A2 – I. Hence otherwise determine A50

143
 Exercise-1
1. Show that the set
C2 = {(x1, x2) : x1  C, x2  C}
is a vector space over C with respect to co-ordinate wise addition and scalar multiplication.
2. Show that the set of all 2 × 2 matrices over C is a vector space over C with respect ot matrix
addition and multiplication of a matrix by a scalar.
3. Let V = {(a1, a2, a3, a4) : a1, a2, a3, a4 are integers}
Is V a vector space over R with respect to co-ordinatewise addition and scalar multiplication ?
Justity your answer.
4. Let V = {(x1, x2, x3,) : x1, x2, x3, are complex numbers, and x1, x2, = 0}
It V a vector space of over C with respect to co-oridnatewise addition and scaler multiplication ?
Justify your answer.
 x o  where y  C is a vector space over C with
5. Show that the set of all matrices of the form  ,
o y
respect to matrix addition and multiplication of a matrix by a scalar.
 a b  where a, b  C is a vector space over C.
6. Show that the set of all matrices of the form  ,
 −b a 
With respect to matrix addition and multiplication of a matrix by a scalar.

Exercise-2
1. For each of the following matrices A, verify that At = A:
 1 2 −1  −1 6 −7   a h g
 2 0 −4  ,  6 3 8  ,  h b f
 −1 −4 3   −7 8 −5   g c 
 f

2. For each of the following matrices A, verify that At = – A:

 0 i 1  0 3 1 + i   0 −3 2i 
 −i 0 −2  ,  −3 0 −4  ,  3 0 4 
 −1 2 0   −1 − i 4 0   −2i −4 0 
  
1 2 −2 
1 
3. If A = 2 1 2  , verify that A At = At = A = I3.
3  2 −2 −1
 
4. If A be any square matrix, verify that he matrix A + At is symmetric and the matrix A – At is
skew-symmetric.
5. If A be any square matrix, prove that A + A, AA, AA are all hermitian and A – A is
skew-hermitian.
 3 1 −7 
6. Express the matrix  2 4 8  as X + Y where X is symmetric and Y is skew-symmetric.
 6 −1 2 
 

144
7. Find remain matrices A and B such that
 1 1+ i 3
A+ iB =  4 2 1 − i
 −3 2 + 3i i 


Exercise-3
1. Apply the elementary operation R2  R3 to the matrix
 1 −1 3 
 4 2 −6 
 5 8 9
 
2. Apply the elementary operation C2 → 2C2 to the matrix
 −1 3 7 6 
 5 −1 4 −2 
 
3. Write down the elementary matrix obtained by applying R3 → R3 – 4R1 to I3.
4. Reduce the matrix
 1 2 −3 4 
 3 −1 2 0 
2 1 −1 5 

to triangular from by applying E-row operations
5. Reduce the matrix
 1 −1 −1
 4 1 0
 8 1 1
 
to I3 by E-row operations
6. Verify that the E-row operation R1 → R1 – R3 on the matrix AB is equivalent to the same E-row
operation on A, where
 1 −1 2  2 3 −5 
A = −3 1 4 , B = 1 −2 6  .
  
 0 1 3 3 1 1
  

Exercise-4
1. Find the inverse of the matrix
 3 −15 5
 −1 6 −2 
 1 −5 2 
 
by calculating its adjoint.
2. Show that the matrix
 1 −2 3 
 0 −1 4 
 −2 2 1

possesses an inverse.

145
  1 1 0
3. Given that A =  −2 −1 0  , calculate A2 and A–1
 0 0 2
 
 1 2 −1  2 1 2
4. If A = −4 −7 4 , B =  2 2 1 , verify that (AB)–1 = B–1A–1.
 
 −4 −9 5   1 2 2
   
5. By applying elementary row operations find the inverse of the matrix
 1 1 2
 3 1 −3  .
 2 1 −1 
 

Exercise-5
1. Find the rank of the matrix
 3 −1 2 −6 
 6 −2 4 −9  .
 
2. Find the rank of the matrix.
 −1 2 3 
 0 3 4 .
 0 0 0
 
3. Find the rank of the matrix
 1 1 1
 a b c .
 3 3

3
a b c 
4. Reduce the matrix
 1 −1 2 −3 
 4 1 0 2
 0 3 1 4 .
 
 0 1 0 2
to triangular form and hence find rank.
5. Find the rank of the matrix
 1 22 32 42 
 2 
2 32 42 52 
.
 32 42 52 62 
 42 52 62

7 2 


Exercise-6
1. Does the following system of equations possess a unique common solution?
x + 2y + 3z = 6,
2x + 4y + z = 7,
3x + 2y + 9z = 10.
It so, find it.

146
2. Show that x = y = z = 0 is the only common solution of the following system of equations:
2x + 3y + 4z = 0,
x – 2y – 3z = 0,
3x + y – 8z = 0.
3. Solve the system of equations:
x + y + z = 3,
3x – 5y + 2z = 8,
5x – 3y + 4z = 14.
4. Solve the following system of equations:
x – 3y + 2z = 0,
7x – 21y + 14z = 0,
–3x + 9y – 6z = 0.
5. Which of the following system of equations are consistent?
(a) x – 4y + 7z = 8,
3x + 8y – 2z = 6,
7x – 8y + 26z = 31.
(b) x – y + 2z = 4,
3x + y + 4z = 6,
x+y+z= 1.

Exercise-7
1. Find the characteristics roots of each of the following matrices:
 1 2 3 1 −1 −1
(i)  0 −4 2  (ii) 1 −1 0  .
0 0 7 1 0 −1
   
2. Show that the matrices.
o g f o f h  o h g
g o h ,  f o g ,  h o f
f o   h o   g o 
 h g f
have the same characteristics equation.
3. Show that the matrices
a b c b c a  c a b
 b c a ,  c a b ,  a b c 
 c a b a b c  b c a
     
4. Show that the characteristics roots of A* are the conjugates of the characteristics roots of A.
5. Show that 0 is a characteristics root of a matrix if and only if the matrix is singular.
6. If A and B are n-rowed square matrices and if A be invertible, show that the matrices A–1B and B
A–1 have the same characteristics roots.

147
 Exercise-8
1. Verify that the matrix
 1 2 1
A =  −1 0 3  satisfies its characteristic equation.
 2 −1 1 
 
2. Obtain the characteristic equation of the matrix
 1 0 2
A =  0 2 1 and hence calculate its inverse.
 2 0 6
 
 1 2
3. If A =   , express
 −1 3 
A6 – 4A5 + 8A4 – 12A3 + 14A2 as a linear polynomial in A.
4. Evaluate the matrix polynomial A5 – 27A3 + 65A2 as a 3 × 3 matrix, where
 0 0 1
A =  3 1 0
 −2 1 4 
 
 1 0 0
5. If A =  1 0 1 
0 1 0
 
show that A3 – A = A2 – I.
Deduce that for every integer n  3, An – An – 2 = A2 – I. Hence otherwise determine A50.

148
 UNIT 3

LESSON 1

GROUPS

In this chapter we study the properties of an important algebraic structure called Groups.

An algebraic structure is any set which is endowed with one or more binary operations satisfying
some given axioms.
The motivation for the study of these structures stems from our familiarity with the number
system. In the set I, of all integers, we have the following properties of addition:
1. a + b I  a, b  I (Closure Axiom)
2. a + (b + c) = a + (b + c)  a, b, c I (Associative Law)
3. a+0=aaI (Existence of identity)
4. a + (–a) = 0  a  I (Existence of inverse)
Similarly in the sets of all non-zero rational, real or complex numbers, the operation of
1
multiplication has similar properties, with 1 playing the role of the identity and 𝑎
playing the role of the
inverse of any element a.
We would like to introduce a binary operation in any abstract set, satisfying the properties
mentioned above.
Definition and examples of Groups
Let G be an arbitary set in which there exists a binary operation () satisfying the following
axioms.
G1 a  b  G  a, b  G
G2 a  (b  c) = (a  b)  c  a, b, c  G
G3  e  G such that a  e = e  a = a  a  G
G4 To every a  G,  a'  G a  a' = a'  a = e
Then G together with this operation () is called a group. We say that (G, ) is a group. Note that a
group is actually a pair – the set together with the operation. With respect to the binary operation, we may
say that G is a group.
Obvious examples of groups are [I, +], [Q – {0},.], {R –[0],.} {C –[0],.} where I, Q, R and C are
the sets of all integers, rational numbers, real numbers and complex numbers respectively.
We give some other examples of groups.
Example 1. Let G be the set of all even integers i.e. G = {2 p, p  Z} where Z is the set of
integers. It is easy to see that G forms a group with respect to addition. The closure and associative

149
axioms are obviously true. 'o' serves as the identity element and for any 2p  G, –2p is the additive
inverse and as such all the group axioms hold. Therefore the set of all even integers form a group under
ordinary addition.
Example 2. Consider the set S of all fourth roots of unity.
i.e. S = [1, –1, i, –i] where i = √−1
Let the binary composition defined on S be ordinary multiplication obviously closure and
associative axioms hold in S. It is easy to see that 1 is the identity element. The inverse of 1 is 1 itself. –1
is the inverse of itself. The inverse of i is –i and that of –i is i [ i2 = –1]
 S forms a group under ordinary multiplication.
Example 3. Let G be the set of all non-singular n × n matrics, with real or complex entries. Let the
group operation be the usual matrix multiplication. it is easy to see that all the axioms are satisfied. The
unit matrix is the identity and the inverse of any matrix is the inverse.
Example 4. Let (G, ) be the group given by the following:
 e x y z
e e x y z
x x e z y
y y z e x
z z y x e
TABLE-1

Verify that all the group axioms are satisfied.

Example 5. Let (G, ) be the group given by the following table where e is the identity element.

 e a b c
e e a b c
a a e c b
b b c a e
c c b e a

TABLE-2
It is easily verified that all the group axioms are satisfied.
Example 6. Consider the following Table 3:
 e a b c d
e e a b c d
a a e c d b
b b d e a c
c c b d e a
d d c a b e

150
 TABLE-3

Note that (b  c)  d = a  d = b

and b  (c  d) = b  a = d

so that the associative property is not satisfied. Thus the set {e, a, b, c, d} under the given operation is not
a group.

Example 7. Consider the set of all vectors in space. Let the binary operation be the usual scalar or
dot product. Then this set together with this operation is not a group. The dot product of two vectors,
being a scalar, the closure property G1 is not satisfied.

Example 8. Consider the set V of all vectors in space, together with the operation a  b = a × b
(the cross product). This is not a group as a × (b × c)  (a × b) × c i.e. Associative Axiom is not true.

Remark: Note that one of the properties of addition and multiplication in the number system,
namely the commutative property (a + b = b + a, ab = ba) has not beeh stipulated as an axiom in the
definition of a group.

If a group (G, *) satisfies the additional property a * b = b * a  a, b  G, then G is called an

Abelian or Commutative of a group.

[I, +], [Q – {0},], {R –[0],} {C –[0],} are all Abelian groups.

The groups of Ex. 3 is non Abelian as matrix multiplication is not communtative.

Remark: If the group G is not abelian, it is not true that a * e = e * a  a, b  G. However, note
that G3 requires that the identity should commute with all elements of G. The axiom G4 requires that
element a should commute with its inverse.

Remark: Note that G3 merely states that  e  G satisfying a * b = b * a  a  G. It does not

insist on the uniqueness. For all that we know, there may be more than one element satisfying the
requirement of G3. Similarly G4 does not insist on the uniqueness of the inverse element of any a  G.

We now prove that the axioms G1 to G4 ensure the uniqueness of the identity and the inverse.

Theorem: The identity of a group is unique.

Prof: Suppose  two elements e,   G such that

e*a=ae=aaG

and a=a=aaG

Treating e as in identity, e   = 

and treating f as an identity, e   = e

 =e=e

151
 So that the identity of a group is unique.
Theorem 2: The inverse of any element a  G is unique.
Prof: Suppose  ā and a' such that
āa=aā=e
and a'  a = a  a' = e
We have ā = ā  e = ā  (a  a') = (ā  a)  a' = e'  a' = a'
so that ā = a'
Thus the inverse of any element is unique.
The inverse of any element is generally written as a–1. If the group operation is +, we denote the
inverse of a by –a.
Remark: (a–1)–1 = a and (ab)–1 = b–1 a–1, for abb–1 a–1 = aea–1 = aa–1 = e
Definition: if the number of elements in a group G is finite, we say that G is a finite group.
Otherwise, G is said to be an infinite group.
The groups of examples, 2, 4, 5 are finite groups. The groups of example 1 and 3 are infinite
groups.
Definition: Powers of an element.
Let (G, ) be a group. Let x G.
We define x0 = e and xn = x  x  ......  x, n times, for any positive integer n.
Remark: In (I, +), x0 = 0 for any integer x.
and xn = x + x + ... + n times
= nx.
DEFINITION:
1. Cyclic groups
If every element x  G can be expressed as an for some fixed a  G and some +ve integer n, then
G is said to be cyclic group. 'a' is said to be a generator of the group.
The set of all integers under addition is a cycle group. 1 is a generator.
2. Subgroups

Let (G, ) be any group. A subset S of G is said to be a subgroup is the elements of S satisfy all the
axioms of a group under the same binary composition as in G.
Example 1: In (I, +), the set of all even integers forms a subgroup. The set of all multiples of any
given integer is also a subgroup.
Example 2: Consider the group of Table 2.
The subset {a, e} is a subgroup.

152
 The subset {b, e} is not a subgroup. b  b = a and hence the closure axiom is not satisfied.
For every group G, the singleton set {e} and the entire set G, are always subgroups. Thus every
group contains atleast two subgroups. These two subgroups are called trivial subgroups.
A subset S of G is said to be a proper subgroup if it is a non-trivial subgroup. i.e., S  {e}, S  G.
In order that a subset S of a group G, be a subgroup, it is necessary that all the four axioms
G1-G4 are satisfied in S. However, that associative law is automatically satisfied, as every element of S is
also an element of G. Thus it is only necessary to verify the other 3 axioms. However, the next theorem
shows that it suffices to verify only two axioms.
Theorem 3: Let (G, ) be group. Let H be a non-empty subset of G. Then H is a subgroup of
G if,
(i) a, b  H  a  b  H
(ii) a  H  a-1  H
Proof: If H is a subgroup, the conditions (i) and (ii) are axiomatically satisfied.
Conversely, let (i) and (ii) be satisfied. As the associative property is automatically satisfied, we
have to only verify the existence of the identity in H.
By (ii) a  H  a–1  H. Also a  a–1  H by (i) i.e. e  H.
i.e. identity e exists in H.
 H is a subgroup of G.
Theorem 4: If H is a nonempty finite subset of a group G, and H is closed under multiplication
then H is a subgroup of G.
Proof: Let a  H. Since H is closed under multiplication
 a2, a3, a4 .............................. all belong to H.
Again H being a finite set, this is possible only if there exists integers m and n such that am = an,
where m > n > 0.
 am–n = e,  e  H
Also m>n>0m–n1m–n–10
 am–n–1 H
Also a. am–n–1 = am–n = e
Similarly am–n–1 . a = a m–n = e,
 am–n–1 is the inverse of a  H
For every a  H, a–1 = am–n–1  H
 H is a subgroup G.

153
 Let G be a given group and H and K be two subgroups of G. What can we say about the subset
H  K. It is interesting to note that H  K need not necessarily be a subgroup of G. To illustrate this let
us consider the set I of all integers which is a group under addition.
Let I2 = [2P | p  I ] and I3 = [3q | q  I ] be subsets of I obviously I2, I3 are subgroups of I under
addition composition. But I2  I3 is not a subgroup of I.
∵ We see that 2, 3  I2  I3. But 2 + 3 = 5  I2  I3, and as such closure axiom is not valid in
I2  I3 we conclude that the union of two subgroups of a given group G need not necessary be a subgroup
of G. However the intersection of two subgroups of a given is necessarily a subgroup as proved in the
ensuing Theorem.
Theorem 5: Let G be a group. Let H, K be 2 subgroups of G. Then S = H  K is also a subgroup.
Proof: Let a, b,  S
Then a, b  H, a, b  K
As H is a subgroup, ab  H.
Similarly ab  K
Hence ab  H  K = S
Similarly a  S  a  H, a  K
So that a–1  H, a–1  K as H and K are subgroups of G
Hence a–1  H  K = S
Conditions (i) and (ii) of theorem 3 are satisfied.
Hence S is a subgroup
Note: We will write ab for a  b in future.
Definition: Congruence modulo a subgroup.
Let G be a group and let H be a subgroup of G. Define a relation, called congruence modulo H,
written as, a  b (mod H). [a is congruent to b modulo H]
if ab–1  H.
Theorem 6: The relation a  b (mod H) is an equivalence relation in G.
Proof: We show that this relation satisfies the properties of reflexivity, symmetry and transitivity.
Reflexivity: a  a (mod H)
for aa–1 = e  H. (H, being a subgroup, e  H)
Symmetry: Let a  b (mod H)
Then ab–1  H. As H is a subgroup, (ab–1) –1 = ba–1  H.
i.e. b  a (mod H).

154
 Transitivity: Let a  b (mod H) and b  e (mod H).
i.e. ab–1  H and bc–1  H.
Because of the closure property in H, (ab–1) (bc–1) = ac–1  H
 a  c (mod H).
Hence the relation of congruence is an equivalence relation.
Theorem 7: A non-empty subset H of a group G is a subgroup of G if ab–1  H for all a, b  H
Proof: Firstly we show that the condition is necessary. Let H be a subgroup of G and a, b  H.
Since H itself is a group.
 b  H  b–1 H
Again a  H, b–1  H  ab–1  H
Now we shall show that the condition is sufficient as well.
Let ab–1  H  a, b  H
Putting b = a in the above we get aa–1 = e  H i.e. identity exists in H.
Taking a = e, we find that eb–1 = b–1  H  b  H
Let a, b  H  a H, b–1  H.  a(b–1)  H i.e. ab H
H is a subgroup of G.

SOLVED EXAMPLES
Example 1: If a group G is such that (ab)2 = a2 b2 a, b  G, prove that G is abelian.

Solution: We have (ab)2 = a2b2

i.e. abab = aabb
Premultiply by a–1 and postmultiply by b–1, on both sides.
a–1 (ab ab)b–1 = a–1 (aabb)b–1 Using Association axiom and remembering
a-1 a = e, eb = b etc. ]
ba = ab  a, b  G
Hence G is abelian.
Example 2: Prove that in a group G, the equation
ax = b and xa = b
have unique solution
Solution: Consider the equation
ax = b.

155
It is easily verified that a–1 b is solution, for
a(a–1b) = (aa–1) b = eb = b
To prove uniquences, suppose x and y are solutions.
Then ax = ay = b
Premultiplying by a–1, we have
a–1(ax) = a–1 (ay) = a–1 b
(a–1 a) x = (a–1 a)y  ex = ey
giving x = y
Example 3: Let Q be the set of non-zero rational numbers and let the group operation be the
usual multiplication. Is the subset of integers I, a subgroup of Q ?
Solution: Note that the subset I satisfies the three axioms G1, G2 and G3 required of a group
namely, the closure axiom, the associative property and the existence of the identity.
1
However, the inverse of any integer a (under multiplication) is which does not belong to I.
𝑎

Hence I is not a subgroup of (Q,.)

Example 4. Let G be the group of all 2×2 non singular matrics with real entries.
Let
𝑎 𝑜
𝑆 = [( ) 𝑎, 𝑏, 𝑐, ∈ 𝑅, 𝑎𝑐 ≠ 0]
𝑏 𝑐
Prove that S is a subgroup of G, the group operation being the usual matrix multiplication.
Solution: (i) The closure axiom is satisfied, for if A, B  S
𝑎1 𝑜 𝑎2 𝑜
𝐴 = (𝑏 𝑐1 ) 𝐵 = ( 𝑏2 𝑐2 )
1

𝑎1 𝑎2 𝑜
𝐴𝐵 = (𝑏 𝑐1 𝑐2 )
1 𝑎2 +𝑐1 𝑏2

a1c1  0, a2 c2  0  a1a2c1c2  0
hence AB  S

(ii) The identity (10 01)  S

𝑎 0
(iii) Consider any element A of S where A = ( )
𝑏 𝑐
1 𝑐 0
(
𝑎𝑐 −𝑏
) which is also an element of S is the inverse of A.
𝑎
Note the characteristic property of elements of S is that the entry in the first row and second
column should be zero.
Hence S is a subgroup of G.

156
 LESSON 2

SUBGROUP

Let G be a group and H, a subgroup of G. Let a  G. The set Ha = (h a|h  H) is called a right coset of H
in G. Similarly the set

a H = (h a|h  H) is called a left coset of H in G

In an abelian group, as ha = ah, the distinction between left and right cosets disappears. However,
if G is non abelian, the coset Ha, will in general, be different from aH.
Example 1 : Consider the group (I, +) and the subgroup H, consisting of all multiples of 4. (As the
group operation is +, we will denote a coset by H + a, instead of Ha). The coset H + 1 is
H + 1 = (................, –3, 1, 5, 9, ................)
H + 2 = (................, –2, 2, 6, 10, ................)
Note that the cosets H + 1, H + 2, are merely subsets of I.
They are not subgroups.
H + 5 = (................, –3, 1, 5, 9, ................)
Note that the coset H + 1 = the coset H + 5.
We thus see that, in general, the cosets Ha and Hb may be equal even if a  b.
Note also that if h  H, Hh = H
Theorem 1 : Let G be a group and H a subgroup of G. If a, b  G, then the cosets Ha and Hb are
equal if ab–1  H i.e. a  b (mod H)
In other words Ha = Hb  a  b (mod H)
Proof : Firstly we shall show that Ha = Hb  a  b (mod H)
Ha = Hb  h1 a = h2b for some h1, h2  H
 a = h1–1 h2 b = h3 b (say) [ ∵ h1–1 h2  H]
 ab–1 = h3 bb–1 = h3 i.e. ab–1  H or a  b (mod H)
Conversely if a  b (mod H)
i.e. ab–1  H.
Let ab–1 = h2
Consider any element x  Ha
x  Ha  x = h1 a for some h1 a  H
 xb–1 = h1 ab–1

157
 = h1 (ab–1) = h1 h2 = h3 (say) [ab–1 = h2]
 x = h3b  x  Hb
It follows that Ha  Hb. Similarly, it can be shown that Hb  Ha
Ha  Hb, Hb  Ha  Ha = Hb
 Ha = Hb  a  b (mod H)
Theorem 2 : Two right (or left) cosets of H are either identical or disjoint i.e. If Ha, Hb are any
two right cosets of H in a group G then Ha = Hb or Ha  Hb = .
Proof : If Ha  Hb = , there is nothing to prove. Let us assume that Ha  Hb  . We shall show
that Ha = Hb.
Let x  Ha  Hb
 x  Ha, x  Hb
 h1h2  H such that
x = h1a, x = h2b
so that h1 a = h2b
Premultiplying by h1–1, we have
h1–1h1a = h1–1h2b
i.e. a = (h1–1h2) b = h3b where h3  H
 a = h3b
 ab–1 = h3  H
 a  b (mod H)
 By theorem 1, Ha = Hb
Similarly it can be shown that if aH, bH are any two left cosets H of in a group G, then aH = bH or
aH  bH = .
Theorem 3 : Let G be a group. Let He be a subgroup of G. Let Ha, Hb be two distinct cosets of H
in G. Then there exists a one-one correspondence between Ha and Hb.
Proof : Any element x of Ha is of the form ha, for some h  H.
If x  Ha, define a function f as follows.
f(x) = hb. where x = ha
 f(x)  Hb  f is a function from Ha to Hb.
f is obivously onto, for f–1(hb) = ha, for any element hb of Hb.
f is one-one. For if h1 a and h2 a are two different elements of Ha, their images h1 b and h2 b are
different. i.e. there exists a one-one correspondence between Ha and Hb.

158
 Remark : The above theorem implies that if G is a finite group, then the number of elements in
any two cosets is the same.
Definition : Order of a group.
If G is a finite group, the number of elements in the group is called the order of G, written as o(G).
Lagrange’s Theorem : Let G be a finite group and H any subgroup of G. Then the order of H
divides the order of G. i.e. o(H) divides o(G).
Proof : Let o(G) = n, o(H) = m
Let H = {h1, h2 ....................... hm}. Consider any element a  G. Firstly we shall prove that
o(aH) = m  a  G.
aH = {ah1 ah2 .............a hm}.
It is easy to see that the elements of aH are all distinct.
 ahi = a hj, i  j
 hi = hj [by the cancellation law in a group] which is a contradiction for H is a finite group of
order m  o(aH) = m
Let all possible distinct left cosets of H in G be a1 H, a2 H, ................ ak H
where a1 a2 .................. ak  G
We shall prove that
G = a1H  a2H  ..................  ak H.
obviously a1H  G, a2H  G .............. ahH  G for each coset of H in G is a subset of G.
 a1H  a2H  a3H  ....................  akH  G
Let g  G, be any element of G.
Then g  aiH for some i satisfying 1  i  k
For otherwise g  g H for where g H is a left coset of H distinct from a1H,
a2H, .......ak H which is a contradiction
 G  a1H  a2H  a3H  ...................  akH
 G = a1H  a2H  a3H  ...................  akH
 o(G) = 0(a1H) + 0(a2H) + ...........0(akH)
= k0(H)  0(H) divides 0(G)
Definition : Index of a subgroup.
Let G be a group and H, a subgroup of G. The number of distinct cosets of H in G is called the
index of H in G.
If G is a finite group, the index of a subgroup H = 0(G) / 0(H).

159
 Note, however, that the notion of index is valid even if G is an infinite group.
For example in (I, +), if H is the subgroup consisting of multiples of 4, the distinct cosets of H in I
are H, H + 1, H + 2 and H + 3.
Hence the index of H in I is 4

Normal subgroups
While defining the concept of cosets, we have observed earlier that left cosets are in general,
different from right cosets. However, subgroups for which left cosets are identical with right cosets
deserve to be studied specially. A subgroup H of G which is such that every left coset is also right coset
will be called a Normal subgroup. However, we start the definition of normal subgroups in a slightly
different manner and ultimately prove that the two definitions are equivalent.
Definition : A subgroup N of a group G is said to be a normal subgroup of G if gng–1  N  g 
G and n  N
Remark : If G is abelian, gng–1 = gg–1n = n  N  g  G, n  N
Hence every subgroup of an abelian group is a normal subgroup.
In an arbitrary group, if N is a normal subgroup, gng–1 is some element of N
Remark : It is not necessary that gng–1 is equal to n. However, if G is abelian then the element
–1
gng is the same as n.
We state some important consequences of our definition in the form of a series of theorems.
Definition : Let G be any group and N is a subgroup of G, we define
gNg–1 = {gng–1 : n  N}
where g is any element of G.
Theorem 4 : N is a normal subgroup of a group G if gNg–1 = N g  G
Proof : In order to show that gN–1 g = N, it is sufficient to prove that gN–1g  N and N  gNg–1.
From the definition of normal subgroup it is clear that
gng–1  N g  G, n  N  gNg–1  N
We have only to show that N  gNg–1. Now g  G  g–1  G and N being a normal subgroup of
G it follows that
g–1ng  N where n is any element of N.
Let g–1ng = n1 where n1  N
 g(g–1ng) = gn1
 (gg–1) (ng) = gn1
 eng = gn1 [where e is the identity in G]
 (ng) g–1 = gn1g–1

160
  n = gn1g–1
 N  gNg–1
Now gNg–1  N, N  gNg–1  gNg–1 = N
Theorem 5 : N is normal in G if every left coset of N is also a right coset of N, in G.
Proof : Let N be normal in G.
By Theorem 5, gNg–1 = N  g  G
Consider any n  N. From the definition of normal subgroup, gng–1  N for all g  G
let gng–1 = n1 where n1  N
 gng–1g = n1g
 gn = n1g
 gN = Ng
i.e. every left coset of N is also a right coset of N in G. Conversely if gN = Ng  g  G we can show that
N is a normal subgroup of G. Consider any n  N. For a given g
gN = Ng  There exists n1  N such that gn = n1g
 gng–1 = n1gg–1 = n1
i.e. gng–1  N
 N is a normal subgroup of G.
This theorem establishes the equivalence of the two definitions of a normal subgroup.
Theorem 6 : If G is a group and N, a subgroup of index 2 in G, then N is a normal subgroup of G.
Proof : Let g be any element of G. If g  N then obviously gN = N = Ng
If g  N, then gN  N, Ng  N
From the definition of cosets, it follows that
𝑔𝑁  𝑁 = 
[ ] ..........(1)
𝑁𝑔  𝑁 = 
Since the index of N in G is 2, we have
𝑁  𝑔𝑁 = 𝐺
[ ] ..........(2)
𝑁  𝑁𝑔 = 𝐺
Using (1) we see that gN = Ng  g  G.
 N is a normal subgroup of G.
Theorem 7 : The intersection of two normal subgroups of G is again a normal subgroup of G.
Proof : Let H, K be 2 normal subgroups of G.

161
 Let N = H  K
Let n  N, g  G
Then n  H, n  K
As H is normal, gng–1  H
As K is normal, gng–1  K
so that gng–1  H  K = N
Hence N is normal in g.
Theorem 8 : Let G be any group and H, an arbitrary subgroup of G and N, a normal subgroup of
G. Then H  N is a normal subgroup of H
Proof : Let x  H  N. Let h  H.
hHhG [∵ H < G]
x  H  N  x  H, x  N
The hxh–1  N as N is normal in G.
Moreover, hxh–1  H as h, x and h–1 are in H.
 hxh–1  H  N  h  H, x  H  N
Hence H  N is a normal subgroup of H.

SOLVED EXAMPLES
Example 1 : Let G be the group of integers (I, +). Let S be the subgroup containing multiples of 3.
List out the different cosets of S in I. What can you say about the cosets S + 1 and S + 7?

Solution : The different cosets of S in I are

S + 0 = S = { ........, –3, 0, 3, 6, .........}
S + 1 = { ........., –5, –2, 1, 4, 7, 10, ..........}
and S + 2 = { ........., –4, –1, 2, 5, 8, 11, ..........}
Note that I = S  S + I  S + 2.
The index of S in I is 3, as there are 3 distinct cosets.
Note also that 7  1 (mod S)
as 7 * 1–1 = 7 + (–1) = 6  S.
(Here * is + and a–1 = –a)
Hence S + 7 = S + 1
Example 2 : Let G be a group and let a  G be a fixed element of G. Consider the set
S = [x  g / xa = ax]

162
 Then S is a subgroup of G.
Solution : In order to prove that S is a subgroup of G, it is just sufficient to verify that
 x1, x2  S, x1x2  S
Also x  S  x–1  S
Let x1, x2  S. Then by definition of S.
𝑎𝑥1 = 𝑥1 𝑎

𝑎𝑥2 = 𝑥2 𝑎
]

Now
a(x1x2) = (ax1)x2
= (x1a)x2
= x1 (ax2)
= x1 (x2a) = (x1x2)a
a(x1x2) = (x1x2)a  x1x2  S
i.e. Closure Axiom is valid in S
Also x  S  xa = ax
 x–1xa = x–1ax
[x  G,  x–1 exists and xx–1 = x–1 x = e
where e is the identity in G]
 ea = (x–1a)x
 ea x–1 = (x–1a)xx–1
 ax–1 = x–1a
 x–1  S
 S is a subgroup of G.

EXERCISES
1. Show that each of the following sets is a group with respect to multiplication.

(a) {1}, (b) {1, –1}, (c) {1, , 2} where 3 = 1 (d) {1, i –1, –i} (i2 = –1).

2. Check if the following sets are groups.

(a) The set of all positive integers with respect to addition.

(b) The set of all rational numbers with respect to multiplication. present

(c) The set of all odd integers with respect to addition.

163
3. Let I be the set of integers. Define an operation (o) by a o b = a + b – 1. Prove that (I, o) is a group.
Identify the identity of I and the inverse of any element a  I.

4. Let G be the group of all 2 × 2 non singular matrices with real entries, under the binary operation of
the usual matrix multiplication. Let S consist of all matrices of the form

(𝑎𝑏 01), 𝑎  0. Check if S is a subgroup of G.

5. Let G = {0, 1, 2, 3, 4, 5} and let the binary operation be addition modul 6. Prove that G is a group.
Prove that the subset S = {0, 2, 4} is a subgroup of G. Write down the different cosets of S.

6. In the group of ex 5, consider the subset S1 = {0, 2, 3, 4}. Is S1 a subgroup of G?

164
 LESSON 3

RING

Introduction : In the first two lessons we discussed Groups which are Algebraic structures only
one binary composition satisfying certain axioms. In the present lesson we introduce the notion of a ring
which is an algebraic structure having two binary compositions satisfying certain axioms. We give the
formal definition of a ring as follows:

Definition : A non-empty set R with two binary compositions * and o is to be ring with respect to
these compositions if the following axioms hold.
R–1 For all a, b, c  R
a * (b * c) = (a * b) * c
R–2 There exists an element e  R
such that a * e = e * a = a  a  R
R–3  a  R, there exists b  R
such that a * b = b * a = e
R–4  a, b  R, a * b = b * a
R–5  a, b  R, a o (b o c) = (a o b) oc
R–6  a, b c  R
i) a o (b * c) = a o b * a o c
ii) (b * c) o a = b o a * c o a
It is a common practice to use the symbols + and reply for * and a and these compositions are
referred to as addition and multiplication respectively. You will do well to note that when we use the
symbols + and * for the compositions in a ring or refer to them as addition and multiplication, we do not
have to confuse them with addition and multiplication of numbers. They may be quite different from
ordinary addition and multiplication of numbers. They may stand for addition, multiplication of ordered
pairs, or vectors, or of matrices or they may be some other compositions. We choose to call them as
addition and multiplication for reasons to be seen later.
When we use the symbols + and * instead of * and o respectively we give a simple name to the
element e introduced in R-2. We call it the zero of the ring R and denote it by o of course the o of a ring
has nothing to do with the number zero of the number system. In view of our simplified notation as
introduced above, the definition of a Ring could be as follows:
Definition : A non-empty set R along with two binary composition + and * is said to be a Ring if
the following conditions hold:
R–1 For all a, b, c  R [Associative Law]
a + (b + c) = (a + b) + c

165
 R–2  o  R such that  a  R
a+o=o+a=a [Existence of Zero]
R–3 For each a  R,  b  R such that a + b = b + a = o
(Existence of additive inverse)
R–4  a, b  R, a + b = b + a [Commutative Law of addition]
R–5 For all a, b, c  R,
a(bc) = (ab)c [Associative Law of Multiplication]
R–6 For all a, b, c  R
i) a(b + c) = ab + ac [Left Distributive Law]
ii) (b + c)a = ba + ca [Right Distributive Law]
The element b  R in R–3 is called the additive inverse or the negative of ‘a’. We noticed that the
Axioms R–1 to R–4 are simply the Axioms for an Abelian group. We may state that a triple (R, +,.) is
said to be Ring if (R, +) is an Abelian group, (R,.) is a semi-group and multiplication is left as well as
right distributive with respect to addition.

SOME SPECIAL TYPES OF RINGS

1. Commutative Ring

A Ring R is said to be commutative if  a, b  R, ab = ba

A ring is said to be non-commutative if multiplication is not commutative. For a ring to be non-
commutative it is just sufficient to see if the ring possesses at least one pair of elements a, b for which ab
 ba.
In a commutative Ring there is no distinction between left distributive and Right distributive laws
for each implies the other. In such a ring R, R–6 would read as follows. For all a, b, c  R
a(b + c) = ab + ac (Distributive Law)
2. Ring with unity (or Identity)
R is said to be a ring with unity (or Identity) if there exists 1  R such that a.1 = 1.a = a  a  R
1 is called the Identity element of the Ring.
3. Ring without Zero divisors
A Ring (R, +,.) is said to be without zero divisors if the following condition holds.
For all a, b  R, ab = o  a = o or b = o
We say that a ring (R, +,.) has zero divisors if for some pair a, b of elements of the ring which are
both different from the zero of the ring, we have
ab = o

166
4. Integral Domain
A commutative ring with Identity and without zero divisors is called an Integral Domain.

SOME EXAMPLES OF RINGS

Example 1. The set Z of integers is a commutative ring with unit element under the usual addition
and multiplication composition. Similarly the set R of all real numbers and the set C of all complex
numbers are commutative rings with identity 1 under the usual addition and multiplication of real
numbers and complex numbers respectively.

Example 2. The set E of all even integers is a commutative ring without identity under the usual
addition and multiplication of numbers.
Note : The set F of all odd integers is not even group under addition, for the closure axiom is not
valid [ The sum of two odd integers is an even integer which is not a member of F]
Example 3. The set S of all matrices of the form
𝑥 𝑜
(𝑦 𝑜) , 𝑥, 𝑦𝑄,
where Q is the set of rational numbers is a non-commutative ring without identity with respect to addition
and multiplication of matrices. Consider any two elements of S.
𝑥1 𝑜 𝑥2 𝑜
A = (𝑦 𝑜) , 𝐵 = (𝑦2 𝑜)
1

𝑥 +𝑥 𝑜
Then A + B = (𝑦1 + 𝑦2 𝑜)
1 2

[ x1 + x2 y1 + y2  Q for all x1 x2 y1 y2  Q]
Obviously Associative Law is true for matrix addition.
i.e. For All A, B, C  S
(A + B) + C = A + (B + C)
𝑜 𝑜
The matrix ( ) serves as the additive identity in S.
𝑜 𝑜
𝑥1 𝑜 –𝑥 𝑜
If A = (𝑦 0), then 𝐵 = ( 1 )
1 – 𝑦1 0
is the additive inverse of A.
Also A + B = B + A  A, B  S
for matrix addition is commutative
𝑥1 𝑜 𝑥2 𝑜
If A = (𝑦 0) , 𝐵 = (𝑦2 0)
1
𝑥1 𝑥2 0
then AB = ( ) 𝑆
𝑦1 𝑥2 0

167
 𝑥2 𝑥1 0
Also BA = ( )
𝑦2 𝑥1 0

AB  BA
For all A, B, C  S
A (BC) = (AB)C
as matrix multiplication is associative
𝐴(𝐵 + 𝐶) = 𝐴𝐵 + 𝐴𝐶
Also [ ]  𝐴, 𝐵, 𝐶 𝑆
(𝐵 + 𝐶)𝐴 = 𝐵𝐴 + 𝐶𝐴
Since matrix multiplication distributes over addition so we conclude that S is a ring. The ring is non
commutative and does not have the identity element.
Example 4. The set S = {0, 1, 2, 3, 4} is a commutative ring with unity I under addition and
multiplication modulo 5.
Definition : We define the binary compositions n and ʘn in the set Z of integers as follows:
a n b = least non negative remainder obtained when a + b is divided by n.
a ʘn b = least non negative remainder obtained when a b is divided by n.
The compositors n and ʘn are known as addition and multiplication modulo n respectively. Now
consider the set S = {0, 1, 2, 3, 4}. It is easy to verify that
2 5 4, since the least non negative remainder.
when, 2 + 4 = 6 is divided by 5 is 1.
2 5 3 = 0, as the least non negative remainder when 2 + 3 = 5 is divided by 5 is zero.
Also 0 5 1 = 1, 2 5 0 = 2 etc.
 0 serves as the identity element under the operation addition modulo 5.
2 5 3 = 0, 1 5 4 = 0  2 is the inverse of 3 and 4 is the inverse of 1 under addition modulo 5.
Also the inverse of 5 is 0 5 5 0 = 0
Also 2 ʘ5 3 = 1, 4 ʘ5 2 = 3 etc.
 The set S = {0, 1, 2, 3, 4} is a ring under the operations of addition and multiplication modulo 5.
You may easily verify that the associative law is valid under the binary operations addition and
multiplication modulo 5. 1 serves as the identity element under multiplication modulo 5.
Example 5. Let C = {(a, b) : a, b  R} where R is set of all real numbers i.e. C is the set of all
complex numbers. We define addition and multiplication is C as follows:
(a, b) + (c, d) = (a + c, b + d)
(a, b) . (c, d) = [ac – bd, ad + bc]
Then C is a commutative Ring with identity (1, 0) with respect to the operations of addition and
multiplication. Also (0, 0) is the zero of C and the additive inverse of (a, b) is (–a, –b).

168
 The associative law of multiplication and one of the distributive laws are verified as follows:
(a, b) [(c, d) – (e, f)] = (a, b) . (ce – df, cf + de)
= [ace – adf – bcf – bde, acf + ade + bce – bdf]
Also {(a, b) . (c, d)} (e, f) = {ac – bd, ad + bc} . (e, f)
= {ace – bde – adf – bcf, acf – bdf + ade + bce}
and these are equal elements of C. Again, for one of the distributive laws, we get
(a, b) [(c, d) + (e, f)] = (a, b) . (c + e, d + f)
= (ac + ae – bd – bf, ad + af + bc + be)
Also (a, b) . (c, d) + (a, b) (e, f) = (ac – bd, ad + bc) + (ae – bf, af + be)
= (ac + ae – bd – bf, ad + af + bc + be)
The commutative law of multiplication follows easily as shown below:
(c, d) (a, b) = (ca – db, cb + da)
= (ac – bd, ad + bc)
= (a, b) . (c, d)
𝑎 𝑏
Example 6. Let R = ( ) = a, b, c, d  Z i.e. the set of all 2 × 2 matrices with integers as
𝑐 𝑑
elements. We define the usual addition and multiplication in R as follows:
𝑎 𝑏1 𝑎 𝑏2 𝑎 + 𝑎2 𝑏1 + 𝑏2
( 1 )+( 2 ) =( 1 )
𝑐1 𝑑1 𝑐2 𝑑2 𝑐1 + 𝑐2 𝑑1 + 𝑑2
𝑎 𝑏1 𝑎 𝑏2 𝑎 𝑎 + 𝑏1 𝑐2 𝑎1 𝑏2 + 𝑏1 𝑑2
( 1 )×( 2 ) =( 1 2 )
𝑐1 𝑑1 𝑐2 𝑑2 𝑐1 𝑎2 + 𝑑1 𝑐2 𝑐1 𝑏2 + 𝑑1 𝑑2
Then R is a ring with respect to addition and multiplication as defined above. It is called the ring of
matrices of order 2 over integers. It is easy to see that R is a non-commutative ring with identity and with
zero divisors. The zero element and the identity of R are given by
0 0 1 0
O=( ) and e = ( ) respectively
0 0 0 1
1 1 1 0 0 0
Note that ( )( ) =( )
0 0 –1 0 0 0
1 1 1 0
We find that ( ) and ( ) are divisors of zero in R
0 0 –1 0
Theorem 1 : If R is a ring then for all a, b  R
i) a.0 = 0.a = 0
ii) a(–b) = (–a)b = –ab
iii) (–a) (–b) = ab
If in addition R has an identity 1, then (–1) a = –a

169
 (–1) (–1) = 1
Proof : If R has additive identity 0, then a + 0 = a  a  R
Also, in particular
0+0 =0
 a(0 + 0) = a.0
By distributive law a.0 + a.0 = a.0
= a.0 + 0
By Left Concellation Law in the group (R, +) we get
a.0 = 0
Similarly 0.a = 0
Also b + (–b) = 0  b  R
 a[b + (–b)] = a.0 = 0 [as proved above]
i.e. ab + a(–b) = 0 [By Distributive Law]
 a(–b) = –ab
Similarly starting with
[a + (–a)] b = 0.b = 0 we can prove that (–a) b = –ab
To prove (iii) (–a) (–b) = ab
We have (–a) (–b) = (–a) c where c = –b
= –ac
= –[a (–b])
= –[–ab]
= ab
 –(–x) = x is true in the additive group (R, +)
Definition : Integral Domain
Def : A commutative ring with identity and without zero divisions is called an Integral Domain.
If {R, +,.} is an integral Domain
i) R is commutative
i.e.  a, b  R, ab = ba
ii) ab = 0  a = 0 or b = 0
i.e. a  0, b  0  ab  0

170
 i.e. R does not have zero divisors. For example, the set Z of integers is an integral Domain, as we
know that commutative law is true for multiplication of integers.
Also ab = 0  a = 0 or b = 0 where a, b  Z
It is not necessary that every Ring be an integral Domain for example, you may verify that the set
Z6 = {0, 1, 2, 3, 4, 5} forms a Ring with respect to the binary operations of Addition modulo 6 and
multiplication modulo 6.
2  Z6 , 3  Z6 2  0
Also ]
30
But 2 ʘ6 3 = 0
 Z6 is not an integral Domain.
Theorem 2. A finite integral Domain is a field.
Before proving the theorem, we shall define a field. A field is an Integral Domain in which every
non zero element has the multiplicative inverse. It is easy to see that the set of all rational numbers
𝑝 𝑝 𝑞
𝑄 = [𝑞 /𝑝, 𝑞 𝑍, 𝑞 0] is a field. Every non-zero rational number 𝑞
has the multiplicative inverse 𝑝
as
𝑝 𝑞
𝑞
×𝑝=1

Proof : Let R = {1, 2, 3, .......n

be a finite integral Domain
 ai  aj, i  j .....(i)
Let a  o be a non zero element of R [i.e. a = k for some k, 1  k  n]
Consider the set aR.
aR = {a1, a2, a3, ..............an}
It is easy to see that the elements of aR are all distinct.
For if ai = aj  i = j
ij ij
which contradicts (1)
By the closure property in R it follows that
R = aR [ ai  R for all i  R]
If 1 denotes the identity of R then for some k  R
1 = adh  dh is the inverse of a
 R is a field.

171
 Theorem 3. If p is a prime number then Jp the ring of integers modulo p is a field.
Proof : Let Jp = {0, 1, 2, 3, ............, p – 1}. It is easy to see that Jp is a ring under the operations of
addition and multiplication modulo p. As we have proved in Theorem 2 that every finite integral domain
is a field, it is just sufficient to show that Jp is an integral domain.
Let x, y  Jp be such that
xy = 0 then x = 0 or y = 0
or p divides xy. If p divides xy, then p divides x or p divides y since p is prime.
 x = 0 or y = 0 in Jp
 xy = 0 in Jp  x = 0 or y = 0 in Jp
It follows that Jp is a finite integral domain and therefore a field.
SUBRINGS : Let R be a ring. Any nonempty subset S of R is said to be subring if it forms a ring
under the binary compositions of R. Every ring R has two trivial subrings vis {0} and the ring R itself.
Example 1. The set of all multiples of three
viz S {3, 6, 9, ........ 0, –3, –6, –9} is a subring of the ring of integers.
You may easily verify that S satisfies all the Axioms required for a ring.
𝑜 𝑥
Example 2. The set of all 2 × 2 matrices of the form (𝑜 𝑦) where x, y  R, the set of all real
numbers, is a subring of the ring of all 2 × 2 matrices
We shall prove a theorem which gives sufficient conditions for a non empty subset S of a ring R to
be a subring of R.
Theorem 4. A non empty subset S of a ring is a subring of R, if and only if
(i) x – y  S
(ii) xy  S
 x, y  S
Proof : Firstly we shall show that the conditions are necessary. If S be a subring of R we have x, y
 S, xy  S [By Def. of a ring]
Also {S, +} is an abelian group
 y  S  –y  S  x – y  S For all x, y  S.
The conditions are also sufficient if x – y  S,  x, y  S it follows that S is an Abelian group with
respect to addition. Also (S, +) is subgroup of (R, +).
Let x, y, z  S Then xy  S by (ii) obviously x(yz) = (xy)z
x(y + z) = xy + xz
and (y + z)x = yx + zx, since S is a subset R and associative and distributive laws are true in R
 S is a subring of R.

172
 Theorem 5. The intersection of two subrings of a ring is a also a subring i.e. If S1, S2 are two
subrings of a griven ring R, then S1  S2 is also a subring in R.
Proof : Let x, y S1  S2
 x  S1, x  S2, y  S1, y  S2
 x – y  S1, x – y  S2 and xy  S1, xy  S2
as S1, S2 are subrings of R {By Theorem 4}
 x – y  S1  S2
and xy  S1  S2 x, y, S1  S2
 S1  S2 is a subring of R.
Note : The union of two sub-rings need not necessarily be a subring as is evident from the
following example.
Example 3. Consider the set of integers Z. Obviously Z is a ring. It is easy to see that
R2 = [2p / p Z]
R3 = [3p / p  Z]
are subrings of Z. Also R2  R3 = [2, 3, 4, 6, ........0, –2, –3, –4, –6, ........]
2, 3  R2  R3
But 2 + 3 = 5  R2  R3
 R2  R3 is not even a group with respect to addition in Z
 R2  R3 is not subring of Z although R2, R3 are subrings of Z.

173
 LESSON 4

VECTOR SPACES

1.1 INTRODUCTION
You are already familier with several algebric structures such as groups, rings, integral domains and
fields. In this lesson we shall tell you about another equally important algebric structures, namely, a
vector space.
Let V be a non-emply set and let F be a field. Let us agree to call elements of V vectors and
elements of F scalars.
A maping from V × V to V will be called addition in V and a mapping from F × V to V will be
called multiplication by a scalar multiplication, V is said to a vector space over F if addition and scalar
multiplication satisfy certain properties. Of course, these conditions are to be chosen in such a manner
that the resulting algebric structure is rich enough to be useful. Before presenting the definition of a vector
space, let us note that addition in V is denoted by the symbol ‘+’, and scalar multiplication is denoted by
juxtaposition, i.e., if x  V, y  V, and   F, the x + y denoted the sum of x and y, and x denotes the
scalar multiople of x by .
Definition 1. A non-empty set V is said to be a vector space over a field F with respect to addition
and scalar multiplication if the following properties hold.
V 1 Addition in V is associative, i.e.,
x + (y + z) = (x + y) + z, for all x, y, z,  V
V2 There exists of natural element for addition in V, i.e., there exists an element 0 Î V such that
x + 0 = 0 + x = x, for all x  V
V3 Every element of V possesses a negative (or addition inverse), i.e., for each x  V, there
exists an element y  V such that
x + y = y + x = 0.
V4 Addition in V is commulative, i.e., for all elements x, y  V,
x+y= y+x
V5 Associtiavity of scalar multiplication, i.e.,
i.e., (x) = ( ) x, for all , ,  F and x  V
V6 Property of 1. For all xV,
1x = x, where 1 is the multiplicative identity of F.
V7 Distributivity properties for all ,   F and x, y  V.
( + )x = x + x
(x + y) = x + y
Remarks 1. The first of the two distributivity properties stated in V 7 above is generally called
distributivity of scalar multiplication over addition in F, and the second of the two distributibity
properties is called distributivity of scalar multiplication over addition in V.

174
 2. We generally refer to properties V 1 – V 7 above by saying that (V, +) is a vector space over F. If
the underlying field F is fixed, we simply say that (V, +,) is a vector space, and do not make an explicit
reference to F.
In case, the two vector space compositions are known, we denote a vectors space over a field’ F by
the symbol V(F). If there is no chance of confusion about the underlying field, then we simply talk of ‘the
vector space V’.
3. You might have observed that the axions V 1 to V 4 simply assert the V is an abelian group for
the composition ‘+’. In view of the we can re-state the definition of a vector space as follows:

2. DEFINNITION AND EXAMPLS OF A VECTOR SPACE

Defintion 2. A triple (V, +,) is said to be a vector space over a field F if (V, +) is an abelian group,
and the following properties are satisfied :
(x) = ()x, ,   F and  x  V
1x = x, x,   V, where 1 is the multiplicative identity of F
( + )x = x + x, ,   F, and x, y  V
 (x + y) = x + y,   F and  x, y  V
We shall now consider some examples of vector spaces.
Example 1. Let R be the set of number (R, +) is vector space over R. The addition is addition in R
and scalar multiplication is simply multiplication of real numbers.
It is easy to verify that all the vector space axioms are verified. In fact, V1-4 are satisfied because R
is an abelian group with respect to addition, V5 is nothing but the associative property of multipication,
V6 is the property of the multiplicative identity in (R, +,) and the properties listed in V7 are nothing but
the distributivity of multiplication over addition.
Example 2. (C, +, ) is a vector space over C
Example 3. (Q, +, ) is a vector space over Q.
Example 4. Let F be any field. F is a vector space over itself for the usual compositions of addition
and multipication (to be called scalar multiplication) in F.
Example 5. C is a vector space over R, and R is a vector space over Q.
Example 6. R is not a vector space over C. Observe that if   C and x  R, the x is not in R.
Therefore the multiplication composition in R fails to give rise to the scalar multiplication composition.
The examples considered above are in a way re-labelling of the field properties C, R or Q. We shall
now consider some examples of a different type.
Example 7. Let V be the set of all vectors in a plane. You know that addition of two vectors is a
vector, and that V is a group with respect to sum of vectors. Let us take addition of vectors as the first
compostion for the purpose of our example. Also, we know that if d be any vector and k be any real
numbers, then k d is a vector. Let us take R as the underlying field and multiplication of vector by a
scalar as the second vector space composition. It is easy to see that V is vector space over R for these two
compositons.
Example 8. Let R3 be the set
{( x1, x2 , x3 ) : x1, x2, x3  R}
and the addition and scalar multiplication R3 be defined as follows:

175
 If x  R3 and y  R3 let
x + y = ( x1 + y1 , x2 + y2 , x3 + y3 ).
Also if x  R3 and c  R, let
cx = (cx1 , cx2 , cx3 ).
It can be seen that R3 is a vector space over R for the two compositions—addition and scalar
multiplication, as defined above.
We may note before passing on to the next example that the vector space being considered here is
nothing but the space of the vectors (in space) with addition and scalar multiplication as the composition.
This example is of special interest because it was in fact motivation for the present terminology of vector
spaces.
The next three examples are a little abstract in nature but are quite important.
Example 9. Let Rn be the set of ordered n-tuples of real numbers, so that a typical element of Rn is
We shall denote this element by x (printed as a bold-face better), and write
x = ( x1 , x2 , ..., xn )
Let us take R to be the underlying field and define addition and scalar multiplication in Rn by setting
x + y = ( x1 + y1 , x2 + y2 , ..., xn + yn ),
where x = ( x1, x2 , ... xn ), y = ( y1, y2 , ... yn )
and x = ( x1, x2 , ... xn ) ,   R
Let us first of all see that addition scalar multiplication as defined above are meaningful in the sense
that they define the two compositions that we need for making Rn vector space.
Since x1 , x2 , x3 , xn and y1 , y2 , ... yn are all real numbers, therefore x1 + y1 , x2 y2 , ..., xn + yn are all
real numbers and therefore ( x1 + y1 , x2 + y2 , ..., xn + yn ) is an ordered n-type of real numbers and
consequently it is in Rn. Again, since  is a real number and ( x1 , x2 , ..., xn ) are also real numbers,
therefore  x1 , x2 , ...,  xn are also real numbers, and consequently (x1 , x2 , ..., xn ) is an n-type of real
numbers and so is Rn.
Having defined addition and scalar multiplication in Rn, let us see in some detail that all the
properties needed for Rn to be a vector space are actually satisfied.
1. Let x = ( x1, x2 , ..., xn ), y = ( y1, y2 , ..., yn ), z = ( z1, z 2 , ..., z n ) be any three elements of Rn.
Then (x + y) + z = ( x1 + y1 , x2 + y2 , ..., xn + yn ) + ( z1 , z2 , , zn )
= [( x1 + y1 ) + z1 , ( x2 + y2 ) + z2 , ..., ( xn + yn ) + zn ]
= [ x1 + ( y1 + z1 ), x2 + ( y2 + z2 ), , xn + ( yn + zn )]
= ( x1 , x2 , x3 , , xn ) + ( y1 + z1 , y2 + z2 , ... yn + zn )
= x + (y + z).
2. Let o = (0, 0, ..., 0). so that o  Rn and x + 0 = o + x = x for all x  Rn
3. Let x = ( x1 , x2 , ..., xn ) be any element of Rn. If y = (–x1, –xn, ..., –xn) then y  Rn and y is an
element of Rn such that
x+y= y+x=0
4. If x = ( x1 , x2 , , xn ), y = ( x1 , y2 , ..., yn ) are any two elements of Rn then

176
 x + y = ( x1 , x2 , , xn ), + ( y1 , y2 , ..., yn )
= ( x1 + y1 , x2 + y2 , , xn + yn )
= ( y1 + x1 , y2 + x2 , , yn + xn )
= ( y1 , y2 , ... yn ) + ( x1 , x2 , ..., xn )
= y + x.
5. If x = ( x1 , x2 , ... xn ), be any element of Rn and p, q be any real numbers, then
(pq)x = ( pq ) ( x1 , x2 , ... xn )
= [( pq) x1 , ( pq) x2 , ..., ( pq) xn ]
= [ p(qx1 ), p(qx2 ), ... p(qxn )]
= p (qx1 , qx2 , ... qxn )
= p(qx)
6. If x = (x1, ..., xn) be any element of Rn and p, q be any real numbers, then
(p + q) x = [( p + q) x1 , ( p + q) x2 , ..., ( p + q) xn ]
= ( px1 , px2 , ..., pxn ) + (qx1 , qx2 , ..., pxn )
= px + qx
7. If x = ( x1 , x2 , ..., xn ) and y = ( y1 , y2 , ..., yn ) be any two elements of Rn, and p be any real number,
then
p(x + y) = p(x1 + y1, x2 + y2, ..., xn + yn)
= [p(x1 + y1), p(x2 + x2), ..., p(xn + yn)]
= (px1, px2, ..., pxn) + (py1, py2, ..., pyn)
= p(x1, x2, ..., xn) + p(y1, y2, ..., yn)
= px + py.
8. If x = ( x1 , x2 , ..., xn ) be any element of Rn, then
1x = 1( x1 , x2 , ..., xn )
= (1  x1 ,1  x2 , ,1  xn )
= ( x1 , x2 , ... xn )
= x.
From 1-8 above we find that Rn is a vector space over R with co-ordinatewise addition and
co-ordinatewise scalar multiplication as the two vector space compositions.
The use of the word co-otherwise, is due to the fact that if x = ( x1 , x2 , ..., xn ) be any element of Rn
then x1 , x2 , ..., xn are called the co-ordinates of x, and while defining x + y, we add the corresponding
co-ordinates of x and y, and while defining cx we multiply the co-ordinates of x by c.
We may note that the space in example 9 is only a spacing case of the example 9 for n = 3.
Example 10. The set Cn of all ordered n-tuples of complex number is a vector space over C for
co-ordinatewise addition and co-ordinatewise scalar multiplication as the two vector space compositions.

177
 Example 11. Let F be any field. The set Fn of all ordered n-tuples of elements of F is a vector space
over F with co-ordinatewise addition and co-ordinatewise scalar multiplication as the two vector space
compositions.
Example 12. Let Mmn is a vector space over C with respect to matrices over C. Mmn is a vector space
over C with respect to matrix addition and multiplication of a matrix by a scalar, for
1. The sum of two m × n matices with complex entries is an m × n matrix with complex entries.
2. Addition of matrices is associative.
3. The m × n zero-matrix is a natural element for addtion.
4. It A be an m × n matrix with complex entries, then –A is also an m × n with complex entries
such that (–A) + A + (–A) = 0
5. Addition of matrices is commutative.
6. If A  Mmn and c be any complex number, then –A is also m × n matrix with complex entries
and so cA  Mmn
7. If p, q be any complex numbers and A, B be any two m × n matrices with complex entries, then
p(q A) = (pq) A,
(p + q) A = pA + qA,
p(A + B) = pA + pB,
1 A = A, for all A  Mmn
 a b  , where a, b are any complex numbers, is
Example 13. The set S of all matrices of the form  
 −b a 
a vector space over C with respect to matrix addition and multiplication by a scalar for the following
reasons:
1. If A, B,  S, the A + B  S.
 a b  c
A= 
d
For, if  c 
,
 −b a   −d
 a+c b+d  p q
A+B=  = , where p = a + c, q = b + d .
a + c   −q p 
then
 −(b + d )
2. Matrix addition is associative.
o o
3. The matrix O =    S , and A + O = O + A = A, for all A  S.
o o
 p q  − p −q 
4. If A =    S , then the matrix B =    S , and is such that A + B = B + A = 0.
 −q p  q −p
5. Addition f matrices is commulative.
 a b
6. If c be any complex number, and A =    S, then cA is the matix given by
 −b a 
 ca cb 
cA =  .
 −cb ca 
It is obvious that cA  S.

178
  a b
7. If p, q be any complex numbers, and A, B be any two matrices in S, say A =  ,
 −b a 
 c d  , then
B= 
 −d c 

(pq)A = pq 
a b   ( pq)a ( pq)b   p(qA) p(qb) 
(i)
−  = ( pq)(−b) ( pq)a  =  p(−qb) p(qa) 
= p(qA).
 b a    
(ii) (p + q) A = pA + qA
(iii) p(A + B) = pA + pB
(iv) 1 A = A.

Example 14. The set of all matrices of the form 

 x y  where x, y, z,  C, is a vector space over
,
 z o
C with respect to matrix addition and multiplication of a matrix by a scalar.
The verification of the vector space axioms is straight forward.
Example 15. The set s of all hermitian matrices of order n is a vector space over R with respect to
matrix addition and multiplication of matrix by a scalar.
To verify that (S, +, ) is a vector space over C, we proceed as follows:
1. Let A, B be two hermitian matrices of order n. Then A + B is a matrix of order n. It is hermitian
because (A + B) = A + B = A + B, since A = A, B = B.
2. Addition of matrices is associative.
3. The n-rowed, zero matrix O is a matrix is a hermitian matrix such that A + 0 = 0 + A = A
4. It A  S, so that At = A, then (–A)t = –At = –A, so that –4A  S, and A + (–A) = (–A) + A = 0.
5. Matrix addition is commutative
6. If C  R, and A  S, then (CA)t = CAt = CA, so that A  S.
7. If p, q,  R and and A B  S, then
(i) (p, q)A = p(qA)
(ii) (p + q)A = pA + qA
(iii) p(A + B) = pA + pB
(iv) 1A= A
In view of the above properties it follows that (S, +, ) is a vector space over R.
Example 16. The set S of all real symmetric matrices of order n is a vector space over R with
respect to matrix addition and multiplication of a matrix by a scalar.
In order to convince ourselves that (S, +, ) is a vector space over R, let us note the following:
1. A, B be n- rowed real symmetric matrices, then At = A, Bt = A. It follows that
(A + B)t = At + Bt = A + B
2. Addition of matrices is associative, so that for all A, B, C  S,
A + (B + C) = (A + B) + C.
3. The n-rowed zero matrices o is a real symmetric matrix, and therefore it is in S.
Also, A+0= 0+A=A
for all A in S.
4. If A  S so that At = A, then (–A)t = –At = –A, so that—17 A  S. Also, A + (–A) = (–A) + A = 0.

179
 5. Matrix addition is commulatative.
6. If C  R and A  S, then (cA)t = cAt = cA, so that A  S.
7. If p, q  R, and A, B  S, then
(i) (p, q) A = p(qA)
(ii) (p + q)A = pA + qA
(iii) P(A + B) = pA + pB
(iv) 1A= A

3. SOME DIRECT CONSEQUENTS OF VECTOR SPACE AXIOMS

We shall now state and prove some elementary consequences of the vector space axioms. These will
help us in dealing with vector in a convenient way in many situations.
Theorem. Let V be a vector over a field F. Then for all a  F and x  V,
(i) 0 = 0,
(ii) 0x = 0,
(iii) (–) x = –ax
(iv) (–) (–x) = x
(v) x = 0 if either  = 0 or x = 0.
Proof.
(i) 0 = (0 + 0), by the property of 0  V
= 0 + 0, by distributivity of scalar multiplication over addition in V
Since 0 + 0 = 0 = 0 + 0, by the property of 0 in V by cancellation law in (V, +), it
follows that 0 = 0
(ii) 0 x = (0 + 0) x, by the property of 0 in F
= 0x + 0x, by distributivity of scalar multiplication over addition in F.
Since 0x + 0x = 0x + 0, by, the property of 0 in V, therefore by cancellation law in (V, +),
it follows thats
0x = 0
(iii) 0 = a + (–)  0x = ( + (–))x
 0 = x + (–) x, since ox = 0.
Now x, (–)x are two elements of V such that ax + (–)x = 0, therefore (–) x is the negative of.
x, i.e, (–) x = –(x).
(iv) Now –(0) = – (x + (–x))
 0 = –x + ((–) (–x)), since –(0) = 0, by (i) above
 (–) (–x) is the negative of –x in V
 (–) (–x) = –(–x) –x,
because –x  V and therefore negative of –x in V is simply x.
(v) Let us first suppose that ax = 0. If  = 0, we are done. If   0, then –1  F, because   F
and F is a field
Therefore 0 = –1 0

180
 = –1 (x), because x = 0 by hypothesis
= (–1 )x, by associativity of scalar multiplication
= 1x
= x.
Thus x = 0  either  = 0 or x = 0. Conversely, let us assume taht either  = 0 or x = 0
In case  = 0, by (ii) above x = 0x = 0.
In case x = 0, by (i) above x = 0 = 0.
Thus in both cases we find that x = 0.
It is now obvious that x = 0 if either x = 0 or  = 0.
Exercise
1. Show that the set
C2 = {(x1, x2) : x1  C, x2  C}
is a vector space over C with respect to co-ordinateswise addition and scalar multiplication.
2. Show that the set of all 2 × 2 matrices over C is a vector space over C with respect to matrix
addition and multiplication of a matrix by a scalar.
3. Let V = {(a1, a2, a3, a4) : a1, a2, a3, a4 are integers}
Is V a vector space over R with respect to co-ordinatewise addition and scalar multiplication?
Justify your answer.
4. Let V = {(x1, x2, x3) : (x1, x2, x3, are complex numbers, and x1, x2, = 0}
Is V a vector space over C with respect to co-ordinateswise addition and scalar multiplication?
Justify your answer.
 x o  where y  C is a vector space over C with
5. Show that the set of all matrices of the form  ,
o y
respect to matrix addition and multiplication of a matrix by a scalar.

6. Show that the set of all matrices of the form 

 a b  where a, b  C is vector space over C
,
 −b a 
with respect to matrix addition an multiplication of matrix by a scalar.

181