SCHOOL OF SCIENCE AND HUMANITIES

DEPARTMENT OF MATHEMATICS

UNIT – I – ENGINEERING MATHEMATICS-I - SMTA1101

1
Sathyabama University

Application of Matrices
Applications of matrices are found in many fields. The work of the
Mathematician Olga Taussky Todd (1906–1995) elevated the matrix from a
mere tool to an important mathematical theory. She used matrices to analyze
vibrations on airplanes during World War II.
Matrices are a very useful way of studying graphs, since they turn the
picture into numbers. Computer graphics use matrices both to represent
objects and to calculate transformations of objects to accomplish tasks such as
projecting a three-dimensional object onto a two-dimensional screen. In
probability theory and statistics, stochastic matrices are used in the page rank
algorithm that ranks the pages in a google search. In Cryptography, early
encryption techniques such as the Hill cipher, used matrices.
Traditional mesh analysis in electronics leads to a system of linear
equations that can be described with a matrix. The behaviour of many
electronic components can be described using matrices. Matrices over a
polynomial ring are important in the study of control theory. In civil
engineering field, in problems such as trusses, beams, supports, material
mechanics, fluid dynamics and mechanical stresses, the use of matrix is
common.
In every branch of physics, including classical mechanics, optics,
electromagnetism, quantum mechanics and quantum electrodynamics,
matrices are used to study physical phenomena. Planetary theory and atomic
theory uses infinite matrices. Chemistry makes use of matrices in various
ways, particularly in the use of quantum theory to discuss molecular bonding
and spectroscopy.
Matrix decomposition methods simplify computations, both theoretically
and practically.
Matrix: A set of mn numbers (real or imaginary) arranged in the form of a
rectangular array of m rows and n columns is called a matrix of order m × n
(to be read as ‘m by n’). A m×n matrix is usually written as
ª a11 a12 " a1 j " a1n º
«a a22 " a2 j " a2 n »»
« 21
« # # # # »
A=« »
« ai1 ai 2 " aij " ain »
« # »
« »
«¬ am1 am 2 " amj " amn »¼

In compact form it is written as A = ( aij ) or A = ( aij ) where

m× n

i = 1, 2, ... m and j = 1, 2, ... n

2
 Sathyabama University

Here a11, a12, … are called the elements of the matrix. In the element aij
the first suffix i indicates the row and the second suffix j indicates the column
in which it is present.

Types of Matrices
Row Matrix: A matrix which has only one row and any number of columns
is called a row matrix or row vector.
Example: A = [3 4 5] is a row matrix of order 1 × 3

Column matrix: A matrix which has only one column and any number of
rows is called a column matrix or column vector.

ª1 º
«5»
Example: A = « » is a column matrix of order 4 × 1
«9 »
« »
¬ 4¼

Square Matrix: A matrix in which the number of rows is equal to the

number of columns is called a square matrix. Otherwise it is said to be a
rectangular matrix.
ª2 2 1º
Example: ««1 3 1 »» is a square matrix of order 3.
«¬1 2 2 »¼

The elements a11, a22, a33 are called the diagonal elements. The diagonal along
which these elements lie is called the principal diagonal or leading diagonal
of the matrix.
Trace of a Matrix: The sum of the leading diagonal elements of a square
matrix is called its trace or spur.

ª a11 a12 a13 º

Example: If A = «« a21 a22 a23 »» then Trace of A = a11 + a22 + a33
«¬ a31 a32 a33 »¼

Diagonal Matrix: A square matrix in which all the elements except those in
the leading diagonal are zero is called a diagonal matrix.

ª2 0 0º
Example: A = «« 0 3 0 »»
«¬ 0 0 4 »¼

3
Sathyabama University

Scalar Matrix: A diagonal matrix in which all diagonal elements are equal is
called a scalar matrix.

ª −3 0 0 º
Example: A = «« 0 −3 0 »»
«¬ 0 0 −3»¼

Identity or Unit Matrix: A scalar matrix in which all the leading diagonal
elements are unity (i.e., 1) is called a unit matrix or identity matrix.

ª1 0 0 º
Example: I = «« 0 1 0 »» is an identity matrix of order 3.
«¬ 0 0 1 »¼

Null Matrix: A matrix in which all elements are zero is called a Null or zero
matrix.
ª0 0 0 º
Example: A = « » is a null matrix of order 2 × 3.
¬0 0 0 ¼
Upper triangular matrix: A square matrix in which all the elements below
the principal diagonal are zero is called an upper triangular matrix.
ª2 3 4º
Example: «« 0 5 6 »»
«¬ 0 0 7 »¼

Lower triangular matrix: A square matrix in which all the elements above
the principal diagonal elements are zero is called a lower triangular matrix.
ª1 0 0º
Example: «« 2 3 0 »»
«¬ 4 5 6 »¼

Triangular matrix: A square matrix in which all the elements either below
or above the principal diagonal are zero is called a triangular matrix. Thus a
triangular matrix is either upper triangular or lower triangular.
Nonsingular and Singular Matrix: A matrix A is said to be non singular if
| A | ≠ 0. Otherwise it is said to be a singular matrix.

ª1 2 1 º
Example: Let A = «« 3 2 1 »»
«¬ 2 4 2 »¼

4
 Sathyabama University

Then | A | = 1( 4 − 4 ) − 2 ( 6 − 2 ) + 1(12 − 4 ) = 0
∴ A is a singular matrix.

Addition of Matrices:
Let A & B be two matrices of the same order m × n. Then their sum denoted
by A + B is defined to be the matrix of order m × n, obtained by adding the
corresponding elements of A and B.

ªa a12 a13 º ªb b b13 º

Example: Let A = « 11 » and B = « 11 12
¬ a21 a22 a23 ¼ ¬b21 b22 b23 »¼

ªa + b a12 + b12 a13 + b13 º

Then A + B = « 11 11
¬ a21 + b21 a22 + b22 a23 + b23 »¼

Note: If two matrices are not of the same size, we cannot find their sum.

Multiplication of a matrix by a scalar:

Let A be any matrix and k be any number called scalar. The matrix obtained
by multiplying every element of the matrix A by k is called the scalar multiple
of A by k and is denoted by kA.
ª 1 2 −3º
Example: If k = 2 and A = « »
¬4 5 6 ¼
ª 2 4 −6 º
Then 2 A = « »
¬ 5 10 12 ¼

Multiplication of Matrices:
Let A be a m × n matrix and B be an n × p matrix such that the number of
columns in A is equal to the number of rows in B. The m × p matrix whose (i,
j)th element is obtained by multiplying corresponding elements of the ith row
of A and jth column of B and then adding the products, is called the product of
the matrix A and B and is denoted by AB.

ª a11 a12 " a1n º ª b11 b12 " b1 p º ª c11 c12 " c1 p º
«a « » « " c2 p »»
« 21 a22 " a2 n »» «b21 b22 " b2 p » « c21 c22
i.e., × =
« # # » «# # » « # # »
« » « » « »
¬ am1 am 2 " amn ¼ ¬«bn1 bn 2 " bnp ¼» ¬« cm1 cm 2 " cmp ¼»
where cij = ai1b1 j + ai 2 b2 j + " + ain bnj

5
Sathyabama University

Example: The products AB and BA of matrices

ª −3 1 º
ª2 1 3º
A=« » and B = «« 2 5 »» is given by
¬ 4 −1 2 ¼ «¬ 0 2 ¼»
ª −3 1 º
ª2 1 3º «
AB = « » « 2 5 »»
¬ 4 −1 2 ¼ «
¬ 0 2 »¼
ª( 2 × −3) + (1× 2 ) + ( 3 × 0 ) ( 2 ×1) + (1× 5 ) + ( 3 × 2 ) º
=« »
«¬( 4 × −3) + ( −1× 2 ) + ( 2 × 0 ) ( 4 ×1) + ( −1× 5 ) + ( 2 × 2 )»¼
ª −4 13º
=« »
¬ −14 3 ¼

ª −3 1 º
ª2 1 3º
Consider BA = «« 2 5 »» × «
¬ 4 −1 2 »¼
«¬ 0 2 »¼
ª( −3 × 2 ) + (1× 4 ) ( −3 ×1) + (1× −1) ( −3 × 3) + (1× 2 )º
« »
= «( 2 × 2 ) + ( 5 × 4 ) ( 2 × 1) + ( 5 × −1) ( 2 × 3) + ( 5 × 2 ) »
«
¬( 0 × 2 ) + ( 2 × 4 ) ( 0 ×1) + ( 2 × −1) ( 0 × 3) + ( 2 × 2 ) »¼
ª −2 −4 −7 º
BA = «« 24 −3 16 »»
«¬ 8 −2 4 »¼

We note that AB is a 2 × 2 matrix and BA is a 3 × 3 matrix.

Note: AB  BA

Transpose of a matrix

Let A = (aij) be an m × n matrix. The transpose of A denoted by AT or A ' is

obtained from A by changing its rows into columns and its columns into rows.

AT = ( a ji ) where i =1, 2! m and j = 1, 2! n

ª1 2 º
ª1 2 3º
Example: If A = « » then A = «« 2 3 »»
T

¬2 3 4¼ «¬ 3 4 »¼

6
 Sathyabama University

Sub-matrix: A matrix obtained from a given matrix A by deleting some of its

rows or columns or both is called a sub matrix.
ª1 3 4º
ª3 4 º
Example: B = « » is a submatrix of A = «« 2 5 6 »»
¬5 6 ¼ «¬ 7 8 9 »¼

Minor and cofactor

Let A = ( aij ) be a square matrix of order n. The minor Mij of aij in A is the
determinant of the square sub matrix of order n-1 obtained by leaving ith row
and jth column of A.
i+ j
The cofactor cij of aij in A = ( −1) M ij where Mij is the minor of aij in A

ª1 2 2 º
Example: A = «« 2 1 2 »»
«¬ 2 2 1 »¼
1 2
M 11 = Minor of a11 = = −3
2 1
2 2
M 12 = Minor of a12 = = −2
2 1
2 1
M 13 = Minor of a13 = =2
2 2
2 2
M 21 = Minor of a21 = = −2
2 1
1 2
M 22 = Minor of a22 = = −3
2 1
1 2
M 23 = Minor of a23 = = −2
2 2
2 2
M 31 = Minor of a31 = =2
1 2
1 2
M 32 = Minor of a32 = = −2
2 2
1 2
M 33 = Minor of a33 = = −3
2 1

7
Sathyabama University

1+1
c11 = cofactor of a11 = ( −1) M 11 = M 11 = −3
1+ 2
c12 = cofactor of a12 = ( −1) M 12 = − M 12 = 2
1+ 3
c13 = cofactor of a13 = ( −1) M 13 = M 13 = 2
2 +1
c21 = cofactor of a21 = ( −1) M 21 = − M 21 = 2
2+ 2
c22 = cofactor of a22 = ( −1) M 22 = M 22 = −3
2+3
c23 = cofactor of a23 = ( −1) M 23 = − M 23 = 2
3 +1
c31 = cofactor of a31 = ( −1) M 31 = M 31 = 2
3+ 2
c32 = cofactor of a32 = ( −1) M 32 = − M 32 = 2
3+ 3
c33 = cofactor of a33 = ( −1) M 33 = M 33 = −3

Adjoint of a square matrix:

Let A = ( aij ) be a square matrix and let cij be the cofactor of aij in A. The
transpose of the matrix of cofactors of elements of A is called the adjoint of A
and is denoted by adj A.

Inverse of a square matrix

Let A be a non singular square matrix. Then inverse of matrix A denoted by A-

1 1
is defined as A−1 = adj A
A

ª1 2 2 º
Example: Find the inverse of the matrix A = «« 2 1 2 »»
«¬ 2 2 1 »¼

Solution :
ª1 2 2º
A = «« 2 1 2 »» = 1(1 − 4 ) − 2 ( 2 − 4 ) + 2 ( 4 − 2 )
«¬ 2 2 1 »¼
=5≠0

A is a non singular matrix and therefore A-1 exists.

1
A −1 = adj A.
A

8
 Sathyabama University

T
ª c11 c12 c13 º
Where adj A = «« c21 c22 c23 »»
«¬ c31 c32 c33 »¼
T
ª −3 2 2 º
= «« 2 −3 2 »» ( refer previous example )
«¬ 2 2 −3»¼

ª −3 2 2 º
i.e., adj A = «« 2 −3 2 »»
«¬ 2 2 −3»¼
1
Hence A−1 = adj A
A
ª −3 2 2 º
= « 2 −3 2 »»
1«
5
«¬ 2 2 −3»¼

Rank of a Matrix:

Let A be a matrix of order m × n . Then rank of matrix A is said to be r if

(i) There is atleast one minor of A of order r which does not vanish and
(ii) Every minor of A of order greater than r vanishes.
Equal Matrices: Two matrices A and B are said to be equal if and only if
they have the same order and their corresponding elements are equal.
Similar Matrices: Let A and B be square matrices of order n. Then matrix B
is similar to matrix A, if there exists a non singular matrix M such that
B = M −1 AM
Real Matrix: A matrix is said to be real, if all its elements are real numbers.
ª 3 2º
Example: A = « »
¬ 4 1¼

Symmetric Matrix: A square matrix A = (a )

ij is called a symmetric matrix
if aij = a ji , for all i,j.
In other words a square matrix A is symmetric if and only if AT = A

9
Sathyabama University

ª 9 −1 4 º
Example: A = «« −1 8 2 »» is a symmetric matrix because
«¬ 4 2 7 »¼

a12 = a21 = −1
a13 = a31 = 4
a23 = a32 = 2

Skew Symmetric Matrix: A matrix A = ( aij ) is called a skew symmetric

matrix if aij = − a ji for all i, j.
In other words if AT = − A , then A is skew symmetric.

ª 0 −2 3 º
Example: «« 2 0 5 »» is skew symmetric
«¬ −3 −5 0 »¼

Complex matrix: If atleast one element of matrix A is a complex number,

then the matrix is called a complex matrix.
Conjugate of matrix A: The matrix obtained by replacing the elements of a
complex matrix A by the corresponding conjugate complex numbers is called
the conjugate of the matrix A and is denoted by A .
ª1 + 3i 9 − 7i º ª1 − 3i 9 + 7i º
Example: If A = « » then A = «
¬4 3 ¼ ¬4 3 »¼

Note: The conjugate of the transpose of matrix A i.e., ( AT ) and the transpose
T
of conjugate of matrix A i.e., A ( ) are equal and is denoted by A*

T
Thus A* = ( AT ) = A ( )
Hermitian Matrix: A square matrix A is said to be Hermitian if A* = A.
Thus in a Hermitian matrix, the diagonal elements are all real, while every
other element is the complex conjugate of the element in the transposed
position.

ª 9 1− i 5i º
«
Example: A = «1 + i 8 2 + 3i »» is a Hermitian matrix.
¬« −5i 2 − 3i 9 »¼

10
 Sathyabama University

Skew Hermitian Matrix: A matrix A is said to be skew Hermitian if

A* = − A . Thus in a skew Hermiton matrix, the diagonal elements are purely
imaginary of the form iȕ or zero. While every other element is the negative of
the complex conjugate of the element in the transposed position.

ª 2i 3+i 4 º
Example: A = «« −3 + i −7i −1 − i »»
«¬ −4 1 − i 0 »¼

Unitary Matrix: A square complex matrix A is said to be unitary if

AA* = I = A* A .
Orthogonal Matrix: A square real matrix A is said to be orthogonal if
AAT = I = AT A .

Vector
A set of n numbers xi , x2 , … xn written in a particular order (or an ordered
set of n numbers) is called an n-dimensional vector or a vector of order n. The
numbers xi , x2 , … xn are called the components or elements of the vector. A
n-dimensional vector can be written as a row vector as X = [ x1 , x2 … xn ] or
ª x1 º
«x »
as a column vector X = « 2 »
«#»
« »
¬ xn ¼
Linear Combination of Vectors

A vector X is called a linear combination of the vectors X 1 , X 2 , … X r if

there exists scalars k1 , k2 , … kr such that X = k1 X 1 + k2 X 2 …+ kr X r

Linear Dependence of Vectors:

A set of vectors X 1 , X 2 , … X r are said to be linearly dependent if we can

find scalars k1 , k2 , … kr which are not all zero, such that
k1 X 1 + k2 X 2 …+ kr X r = 0 .

−k1 X 1 k2 X 2 k3 X 3 k X
i.e., X m = − − − .... − r r
km km km km

In other words X 1 , X 2 , … X r are linearly dependent vectors if atleast one of

the vectors is a linear combination of the remaining vectors.

11
Sathyabama University

Linear Independence of Vectors

A set of vectors X 1 , X 2 , … X r is said to be linearly independent if it is not

linearly dependent. i.e., if the relation k1 X 1 + k2 X 2 …+ kr X r = 0 is satisfied
only when k1 = k2 = …. = kr = 0.

Method of testing linear dependence or independence of vectors

Method 1: Use the definition to test for dependence or independence of

vectors.

Method 2: Write the given vectors as row vectors and form a matrix. Using
elementary row operations, reduce this matrix to echelon form, i.e., the one in
which all the elements in the rth column below the rth element are zero each. If
the number of non zero rows in the echelon from equals the number of
vectors given, then the vectors are linearly independent. Otherwise they are
linearly dependent

Method 3: When the given vectors written as row vectors and formed as a
matrix is a square matrix, then we find |A|. If |A| = 0, then the vectors are
linearly dependent. If A ≠ 0 , then the vectors are linearly independent.

Example 1:
Show that the vectors X 1 = (1, 0, 4,3) , X 2 = ( 2,1, −1,1) and X 3 = ( 3, 2, −6, −1)
are linearly dependent. Find also the relationship among them.

Solution

Method 1:
If the vectors X 1 , X 2 , X 3 are linearly dependent then we can find scalars
k1 , k2 , k3 (not all zero) such that k1 X 1 + k2 X 2 + k3 X 3 = 0
i.e., k1 (1, 0, 4,3) + k2 ( 2,1, −1,1) + k3 ( 3, 2, −6, −1) = 0

i.e., k1 + 2k2 + 3k3 = 0

0k1 + k2 + 2k3 = 0
4k1 − k2 − 6k3 = 0
3k1 + k2 − k3 = 0

12
 Sathyabama University

Using the rule of cross multiplication, for the first two distinct rows we get
k1 k2 k
= = 3
2 3 1 3 1 2
−
1 2 0 2 0 1

k1 k2 k3
i.e., = =
1 −2 1
Consider 1⋅ X 1 − 2 ⋅ X 2 + 1⋅ X 3 = (1, 0, 4,3) − 2 ( 2,1, −1,1) + 1( 3, 2, −6, −1)

i.e., X 1 − 2 X 2 + X 3 = ( 0, 0, 0, 0 )

∴ The vectors X1 , X 2 , X 3 are linearly dependent and their

relationship is X1 – 2 X 2 + X 3 = 0 .

Method 2:

Writing X 1 , X 2 , X 3 as row vectors, we get

ª 1 0 4 3 º R1
A = «« 2 1 −1 1 »» R2
«¬ 3 2 −6 −1»¼ R3

ª1 0 4 3 º R1
« »
~ «0 1 −9 −5» R2 − 2 R1 = R2' (say)
«¬0 0 0 0 »¼ R3' − 3R2' = R3' ( say )

ª1 0 4 3 º R1
«
~ «0 1 −9 −5»» R2 '
«¬0 0 0 0 »¼ R3' − 2 R2' = R3'' ( say )

In the echelon form of the matrix, the number of non zero vectors = 2 (less
than the number of given vectors)
∴ X1 , X 2 , X 3 are linearly dependent.
Now R3'' = 0
i.e., R3' − 2 R2' = 0

Substituting for R3' and R2' we get

13
Sathyabama University

( R3 − 3R1 ) − 2 [ R2 − 2 R1 ] = 0
i.e., R1 − 2 R2 + R3 = 0
i.e., X 1 − X 2 + X 3 = 0 is the relation between the vectors.

Example 2

Show that the vectors X 1 = (1, −1, 2,3) , X 2 = (1, 0, −1, 2 ) and X 3 = (1,1, −4, 0 )
are linearly independent.

Solution

Method 1:
Consider k1 X 1 + k2 X 2 + k3 X 3 = 0

i.e., k1 (1, −1, 2,3) + k2 (1, 0, −1, 2 ) + k3 (1,1, −4, 0 ) = 0

i.e., k1 + k2 + k3 = 0

− k1 + 0.k2 + k3 = 0
2k1 − k2 − 4k3 = 0
3k1 + 2k2 + 0.k3 = 0

Using the rule of cross multiplication for the first two distinct rows, we get
k1 −k2 k3
= =
1 1 1 1 1 1
0 1 −1 1 −1 0

k1 k2 k3
i.e., = =
1 −2 1
Consider 1 X 1 − 2 X 2 + 1 X 3

= (1, −1, 2,3) − 2 (1, 0, −1, 2 ) + (1,1, −4, 0 )

= ( 0, 0, 0,1)
≠ ( 0, 0, 0, 0 )

∴ The vectors are linearly independent.

14
 Sathyabama University

Method 2:

We write X 1 , X 2 , X 3 as row vectors, we get

ª1 −1 2 3 º
A = ««1 0 −1 2 »»
«¬1 1 −4 0 »¼
ª 1 −1 2 3 º R1
~ ««0 1 −3 −1»» R2 − R1 = R2'
«¬0 2 −6 −3»¼ R3 − R1 = R3'
ª1 −1 2 3 º
~ ««0 1 −3 −1»»
«¬0 0 0 −1»¼ R3' − 2 R2'

In the echelon form of the matrix, the number of non zero vectors = 3 = The
number of given vectors.
∴ X 1 , X 2 , X 3 are linearly independent.

Example 3
Test whether the following vectors are linearly dependent or independent
X 1 = (1,1,1) , X 2 = ( 3,1, 2 ) and X 3 = ( 2,1, 4 )

Solution
Writing the given vectors as row vectors, we get

ª1 1 1 º
A = «« 3 1 2 »»
«¬ 2 1 4 »¼

Since matrix A is a square matrix, we find |A| consider

1 1 1
Consider A = 3 1 2
2 1 4
= 1( 4 − 2 ) − 1(12 − 4 ) + 1( 3 − 2 )
= −5
| A |≠ 0 Ÿ the vectors are linearly independent.

15
Sathyabama University

CHARACTERISTIC EQUATION

Let A = (aij ) be a square matrix of order n, we can form the matrix

A − λ I = 0 , where I is the nth order unit matrix. The determinant of this
matrix when equated to zero is called the characteristic equation of the matrix
A.
i.e., A − λ I = 0

a11 − λ a12 ! a1n

a21 a22 − λ ! a2 n
Ÿ =0
# # ! #
an1 an 2 ! ann − λ

On expanding the determinant, the characteristic equation takes the form

Co λ n + C1λ n −1 + ... + Cn −1λ + Cn = 0

where pi’s are expressible in terms of the elements aij

Note:

§a a12 ·
1. Let A = ¨ 11 ¸ be a 2 × 2 matrix.
© a21 a22 ¹
Then the characteristic equation of A is A − λ I = 0
The general form of the characteristic equation is given by
λ 2 − C1λ + C2 = 0
where C1 = sum of the principal diagonal elements of A = a11 + a22
C2 = A
§ a11 a12 a13 ·
¨ ¸
2. Let A = ¨ a21 a22 a23 ¸ be a 3x3 matrix.
¨a ¸
© 31 a32 a33 ¹
Then the characteristic equation of A is A − λ I = 0
The general form of the characteristic equation is given by
λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = sum of the principal diagonal elements of A
= a11 + a22 + a33

C2 = sum of the minors of the principal diagonal elements of A

16
 Sathyabama University

a22 a23 a11 a13 a11 a12

= + +
a32 a33 a31 a33 a21 a22

C3 = A

CAYLEY HAMILTON THEOREM

Statement:

Every square matrix satisfies it’s own characteristic equation.

Let the characteristic equation of a matrix A of order n be
A − λ I = Co + C1λ + C2 λ 2 + ... + Cn λ n = 0 . Then Cayley Hamilton Theorem
states that Co I + C1 A + C2 A2 + ... + Cn An = 0

Proof:
The elements of ( A − λ I ) are of degree n or less in λ . Hence the elements of
adj ( A − λ I ) are of degree (n-1) or less, in λ .

Therefore adj ( A − λ I ) = Q0 + Q1λ + Q2 λ 2 + ... + Qn −1λ n −1

where Qi ’s are square matrices of order n their elements being polynomials

in λ .

We know
( A − λ I ) adj ( A − λ I ) = A − λ I I

Ÿ ( A − λ I )(Q0 + Q1λ + Q2 λ 2 + ... + Qn −1λ n −1 ) = Co + C1λ + C2 λ 2 + ... + Cn λ n

= ( Co + C1λ + C2 λ 2 + ... + Cn λ n ) I

Equating the coefficient of like powers of λ on both sides, we get

AQ0 = C0 I
AQ1 − Q0 = C1 I
AQ2 − Q1 = C2 I
......
−Q n −1 = Cn I

Pre multiplying these equations by I , A, A2 ,.... An respectively, we obtain

17
Sathyabama University

IAQ0 + A2 Q1 − AQ0 + A3Q2 − A2 Q1 + ... − An Q n −1

= C0 I + AC1 I + A2 C2 I + ... + An Cn I
Ÿ C0 I + C1 A + C2 A2 + ... + Cn An = 0

This proves Cayley-Hamilton theorem.

Applications of Cayley-Hamilton theorem

Two important applications of Cayley-Hamilton theorem are

1. To find the inverse of a non-singular matrix.

If A is a non singular matrix whose characteristic equation is

Co λ n + C1λ n −1 + " + Cn −1λ + Cn = 0

Then by Cayley-Hamilton Theorem, we have

Co An + C1 An −1 + " + Cn −1 A + Cn = 0 (1)

Multipying both sides by A−1 , we have

C0 An -1 +C1 An -2 + ...+Cn -1 I + Cn A-1 = 0

–1
A-1 = [C0 An -1 + C1 An -2 + ...+Cn -1 I]
Cn

2. To obtain the higher powers of a matrix.

Multiplying both sides of equation (1) by A, we have

C0 An +1 + C1 An + ... + Cn −1 A2 + Cn A = 0

−1
Ÿ An +1 = [C1 An + " + Cn −1 A2 + Cn A]
C0

18
 Sathyabama University

WORKED EXAMPLES

Example 1

§7 3· −1
Verify Cayley Hamilton theorem for A = ¨ ¸ and hence find A
© 2 6¹

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 2 − C1λ + C2 = 0

where C1 = 7 + 6 = 13
7 3
C2 = = 42 − 6 = 36
2 6

Ÿ λ 2 − 13λ + 36 = 0 (1)
To verify Cayley-Hamilton theorem, we need to prove that matrix A
satisfies equation(1)
i.e., to prove that A2 − 13 A + 36 I = 0

§ 7 3 · § 7 3 · § 55 39 ·
A2 = ¨ ¸¨ ¸= ¨ ¸
© 2 6 ¹ © 2 6 ¹ © 26 42 ¹

LHS = A2 − 13 A + 36 I
§ 55 39 · § 7 3· §1 0·
=¨ ¸ − 13 ¨ ¸ + 36 ¨ ¸
© 26 42 ¹ © 2 6¹ ©0 1¹
§ 55 − 91 + 36 39 − 39 ·
=¨ ¸
© 26 − 26 42 − 78 + 36 ¹
§0 0·
=¨ ¸
©0 0¹
= RHS

Hence Cayley-Hamilton theorem is verified.

To find A−1 :

Consider equation (2), A2 − 13 A + 36 I = 0

19
Sathyabama University

Post multiply by A−1 on both sides,

A − 13I + 36 A−1 = 0
Ÿ 36 A−1 = 13I − A
§1 0· § 7 3·
Ÿ 36 A−1 = 13 ¨ ¸−¨ ¸
©0 1¹ © 2 6¹
§ 6 −3 ·
=¨ ¸
© −2 −7 ¹
1 § 6 −3 ·
Ÿ A−1 = ¨ ¸
36 © −2 −7 ¹

Example 2
§1 2·
Using Cayley Hamilton theorem find A3 and A−2 for A = ¨ ¸
© 4 3¹

Solution
The characteristic equation of A is given by A − λ I = 0
The general form of characteristic equation is λ 2 − C1λ + C2 = 0
where C1 = 1 + 3 = 4
C2 = 3 − 8 = −5

Ÿ λ 2 − 4λ − 5 = 0 (1)
By Cayley Hamilton theorem, we know that A satisfies equation (1)
i.e., A2 − 4 A − 5 I = 0 (2)
To find A3 :
Post multiply by A on both sides of equation (2)
A3 − 4 A2 − 5 A = 0
Ÿ A3 = 4 A2 + 5 A (3)
From equation (2), A2 = 4 A + 5 I

§1 2· §1 0·
= 4¨ ¸ + 5¨ ¸
©4 3¹ ©0 1¹
§9 8·
=¨ ¸
©16 17 ¹

20
 Sathyabama University

Substituting in equation (3),

§ 9 8 · §1 2·
A3 = 4 A2 + 5 A = 4 ¨ ¸ + 5¨ ¸
© 16 17 ¹ © 4 3 ¹
§ 41 42 ·
Ÿ A3 = ¨ ¸
© 84 83 ¹

To find A-2 :
Post multiply by A−2 on both sides of equation (2)
I − 4 A−1 − 5 A−2 = 0
Ÿ 5 A−2 = I − 4 A−1 (4)
Post multiply by A−1 on both sides of equation (2)
A − 4 I − 5 A−1 = 0

Ÿ 5 A−1 = A − 4 I

§ 1 2 · § 1 0 · § −3 2 ·
=¨ ¸ − 4¨ ¸ =¨ ¸
© 4 3 ¹ © 0 1 ¹ © 4 −1 ¹

1 § −3 2 ·
Ÿ A −1 = ¨ ¸
5 © 4 −1¹

Substituting A−1 in equation (4),

§ 1 0 · 4 § −3 2 ·
5 A−2 = ¨ ¸− ¨ ¸
© 0 1 ¹ 5 © 4 −1 ¹
1 ª§ 5 0 · § −12 8 · º
= «¨ ¸−¨ ¸»
5 ¬© 0 5 ¹ © 16 −4 ¹ ¼
1 § 17 −8 ·
= ¨ ¸
5 © −16 9 ¹
1 § 17 −8 ·
Ÿ A−2 = ¨ ¸
25 © −16 9 ¹

Example 3
§1 3 7·
¨ ¸
Verify Cayley Hamilton theorem for A = ¨ 4 2 3 ¸
¨1 2 1¸
© ¹

21
Sathyabama University

Solution
The characteristic equation of A is given by A − λ I = 0
The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 λ = 0
where C1 = 1 + 2 + 1 = 4

2 3 1 7 1 3
C2 = + +
2 1 1 1 4 2
= −4 − 6 − 10
= −20
1 3 7
C3 = 4 2 3
1 2 1
= 1(2 − 6) − 3(4 − 3) + 7(8 − 2)
= 35

Ÿ λ 3 − 4λ 2 − 20λ − 35 = 0 (1)
To verify Cayley-Hamilton theorem, we need to prove that A satisfies
equation (1)
i.e., to prove that A3 − 4 A2 − 20 A − 35 I = 0

§1 3 7·§1 3 7·
¨
2 ¸¨ ¸
Now A = ¨ 4 2 3 ¸¨ 4 2 3 ¸
¨1 2 1 ¸¨ ¸
© ¹© 1 2 1 ¹
§ 20 23 23 ·
¨ ¸
= ¨ 15 22 37 ¸
¨ 10 9 14 ¸¹
©

§ 20 23 23 ·§ 1 3 7 ·
3 2 ¨ ¸¨ ¸
A = A ⋅ A = ¨ 15 22 37 ¸¨ 4 2 3 ¸
¨ 10 9 14 ¸¨ 1 2 1 ¸
© ¹© ¹
§ 135 152 232 ·
¨ ¸
= ¨ 140 163 208 ¸
¨ ¸
© 60 76 111 ¹

22
 Sathyabama University

LHS = A3 − 4 A2 − 20 A − 35 I
§ 135 152 232 · § 80 92 92 · § 20 60 140 · § 35 0 0 ·
¨ ¸ ¨ ¸ ¨ ¸ ¨ ¸
= ¨ 140 163 208 ¸ − ¨ 60 88 148 ¸ − ¨ 80 40 60 ¸ − ¨ 0 35 0 ¸
¨ 76 111 ¸¹ ¨© 40 36 56 ¸¹ ¨© 20 40 20 ¸¹ ¨© 0 0 35 ¸¹
© 60
§0 0 0·
¨ ¸
= ¨0 0 0¸
¨0 0 0 ¸¹
©
= RHS

Hence Cayley-Hamilton theorem is verified.

Example 4

§1 1 1 ·
¨ ¸
Prove Cayley Hamilton theorem for A = ¨ 1 2 −3 ¸ and hence find the
¨ 2 −1 3 ¸
© ¹
inverse of the matrix.

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 1 + 2 + 3 = 6

2 −3 1 1 1 1
C = + +
2 −1 3 2 3 1 2
= 3 +1+1 = 5
1 1 1
C = 1 2 −3
3
2 −1 3
= 1(6 − 3) − 1(3 + 6) + 1(−1 − 4)
= −11

Ÿ λ 3 − 6λ 2 + 5λ + 11 = 0 (1)

23
Sathyabama University

To verify Cayley-Hamilton theorem, we need to prove that matrix A satisfies

equation(1)
i.e., to prove that A3 − 6 A2 + 5 A + 11I = 0 (2)

§ 1 1 1 ·§ 1 1 1 ·
¨
2 ¸¨ ¸
A = ¨ 1 2 −3 ¸¨ 1 2 −3 ¸
¨ 2 −1 3 ¸¨ 2 −1 3 ¸
© ¹© ¹
§4 2 1 ·
¨ ¸
= ¨ −3 8 −14 ¸
¨ 7 −3 14 ¸
© ¹

§4 2 1 ·§1 1 1 ·
¨ ¸¨ ¸
A3 = A2 ⋅ A = ¨ −3 8 −14 ¸ ¨ 1 2 −3 ¸
¨7 −3 14 ¸¹ ¨© 2 −1 3 ¸¹
©
§ 8 7 1 ·
¨ ¸
= ¨ −23 27 −69 ¸
¨ 32 −13 58 ¸¹
©

LHS = A3 − 6 A2 + 5 A + 11I

§ 8 7 1 · § 24 12 6 · §5 5 5 · § 11 0 0 ·
¨ ¸ ¨ ¸ ¨ ¸ ¨ ¸
= ¨ −23 27 −69 ¸ − ¨ −18 48 −84 ¸ + ¨ 5 10 −15 ¸ + ¨ 0 11 0 ¸
¨ ¸ ¨ ¸ ¨ ¸ ¨ ¸
© 32 −13 58 ¹ © 42 −18 84 ¹ © 10 −5 15 ¹ © 0 0 11¹

§0 0 0·
¨ ¸
= ¨0 0 0¸
¨0 0 0¸
© ¹

= RHS
Hence Cayley-Hamilton theorem is verified

To find A–1

Post multiply by A−1 on both sides of (2)

A2 − 6 A + 5 I + 11A−1 = 0
Ÿ 11A−1 = − A2 + 6 A − 5 I

24
 Sathyabama University

§ −4 −1 · § 6 6
−2 6 · §5 0 0·
¨ ¸ ¨ ¸ ¨ ¸
=¨ 3 −8 14 ¸ + ¨ 6 12 −18 ¸ − ¨ 0 5 0 ¸
¨ −7 3 −14 ¸¹ ¨©12 −6 18 ¸¹ ¨© 0 0 5 ¸¹
©
§ −34 5·
¨ ¸
=¨ 9 −1 − 4 ¸
¨5 −3 −1 ¸¹
©
§ −3 4 5 ·
1¨ ¸
Ÿ A = ¨ 9 −1 −4 ¸
−1

11 ¨ ¸
© 5 −3 −1 ¹

Example 5

§1 0 3 ·
¨ ¸
Prove that the matrix A = ¨ 2 1 −1¸ satisfies its own characteristic
¨ 1 −1 1 ¸
© ¹
4 −1
equation and hence find A and A

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 1 + 1 + 1 = 3

1 −1 1 3 1 0
C2 = + + = −1
1 1 1 1 2 1
C3 = A = −9

Ÿ λ 3 − 3λ 2 − λ + 9 = 0 (1)
To prove that A satisfies it’s own characteristic equation, we need to prove
that
A3 − 3 A2 − A + 9 I = 0 (2)

§ 1 0 3 ·§ 1 0 3 ·
2 ¨ ¸¨ ¸
A = ¨ 2 1 −1¸¨ 2 1 −1¸
¨ 1 −1 1 ¸¨ 1 −1 1 ¸
© ¹© ¹

25
Sathyabama University

§ 4 −3 6 ·
¨ ¸
= ¨ 3 2 4¸
¨ 0 −2 5 ¸
© ¹
§ 4 −3 6 ·§ 1 0 3 · § 4 −9 21·
¨ ¸¨ ¸ ¨ ¸
A3 = A2 ⋅ A = ¨ 3 2 4 ¸¨ 2 1 −1¸ = ¨11 −2 11 ¸
¨ 0 −2 5 ¸¨ ¸ ¨ ¸
© ¹© 1 −1 1 ¹ © 1 −7 7 ¹
3 2
LHS = A − 3 A − A + 9 I
§ 4 −9 21· §12 −9 18 · § 1 0 3 · § 9 0 0 ·
¨ ¸ ¨ ¸ ¨ ¸ ¨ ¸
= ¨11 −2 11 ¸ − ¨ 9 6 12 ¸ − ¨ 2 1 −1¸ + ¨ 0 9 0 ¸
¨ 1 −7 7 ¸ ¨ 0 −6 15 ¸ ¨ 1 −1 1 ¸ ¨ 0 0 9 ¸
© ¹ © ¹ © ¹ © ¹
§0 0 0·
¨ ¸
= ¨0 0 0¸
¨0 0 0¸
© ¹
= RHS
Hence Cayley-Hamilton theorem is verified.

To find A4 :

Post multiply by A on both sides of (2), we get

A4 − 3 A3 − A2 + 9 A = 0

Ÿ A4 = 3 A3 + A2 − 9 A
§ 12 −27 63 · § 4 −3 6 · § 9 0 27 ·
¨ ¸ ¨ ¸ ¨ ¸
= ¨ 33 −6 33 ¸ + ¨ 3 2 4 ¸ − ¨18 9 −9 ¸
¨ 3 −21 21 ¸ ¨ 0 −2 5 ¸ ¨ 9 −9 9 ¸
© ¹ © ¹ © ¹

§ 7 −30 42 ·
¨ ¸
= ¨ 18 −13 46 ¸
¨ ¸
© −6 −14 17 ¹

To find A-1 :
Post multiply by A−1 on both sides of equation (2), we get

A2 − 3 A − I + 9 A−1 = 0
Ÿ 9 A−1 = − A2 + 3 A + I

26
 Sathyabama University

§ −4 3 −6 · § 3 0 9 · § 1 0 0 ·
¨ ¸ ¨ ¸ ¨ ¸
= ¨ −3 −2 −4 ¸ + ¨ 6 3 −3 ¸ + ¨ 0 1 0 ¸
¨ 0 2 −5 ¸ ¨ 3 −3 3 ¸ ¨ 0 0 1 ¸
© ¹ © ¹ © ¹
§0 3 3 ·
¨ ¸
= ¨ 3 2 −7 ¸
¨ 3 −1 −1 ¸
© ¹
§0 3 3 ·
1¨ ¸
Ÿ A = ¨ 3 2 −7 ¸
−1

9¨ ¸
© 3 −1 −1 ¹

Example 6

§ 1 2 −1 ·
¨ ¸
Using Cayley Hamilton theorem, find the inverse of the matrix ¨ 3 −3 1 ¸
¨ 2 1 −2 ¸
© ¹

Solution

§ 1 2 −1 ·
¨ ¸
Let A = ¨ 3 −3 1 ¸
¨ 2 1 −2 ¸
© ¹

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 1 − 3 − 2 = −4

−3 1 1 −1 1 2
C2 = + + = −4
1 −2 2 −2 3 −3
C3 = A = 12
Ÿ λ 3 + 4λ 2 − 4λ − 12 = 0
By Cayley-Hamilton theorem, A satisfies it’s own characteristic
equation
i.e., A3 + 4 A2 − 4 A − 12 I = 0
Post multiply by A−1 on both sides,

27
Sathyabama University

A2 + 4 A − 4 I − 12 A−1 = 0
Ÿ 12 A−1 = A2 + 4 A − 4 I (1)

§1 2 −1 · § 1 2 −1 ·
2 ¨ ¸¨ ¸
A = A⋅ A = ¨ 3 −3 1 ¸ ¨ 3 −3 1 ¸
¨2 −2 ¹¸ ©¨ 2 1 −2 ¹¸
© 1
§ 5 −5 3·
¨ ¸
= ¨ −4 16 −8 ¸
¨1 −1 3 ¸¹
©
Substituting in equation (1)

§ 5 −5 3· §4 8 −4 · § 4 0 0 ·
¨
−1 ¸ ¨ ¸ ¨ ¸
12 A = ¨ −4 16 −8 ¸ + ¨12 −12 4 ¸ − ¨ 0 4 0 ¸
¨1 −1 3 ¸¹ ¨© 8 4 −8 ¸¹ ¨© 0 0 4 ¸¹
©
§5 3 −1 ·
¨ ¸
= ¨8 0 −4 ¸
¨9 3 −9 ¸¹
©
§ 5 3 −1 ·
1 ¨ ¸
A−1 = ¨ 8 0 −4 ¸
12 ¨ ¸
© 9 3 −9 ¹

Example 7

§ 7 2 −2 ·
4 ¨ ¸
Compute A using Cayley-Hamilton theorem for A= ¨ −6 −1 2 ¸
¨ 6 2 1¸
© ¹
Solution
The characteristic equation of A is given by A − λ I = 0
The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 7í1+1=7
−1 2 7 −2 7 2
C2 = + + = 19
2 1 6 1 −6 −1
C 3 = A = 13
Ÿ λ 3 − 7λ 2 + 19λ − 13 = 0

28
 Sathyabama University

By Cayley-Hamilton theorem, A satisfies it’s own characteristic equation.

i.e, A3 − 7 A2 + 19 A − 13I = 0 (1)
Post multiply by A on both sides,
A4 − 7 A3 + 19 A2 − 13 A = 0

Ÿ A4 = 7 A3 − 19 A2 + 13 A (2)

§ 7 2 −2 · § 7 2 −2 · § 25 8 −12 ·
2¨ ¸¨ ¸ ¨ ¸
Consider A = ¨ −6 −1 2 ¸ ¨ −6 −1 2 ¸ = ¨ −24 −7 12 ¸
¨ 6 2 1 ¸ ¨ 6 2 1 ¸ ¨ 36 12 −7 ¸
© ¹© ¹ © ¹
From equation (1),
§ 25 8 −12 · § 7 2 −2 · §1 0 0·
¨ ¸ ¨ ¸ ¨ ¸
A3 = 7 A2 − 19 A + 13I = 7 ¨ −24 −7 12 ¸ − 19 ¨ −6 −1 2 ¸ + 13 ¨ 0 1 0 ¸
¨ 36 12 −7 ¸ ¨6 2 1¸ ¨0 0 1¸
© ¹ © ¹ © ¹
§ 55 18 −46 ·
¨ ¸
= ¨ −54 −17 46 ¸
¨ ¸
© 138 46 −55 ¹
Substituting in equation (2),
§ 55 18 −46 · § 25 8 −12 · § 7 2 −2 ·
4 ¨ ¸ ¨ ¸ ¨ ¸
A = 7 ¨ −54 −17 46 ¸ − 19 ¨ −24 −7 12 ¸ + 13 ¨ −6 −1 2 ¸
¨ 138 46 −55 ¸ ¨ ¸ ¨ ¸
© ¹ © 36 12 −7 ¹ © 6 2 1¹
§ 1 0 −120 ·
¨ ¸
=¨ 0 1 120 ¸
¨ 360 120 −239 ¸
© ¹

Example 8

§1 0 0·
¨ ¸
If A = ¨ 1 0 1 ¸ then show that An = An + 2 + A2 − I for n ≥ 3 using Cayley-
¨0 1 0¸
© ¹
Hamilton theorem.

Solution
The characteristic equation of A is given by A − λ I = 0

29
Sathyabama University

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1=1+0+0=1
0 1 1 0 1 0
C2 = + + = −1
1 0 0 0 1 0
C3 = A = − 1
Ÿ λ3 − λ2 − λ +1 = 0
By Cayley-Hamilton theorem, A satisfies it’s own characteristic
equation.
i.e., A3 − A2 − A + I = 0
Ÿ A3 − A2 = A − I

Post multiplying by A,
A4 − A3 = A2 − A
A5 − A4 = A3 − A2
#
An − An −1 = An − 2 − An − 3

Adding, we get
An − A2 = An − 2 − I
Ÿ An = An − 2 + A2 − I

Example 9

§ 1 2·
If A = ¨ ¸, use Cayley Hamilton theorem to express
© −1 3 ¹
A6 − 4 A5 + 8 A4 − 12 A3 + 14 A2 as a linear polynomial in A and hence find the
value of it.

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 2 − C1λ + C2 = 0

30
 Sathyabama University

where C1 = 1 + 3 = 4
C2 = A = 5

Ÿ λ 2 − 4λ + 5 = 0

By Cayley-Hamilton theorem, A satisfies it’s own characteristic equation.

i.e., A2 − 4 A + 5I = 0 (1)

Consider A6 − 4 A5 + 8 A4 − 12 A3 + 14 A2

= A4 ( A2 − 4 A + 5 I ) + 3 A4 − 12 A3 + 14 A2
= 0 + 3 A4 − 12 A3 + 15 A2 − A2 (from equation (1))
= 3 A 2 ( A 2 − 4 A + 5 I ) − A2 \
= − A2 (from equation (1))
= −4 A + 5 I (using equation (1))

§ 1 2· §1 0·
= −4 ¨ ¸ + 5¨ ¸
© −1 3 ¹ © 0 1 ¹
§ 1 −8 ·
=¨ ¸
© 4 −7 ¹

Example 10

Using Cayley-Hamilton theorem find An , n is a positive integer given that

§1 2·
A= ¨ ¸
© 4 3¹

Solution

The Characteristic equation is given by A − λ I = 0

The general form of characteristic equation is λ 2 − C1λ + C2 = 0

where C1 = 1 + 3 = 4
C 2 = A = −5

Ÿ λ 2 − 4λ − 5 = 0

By Cayley-Hamilton theorem, we have

31
Sathyabama University

A2 − 4 A − 5 I = 0 (1)

Let λ n = (λ 2 − 4λ − 5) Q ( λ ) + ( aλ + b ) (2)

where Q ( λ ) is the quotient

n
When λ =-1, ( −1) = −a + b

When λ = 5, 5 n = 5a + b

Solving the above equations, we get

n n
5n − ( −1) 5n + 5 ( −1)
a= and b =
6 6
Replacing λ by matrix A in equation (2), we get

An = ( A2 − 4 A − 5 I )Q ( A ) + ( aA + bI )
= aA + bI ( using eqn (1) )

ª 5n − ( −1) n º § 1 2 · ª 5n + 5 ( −1) n º § 1 0 ·
=« »¨ ¸+« »¨ ¸
«¬ 6 »¼ © 4 3 ¹ ¬« 6 »¼ © 0 1 ¹

Example 11

Use Cayley-Hamilton theorem to find the value of the matrix given by

A8 − 5 A7 + 7 A6 − 3 A5 + A4 − 5 A3 + 8 A2 − 2 A + I if the matrix
§2 1 1·
¨ ¸
A= ¨ 0 1 0 ¸
¨1 1 2¸
© ¹

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 2 + 1 + 2 = 5

1 0 2 1 2 1
C2 = + + =7
1 2 1 2 0 1

32
 Sathyabama University

C3 = A =3

Ÿ λ 3 − 5λ 2 + 7λ − 3 = 0

By Cayley-Hamilton theorem, A satisfies its own characteristic equation.

i.e., A3 − 5 A2 + 7 A − 3I = 0 (1)

Consider A8 − 5 A7 + 7 A6 − 3 A5 + A4 − 5 A3 + 8 A2 − 2 A + I

( )
= A5 A3 − 5 A + 7 A − 3I + A ( A3 − 5 A2 + 7 A − 3I ) + ( A2 + A + I )
2

= A2 + A + I (using equation(1))

§ 2 1 1 ·§ 2 1 1 · § 2 1 1 · § 1 0 0 ·
¨ ¸¨ ¸ ¨ ¸ ¨ ¸
= ¨ 0 1 0 ¸¨ 0 1 0 ¸ + ¨ 0 1 0 ¸ + ¨ 0 1 0 ¸
¨ 1 1 2 ¸¨ 1 1 2 ¸ ¨ 1 1 2 ¸ ¨ 0 0 1 ¸
© ¹© ¹ © ¹ © ¹

§5 4 4· § 2 1 1· §1 0 0·
¨ ¸ ¨ ¸ ¨ ¸
= ¨0 1 0¸ + ¨ 0 1 0¸ + ¨ 0 1 0¸
¨4 4 5 ¸¹ ¨© 1 1 2 ¸¹ ¨© 0 0 1 ¸¹
©
§8 5 5·
¨ ¸
= ¨0 3 0¸
¨5 5 8 ¹¸
©

Example 12

§ 1 2 −1 ·
¨ ¸
Given A = ¨ 3 −3 1 ¸ find Adj A using Cayley-Hamilton theorem.
¨ 2 1 −2 ¸
© ¹

Solution

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 1 − 3 − 2 = −4

33
Sathyabama University

−3 1 1 −1 1 2
C2 = + + = −4
1 −2 2 −2 3 −3

C3 = A = 12

Ÿ λ 3 + 4λ 2 − 4λ − 12 = 0
By Cayley-Hamilton theorem, the given matrix A satisfies it’s own
characteristic equation.
i.e., A3 + 4 A2 − 4 A − 12 I = 0
Post multiplying by A−1 on both sides,

A2 + 4 A − 4 I − 12 A−1 = 0
12 A−1 = A2 + 4 A − 4 I

§ 1 2 −1 ·§ 1 2 −1 · § 1 2 −1 · § 1 0 0 ·
¨ ¸¨ ¸ ¨ ¸ ¨ ¸
= ¨ 3 −3 1 ¸¨ 3 −3 1 ¸ + 4 ¨ 3 −3 1 ¸ − 4 ¨ 0 1 0 ¸
¨ 2 1 −2 ¸¨ 2 1 −2 ¸ ¨ 2 1 −2 ¸ ¨ 0 0 1 ¸
© ¹© ¹ © ¹ © ¹

§5 −5 3· §4 8 −4 · § −4 0 0 ·
¨ ¸ ¨ ¸ ¨ ¸
i.e., 12 A−1 = ¨ −4 16 −8 ¸ + ¨12 −12 4 ¸ + ¨ 0 −4 0 ¸
¨1 −1 3 ¸¹ ¨© 8 4 −8 ¸¹ ¨© 0 0 −4 ¸¹
©
§5 3 −1 ·
1¨ ¸
Ÿ A = ¨8
−1
0 −4 ¸
12 ¨
©9 3 −9 ¹¸
1
A −1 = Adj A
A
Ÿ Adj A = A A−1

§ 5 3 −1 ·
1¨ ¸
= 12 × ¨ 8 0 −4 ¸
12 ¨ ¸
© 9 3 −9 ¹
§ 5 3 −1 ·
¨ ¸
Ÿ Adj A = ¨ 8 0 −4 ¸
¨ 9 3 −9 ¸
© ¹

34
 Sathyabama University

EXERCISE

Part A

1. Give two applications of Cayley-Hamilton theorem.

ª −2 −2 º
2. Verify Cayley-Hamilton theorem for the matrix A = « »
¬ −5 1 ¼
ª1 1 º
3. If A = « » express A4 in terms of A and I.
¬1 2 ¼
ª 2 −1º
4. Use Cayley-Hamilton theorem to find A-1 if A = « »
¬4 3 ¼

ª1 2 º
5. Given A = « » compute A3 and A-3.
¬1 1 ¼
ª1 1 2º
6. Check whether A – 3A – 7A – 11I = 0 where A = «« 3 1 1 »»
3 2

«¬ 2 3 1 »¼

ª1 0 º
7. If A = « » express A3 in terms of A and I and find its value.
¬4 5¼
8. Use Cayley-Hamilton theorem to find (A4 – 4A3 – 5A2 + A + 2I) where
ª1 2 º
A=« »
¬4 3¼
Part B

§2 2 1·
¨ ¸
9. Verify Cayley-Hamilton theorem for A = ¨ 1 3 1 ¸
¨1 2 2¸
© ¹

§1 2 3·
¨ ¸
10. Prove that the matrix ¨ 2 1 4 ¸ satisfies its own characteristic
¨3 1 1¸
© ¹
3
equation and hence find A .
11. Using Cayley-Hamilton theorem, simplify the expression
§ 1 2 −2 ·
( A − A + 5 A − A + A − A + 6 A + A − 2I ) if A = ¨¨ 2 5 −4 ¸¸
8 7 6 5 4 3 2

¨ 3 7 −5 ¸
© ¹

35
Sathyabama University

12. Use Cayley-Hamilton theorem to find the value of the matrix given
§ 0 0 2·
6 2 ¨ ¸
by A − 25 A + 122 A if the matrix A = ¨ 2 1 0 ¸
¨ −1 −1 3 ¸
© ¹

13. Find A4 using Cayley-Hamilton theorem for the matrix

§1 2 3 ·
¨ ¸
A = ¨ 2 −1 4 ¸
¨ 3 1 −1¸
© ¹

14. Verify Cayley-Hamilton theorem and hence find A−1 and A4 ,

§3 1 1·
¨ ¸
A = ¨ −1 5 −1 ¸
¨ 1 −1 3 ¸
© ¹

§a b·
15. Show that the matrix A = ¨ ¸ satisfies its own characteristic
©c d ¹
equation and hence find A-1
ª 5 3º
16. Find An using Cayley-Hamilton theorem, where A = « » . Hence
¬1 3 ¼
find A4.
ª7 3º
17. Find An using Cayley-Hamilton theorem where A = « » . Hence
¬ 2 6 ¼
3
find A .
18. Using Cayley-Hamilton theorem find Adj A of the matrix
ª −1 0 3 º
A = «« 8 1 −7 »»
«¬ −3 0 8 ¼»

ª8 −8 2 º
19. Verify Cayley-Hamilton theorem and hence find A for A = « 4 −3 −2 »
-1
« »
«¬ 3 −4 1 »¼

ª1 1 2º
20. Find A where A = «« 3 1 1 »»
–2

«¬ 2 3 1 ¼»

36
 Sathyabama University

ª1 0 3 º
21. Given that A = «« 2 1 −1»» compute the value of
«¬1 −1 1 ¼»
A6 – 5 A5 + 8 A4 – 2 A3 – 9 A2 + 31A − 36 I using Cayley-Hamilton theorem.

ANSWERS

Part A

3. A4 = 21A – 8 I

1 ª 3 1º
4. A−1 =
10 «¬ −4 2 ¼»

ª 7 10 º
5. A3 = « »
¬5 7 ¼

ª −7 10 º
A−3 = « »
¬ 5 −7 ¼

ª 1 0 º
7. A3 = 31A – 30 I = « »
¬124 125¼

ª3 2º
8. « »
¬4 5¼

Part B

ª 70 49 84 º
10. A3 = «« 76 55 98 »»
«¬ 66 38 70 »¼

11. A2 + 2 A – I

ª −44 0 −20 º
12. «« −20 −54 0 »»
«¬ 10 10 −74 »¼

37
Sathyabama University

ª 248 101 218º

13. A = − A + 18 A + 40 A = «« 212 109 50 »»
4 3 2

«¬104 98 204 »¼

ª 7 −2 −3 º
1 «
−1
14. A = « 1 4 1 »»
20
«¬ −2 2 8 »¼

ª 53 203 37 º
A4 = «« −369 625 −369 »»
«¬ 203 −203 219 ¼»

1 ª d −b º
15. A−1 =
ad − bc ¬« −c a »¼

n § 6n − 2n · ª5 3º § 3.2n − 6n · ª1 0 º
16. A = ¨ ¸« »+¨ ¸« »
© 4 ¹ ¬1 3 ¼ © 2 ¹ ¬0 1¼

ª976 960 º
A4 = « »
¬320 336 ¼

n § 9n − 4n · ª 7 3 º § 9.4n − 4.9n · ª1 0 º
17. A = ¨ ¸« »+¨ ¸« »
© 5 ¹ ¬2 6¼ © 5 ¹ ¬0 1 ¼
ª 463 399 º
A3 = « »
¬ 266 330 ¼
ª −8 0 −3º
« »
18. Adj A = « −43 1 17 »
¬« 3 0 −1»¼

ª11 0 −22 º
1 «
−1
19. A = « 10 −2 −24 »»
22
«¬ 7 −8 −8 »¼

ª −8 −24 29 º
1 «
−2
20. A = « 40 −1 −24 »»
121
«¬ −27 40 −8 »¼

38
 Sathyabama University

ª0 0 0º
21. «« 0 0 0 »»
«¬ 0 0 0 »¼

Eigen values and Eigen vectors

( ) is a square matrix of order n, then

If A = aij A − λ I = 0 is called the
characteristic equation of the matrix A. On expanding the determinant, the
characteristic equation takes the form,
n
( −1) λ n + k1λ n −1 + k2 λ n − 2 + ....... + kn = 0 where the ki’s are expressible in
terms of the elements aij. The roots of the characteristic equation are called
eigen values or characteristic roots or latent values.
ª x1 º ª a11 a12 ! a1n º
«x » «a a22 ! a2 n »»
If X = « 2»
and A = « 21
«#» « # »
« » « »
¬ xn ¼ ¬ an1 an 2 ! ann ¼

then the linear transformation Y = AX transforms the column vector X into the
column vector Y by means of the square matrix A.
In practice, it is often required to find such vectors which transform into
themselves or to a scalar multiple of themselves. Let X be such a vector which
transforms into ȜX by means of the transformation AX.
Then λ X = AX
i.e., AX − λ IX = 0
.
Ÿ ( A − λI ) X = 0

This matrix equation represents n homogeneous linear equations,

( a11 − λ ) x1 + a12 x2 + " + a1n xn = 0
a21 x1 + ( a22 − λ ) x2 + " + a2 n xn = 0
""""""""""""""
an1 x1 + an 2 x2 + " + ( ann − λ ) xn = 0

which will have a non-trivial solution only if the coefficient matrix ( A − λ I )

is singular, i.e., if | A − λ I |= 0 . This is called the characteristic equation of
the transformation and is the same as the characteristic equation of the matrix

39
Sathyabama University

A. It has n roots and corresponding to each root, the equation will have a non-
ª x1 º
«x »
zero solution X = « 2 » which is known as the eigen vector or latent vector.
«#»
« »
¬ xn ¼

Properties of Eigen values

1. A square matrix A and its transpose AT have the same eigen value.

Proof: Let A = [aij] for i, j = 1, 2,…n be a square matrix of order n.

The characteristic polynomial of A is | A − λ I | .

a11 − λ a12 ..... a1n

a21 a22 − λ ..... a2 n
i.e., (1)
" " " "
an1 an 2 ..... ann − λ

The characteristic polynomial of AT is AT − λ I .

a11 − λ a21 ..... an1

a12 a22 − λ ..... an 2
i.e., (2)
" " " "
a1n a2 n ..... ann − λ
Equation (2) is obtained from equation (1) by changing its row into
columns. Since A − λ I = AT − λ I , A and AT have the same
characteristic equation. Hence A and AT have the same eigen values.
2. The sum of the eigen values of a matrix A is equal to the sum of the
principal diagonal elements of A. i.e., trace of A.
Proof: The characteristic equation of an nth order square matrix A may
n
be written as λ n − C1λ n −1 + C2 λ n − 2 + ...... + ( −1) Cn = 0 (1)

where Cr is the sum of all the rth order minors of A, whose principal
diagonals lie along the principal diagonal of A. If Ȝ1, Ȝ2, Ȝ3,…Ȝn are the
eigen values of A, then they are the roots of equation (1).
− ( −C1 )
= = C1
Hence sum of the roots 1
Ÿ λ1 + λ2 + ... + λn = a11 + a22 + .... + ann

40
 Sathyabama University

Hence sum of eigen values = Trace of the matrix A.

3. The product of eigen values of a matrix A is equal to the determinant
of A.
Proof: Let λ1 , λ2 ,....λn be the eigen values of A.
n
Then A − λ I = ( −1) ( λ − λ1 )( λ − λ2 ) ...... ( λ − λn )
n
Put λ = 0. Then | A |= ( −1) ( −λ1 )( −λ2 ) ... ( −λn )
n n
i.e., | A | = ( −1) ( −1) λ1λ2 ...λn = λ1λ2 ...λn

∴ Product of the eigen values = determinant of A

4. If λ1 , λ2 ,....λn are the eigen values of a matrix A, then

i. k λ1 , k λ2 , ! k λn are the eigen values of the matrix kA where k is

a non-zero scalar.
ii. λ1p , λ2p ,! λnp are the eigen values of the matrix AP, where p is a
positive integer.
1 1 1
iii. , ," , are eigen values of the matrix A-1, provided
λ1 λ2 λn

λr ≠ 0 for r = 1, 2, …, n.

Proof: (i) Let Ȝr be an eigen value of A and Xr be the corresponding eigen

vector.
Then AX r = λr X r (1)
Pre multiplying both sides of (1) by k
k ( AX r ) = k ( λr X r )
( kA) X r = ( k λr ) X r
Hence kȜr is an eigen value of kA and the corresponding eigen vector is Xr
(ii) Pre multiplying both sides of (1) by A, we get
A ( AX r ) = A ( λr X r )
i.e., A2 X r = λr ( AX r )
= λr ( λr X r ) using equation (1)
= λr2 X r

41
Sathyabama University

Similarly it can be proved that A3 X r = λr3 X r . In general

A p X r = λrp X r for any positive integer p. Hence λrp is an eigen value
of AP with the corresponding eigen vector as Xr.
(iii) Pre multiplying both sides of (1) by A-1, we get
A−1 ( AX r ) = A−1 ( λr X r )
i.e., IX r = λr ( A−1 X r )
1
∴ . X r = A−1 X r ; provided λr ≠ 0
λr
1
Hence is the eigen value of A-1.
λr

5. If λ1 , λ2 ,....λn are the eigen values of A, then λ1 − k , λ2 – k ,... λn − k

are the eigen values of A – kI
Proof: Let Xr be the eigen vector corresponding to the eigen value Ȝr.
Then
AX r = λ2 X r (1)

subtract kIX r from both sides of (1), we get

AX r − kIX r = λr X r − kIX r
i.e., ( A − kI ) X r = λr X r − kX r
i.e., ( A − kI ) X r = ( λr − k ) X r

Hence λr – k is an eigen value of the matrix A − kI with the

corresponding eigen vector Xr.
6. The eigen values of a real symmetric matrix are all real.
Proof: Let Ȝ be an eigen value of the real symmetric matrix A and X, the
corresponding eigen vector.
Then AX = λ X (1)

Taking complex conjugate of (1) AX = λ X

Since A =A, we get A X = λ X (2)

Taking transpose of (2) we get,

X * AT = λ X *

42
 Sathyabama University

Since AT = A we get X * A = λ X *
Post multiplying by X, we get
X * AX = λ X * X
i.e., X* λ X = λ X * X using (1)
i.e., λ X * X = λ X * X
(λ − λ ) X *
X =0

ª x1 + iy1 º
« x + iy »
If X = « 2 2»
then X T X = x12 + ..... + xn2 + y12 + ... + yn2
« # »
« »
¬ xn + iyn ¼

Since X T X ≠ 0 then λ − λ = 0 Ÿ λ = λ

∴ Ȝ is real.
Hence all the eigen values of a real symmetric matrix are real.
7. The absolute value of the eigen values of an orthogonal matrix is 1.
Proof: Let A be an orthogonal matrix and Ȝ be an eigen value of A. If X
is the corresponding eigen vector, then
AX = λ X (1)

Taking complex conjugate we get, AX = λ X .

Since A is real, A X = λ X
T
( )
T
Taking transpose we get AX = λX

i.e., X * AT = λ X * (2)

Consider (1)×(2) X * AT AX = λ X *λ X
2
Since AT A = I we get, X * X = λ X * X

( )
Ÿ λ −1 X * X = 0
2

λ −1 = 0 Ÿ λ
2 2
Since X * X ≠ 0 = 1or λ = 1.

43
Sathyabama University

8. The eigen values of any triangular matrix (upper or lower) are its
leading diagonals.
Proof: For any triangular matrix or diagonal matrix, the characteristic
equation is

( a11 − λ )( a22 − λ ) ............. ( ann − λ ) = 0.

Hence the eigen values are a11, a22,….,ann Thus for any triangular matrix
or diagonal matrix the entries on its leading diagonal are its eigen values.
Note: The eigen values of I are equal to one.
9. Zero is an eigen value of A if and only if A is a singular matrix.
Proof: The eigen values of A are the roots of the characteristic equation
|A – ȜI| = 0. If 0 is an eigen value then A − 0 I = 0

⇔| A |= 0

⇔ A is a singular matrix.
10. If A and B are two square matrices of order n then AB and BA have
the same eigen values.
Proof: Let Ȝ be an eigen value of AB and X be an eigen vector
corresponding to Ȝ then

( AB ) X = λX

Pre multiplying by B, we get

B ( AB ) X = B ( λ X ) = λ ( BX )
ie., ( BA )( BX ) = λ ( BX )
Let BX = Y . Then ( BA ) Y = λY

Ÿ λ is an eigen value of BA, where BX is the corresponding eigen

vector.
11. If P and A are n×n matrices and P is a non-singular matrix then A
and P −1 AP have the same eigen value.
Proof: Let B = P −1 AP . To prove that both A and B have the same
eigen values, it is sufficient to prove that both have the same
characteristic polynomial.

44
 Sathyabama University

B − λ I = P −1 AP − λ I
= P −1 AP − P −1 P ( λ I )
= P −1 ( A − λ I ) P
= P −1 P A − λ I
= I A − λI
= A − λI

Thus the Characteristic equations of A and P −1 AB are the same.

12. If Ȝ is a characteristic root of A, then f(Ȝ) is a characteristic root of
f(A) where f(x) is any polynomial.
Proof: Let f ( x ) = a0 x n + a1 x n −1 + ..... + an −1 x + an

f ( A ) = a0 An + a1 An −1 + ..... + an −1 A + an I

If Ȝ is a characteristic root of A, with characteristic vector X, then Ȝn is a

characteristic root of An, for any positive integer n. (by property 4(ii)]
∴ An X = λ n X
An −1 X = λ n −1 X etc.
Hence a0 An X = a0 λ n X
a1 An −1 X = a1λ n −1 X etc.
an −1 AX = an −1λ X

Adding the above equation

a0 An X + a1 An −1 X + ...... + an −1 AX = a0 λ n X + a1λ n −1 X + ...... + an −1λ X

Adding an I X to both sides of the above equation we get

( a0 An + a1An−1 + ....... + an−1A + an I ) X

= (a λ n + a λ n−1 + ......... + a λ + an ) X
0 1 n−1
∴ f ( A) X = f ( λ ) X

Hence f(Ȝ) is a characteristic root of f(A).

| A|
13. If Ȝ is an eigen value of a non-singular matrix A, then is an
Ȝ
eigen value of adj A

45
Sathyabama University

Proof: Let Ȝ be an eigen value of with eigen vector X, then AX = ȜX

Pre multiplying by adj A, we get
( adj A) AX = ( adj A) λ X
i.e., ( ( adj A) A) X = λ ( adj A) X
i.e., | A | I X = λ ( adj A ) X since ( adj A ) A =| A | I

∴ | A | X = λ ( adj A ) X
| A|
X = ( adj A ) X
λ
| A|
Hence is an eigen value of adj A.
λ
14. The eigen values of a Hermitian matrix A are all real.
Proof: Let Ȝ be an eigen value of a Hermitian matrix A with eigen vector X.
Then AX = λ X

Pre multiply by X * we get , X * AX = X * λ X

i.e., X * AX = λ X * X (1)
Taking conjugate transpose of (1), we get
* *
(X *
AX ) = ( λ X * X )
* *
X * A* ( X * ) = X * ( X * ) λ
X * AX = X * X λ since A is a Hermitian matrix, A* = A
λX*X = X*X λ using (1)
Ÿλ =λ
Thus λ is real.

15. The eigen values of a skew hermitian matrix is either zero or a

purely imaginary number.
Proof: Let Ȝ be an eigen value of a skew-hermitian matrix with X as its
eigen vector. Then AX = λ X

( iA) X = i (λ X ) = (i λ ) X
Hence iȜ is an eigen value of the hermitian matrix iA. Since the
characteristic roots of a Hermitian matrix are real, iȜ is real. iȜ will be
real if either Ȝ = 0 or Ȝ is purely imaginary.

46
 Sathyabama University

Note: Eigen values of a skew symmetric matrix is either zero or purely

imaginary.

Properties of Eigen vectors

1. The eigen vector corresponding to an eigen value is not unique.
Proof: Let Ȝ be an eigen value of a matrix A with eigen vector X. Then
AX = λ X
For any non zero scalar k
k ( AX ) = k ( λ X )

A ( kX ) = λ ( kX )

Thus kX is a eigen vector of A corresponding to the same eigen value Ȝ

2. Every eigen vector corresponds to a unique eigen value.
Proof: Let there exists two distinct eigen values Ȝ1 and Ȝ2 corresponding
to an eigen vector X of a square matrix A.
Then we have AX = λ1 X

and AX = λ2 X

∴ λ1 X = λ2 X

i.e., (Ȝ1 – Ȝ2)X = 0

Ÿ X = 0 since Ȝ1 – Ȝ2  0.
Hence every eigen vector corresponds to a unique eigen value.
3. If all the eigen values λ1 , λ2 ,....λn of a matrix A are distinct, then
the corresponding eigen vectors X 1, X 2 , ….., X n are linearly
independent.
Proof: Let Ȝ1, Ȝ2,…, Ȝn be distinct eigen values of a matrix A and let Xi be
the eigen vector corresponding to Ȝi
Hence AX i = λi X i ( i = 1, 2,...., n )
Let X 1 , X 2 , …, X n be linearly dependent. Then there exists
scalars α1 , α 2 , …. , α n , not all zero, such
that α1 X 1 + α 2 X 2 + " + α n X n = 0

By rearranging we may assume that the relation with shortest length j is

47
Sathyabama University

α1 X 1 + α 2 X 2 + " + α j X j = 0 (1)

where all Įi’s are not zero

∴ A (α1 X 1 ) + A (α 2 X 2 ) + ....... + A (α j X j ) = 0
α1 ( AX 1 ) + α 2 ( AX 2 ) + ....... + α j ( AX j ) = 0

α1 ( λ1 X 1 ) + α 2 ( λ2 X 2 ) + ....... + α j ( λ j X j ) = 0 (2)

(1) × Ȝ1 – (2) gives

α 2 ( λ1 − λ2 ) X 2 + ......... + α j ( λ1 − λ j ) X j = 0 (3)

Since all Ȝi’s are distinct and Į2,……, Įj are non zero (3) gives a shorter
relation than (1), giving rise to a contradiction. Hence X1, X2, …, Xn are
linearly independent.
Note: We give examples to show that if two or more eigen values of a matrix
are equal, then the eigen vectors may be linearly dependent or linearly
independent.

ª2 1 0º
Example (1): Let A = «« 0 2 0 »»
«¬ 0 0 2 »¼

2−λ 1 0
The characteristic equation of A is 0 2−λ 0 =0
0 0 2−λ

Ÿ ( 2 − λ )( 2 − λ )( 2 − λ ) = 0
Ÿ λ = 2, 2, 2 are the eigen values of A.
ªxº
If X = «« y »» is an eigen vector corresponding to the eigen value λ = 2,
«¬ z ¼»
then AX = 2 X
ª2 1 0º ª x º ª xº
«
i.e., « 0 2 0 » « y » = 2 «« y »»
» « »
«¬ 0 0 2 ¼» ¬« z ¼» ¬« z ¼»
2x + y = 2x Ÿ y = 0. x and z can be arbitrarily chosen.

48
 Sathyabama University

ª1 º
Hence if x = 1 and z = 0 then «« 0 »» is an eigen vector.
«¬ 0 »¼

ª0º
If x = 0 and z = 1 then «« 0 »» is another vector.
«¬1 »¼

ª1 º ª0º
«
Thus both «0 » and «« 0 »» are linearly independent vectors for eigen value Ȝ = 2
»
«¬0 »¼ «¬1 »¼

ª2 1 0º
Example 2: Let A = «« 0 2 1 »» .
«¬ 0 0 2 »¼

ª2 − λ 1 0 º
The characteristic equation of A is «« 0 2−λ 1 »» = 0
«¬ 0 0 2 − λ »¼
Ÿ ( 2 − λ )( 2 − λ )( 2 − λ ) = 0
Ÿ λ = 2, 2, 2 are the eigen values
ª xº
If X = «« y »» is an eigen vector corresponding to Ȝ = 2, then AX = 2X.
«¬ z »¼

ª2 1 0 º ª x º ªxº
« »« » « »
«0 2 1 » « y » = 2 « y »
«¬ 0 0 2 »¼ «¬ z »¼ «¬ z »¼
2x + y = 2x, 2y + z = 2y
Ÿ y = 0, z = 0 and x can be chosen arbitrarily.
ª1 º
Hence the eigen vector for λ = 2 is ««0 »» .
«¬0 »¼

Here the eigen vector corresponding to the eigen value λ = 2 is linearly

dependent.

49
Sathyabama University

4. The eigen vectors corresponding to distinct eigen values of a real

symmetric matrix are orthogonal.
Proof: Let X1, X2 be the eigen vectors corresponding to the distinct
eigen values λ1 and λ2 of a real symmetric matrix.

Both λ1 and λ2 are real numbers

AX 1 = λ1 X 1 (1)

AX 2 = λ2 X 2 (2)

Pre multiplying (1) by X2T we get,

X 2T AX 1 = λ1 X 2T X 1 (3)

Pre multiplying (2) by X1T we get

X 1T AX 2 = λ2 X 1T X 2 (4)

Taking transpose of (3)

X 1T AT ( X 2T )T = λ1 X 1T ( X 2T )T

X 1T AX 2 = λ1 X 1T X 2 (5)

(' A is a symmetric matrix, AT=A)

From (4) & (5) we get,
λ1 X 1T X 2 = λ2 X 1T X 2

(λ1 − λ2 ) X 1T X 2 = 0
Since λ1 ≠ λ2 , X 1T X 2 = 0
Ÿ X 1 and X 2 are orthogonal

ª1 2 º
Example 1: Find the eigen values of « »
¬ 0 3¼
Solution
ª1 2 º
Let A = « »
¬ 0 3¼
For a triangular matrix, the eigen values are its leading diagonal elements.
∴ Eigen values of A are 1 and 3.

50
 Sathyabama University

ª3 1 4 º
Example 2: Find the sum and product of the eigen values of «« 0 2 6 »»
«¬ 0 0 5 »¼
Solution
ª3 1 4 º
Let A = ««0 2 6 »»
«¬0 0 5 »¼
Since A is a triangular matrix the eigen values are its leading diagonal
elements.
∴ 3, 2, 5 are the eigen values of A.
∴ Sum = 3+2+5=10
Product = 3×2×5=30
ª4 1 º
Example 3: Find the eigen values of 3A2, A = « »
¬3 2 ¼
Solution
The characteristic equation of A is given by λ 2 − C1λ + C2 = 0

where C1 = 4 + 2 = 6
C2 = A = 5

∴ Characteristic equation of A is
λ 2 − 6λ + 5 = 0.
(λ − 1) (λ − 5) = 0

Ÿ λ = 1,5 are the eigen values of A,

∴ 12, 52 are the eigen values of A2,

∴ Eigen values of 3A2 are 3(1)2, 3(5)2
i.e., 3, 75 are the eigen values of 3A2.

ª3 5 3 º
Example 4: Find the eigen values of Adj A if A = «« 0 4 6 »»
«¬ 0 0 1 »¼
Solution
Since A is a triangular matrix, its eigen values are its leading diagonal
elements.

51
Sathyabama University

∴ The eigen values of A are 3, 4, 1

1 1
The eigen values of A-1 are , ,1
3 4
Adj A
But A−1 =
A
A =12
∴ Adj A = A−1 A

∴ Eigen values of Adj A are got by multiplying the eigen values of A-1
by 12.
∴ Eigen values of Adj A are 4, 3, 12.

ª 2 2 1º
Example 5: Two eigen values of the matrix A = ««1 3 1»» are equal to
«¬1 2 2 »¼
1 each. Find the eigen values of A-1.

Solution
Let the eigen values of A be 1, 1, λ . Sum of eigen values = Trace of A.
i.e., 1 + 1 + λ = 2 + 3 + 2
2+λ = 7
∴λ = 5

Ÿ 1, 1, 5 are the eigen values of A.

1
Ÿ 1, 1, are the eigen values of A-1.
5
Example 6: Two eigen values of a 3×3 matrix A are 2, 8 and A = 32. Find
the eigen values of A-1

Solution
Let 2 , 8, λ be the eigen values the matrix A.
Given A = 32

Then by property (3), product of eigen values = A

52
 Sathyabama University

i.e., 2 × 8 × λ = 32
i.e., 16 λ = 32
∴ λ = 2 is the third eigen value.
If 2, 8, 2 are the eigen values of matrix A, then
1 1 1
, , are the eigen values of A−1
2 8 2
Example 7: Find the eigen values and eigen vectors of the matrix
ª 2 2 −7 º
« »
«2 1 2 »
«¬ 0 1 −3»¼
Solution
ª 2 2 −7 º
Let A = «« 2 1 2 »» be the given matrix.
«¬ 0 1 −3¼»
The characteristic equation of A is given by A − λ I = 0
The general form of a characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 2 + 1 − 3 = 0
1 2 2 −7 2 2
C2 = + + = −13
1 −3 0 −3 2 1
C3 = A = −12

∴ The characteristic equation is λ 3 − 13λ + 12 = 0

1 0 −13 12
1
0 1 1 −12
1 1 −12 0
2
∴ λ = 1 is a root and λ + λ − 12 = 0
(λ + 4) (λ − 3) = 0
λ = 3, −4
∴λ = 1,3, −4 are the eigen values of A.

The eigen vector corresponding to the eigen value Ȝ is given by

ª x1 º
( A − λ I ) X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

53
Sathyabama University

ª2 − λ 2 −7 º ª x1 º
«
i e., « 2 1− λ 2 »» «x » 0
« 2»
«¬ 0 1 −3 − λ »¼ «¬ x3 »¼

(2 − λ ) x1 + 2 x2 − 7 x3 = 0 ½
°
2 x1 + (1 − λ ) x2 + 2 x3 = 0 ¾ (1)
0 x1 + x2 + (−3 − λ ) x3 = 0 °¿

Case 1: If Ȝ =1, then (1) becomes

x1 + 2 x2 − 7 x3 = 0
2 x1 + 0 x2 + 2 x3 = 0
0 x1 + x2 − 4 x3 = 0

Using the rule of cross multiplication for the first two distinct equations
we get,
x1 − x2 x3
= =
2 −7 1 −7 1 2
0 2 2 2 2 0
x1 − x2 x3
= =
4 16 −4
x −x x
i e., 1 = 2 = 3
1 4 −1

ª1º
∴ The eigen vector corresponding to Ȝ =1 is X 1 = «« −4 »»
«¬ −1»¼
Case 2: If Ȝ = 3, then (1) becomes
− x1 + 2 x2 − 7 x3 = 0
2 x1 − 2 x2 + 2 x3 = 0
0 x1 + x2 − 6 x3 = 0

Using the rule of cross multiplication for the second and third distinct
equations we get,
x1 − x2 x3
= =
−2 2 2 2 2 −2
1 −6 0 −6 0 1

54
 Sathyabama University

x1 x2 x3
= =
10 12 2
x x x
i e., 2 = 2 = 3
5 6 1

§ 5·
¨ ¸
∴ The eigen vector corresponding to Ȝ = 3 is X 2 = ¨ 6 ¸
¨1¸
© ¹

Case 3: If Ȝ = -4, then (1) becomes

6 x1 + 2 x2 − 7 x3 = 0
2 x1 + 5 x2 + 2 x3 = 0
0 x1 + x2 + x3 = 0

Using the rule of cross multiplication for the second and third distinct
equations we get,
x1 − x2 x
= = 3
5 2 2 2 2 5
1 1 0 1 0 1
x1 − x2 x3
= =
3 2 2

§ 3·
¨ ¸
∴ The eigen vector corresponding to Ȝ = – 4 is X 3 = ¨ −2 ¸
¨ 2¸
© ¹
The eigen values of A are 1, 3, -4 and its corresponding eigen vectors are
§ 1 · §5· § 3·
¨ ¸ ¨ ¸ ¨ ¸
¨ −4 ¸ , ¨ 6 ¸ and ¨ −2 ¸
¨ −1 ¸ ¨ 1 ¸ ¨ 2¸
© ¹ © ¹ © ¹
Example 8: Find the eigen values and eigen vectors of the
ª5 0 1 º
matrix «« 0 −2 0 »»
«¬1 0 5 ¼»

55
Sathyabama University

Solution
ª5 0 1 º
Let A= «« 0 −2 0 »» be the given matrix
«¬1 0 5 »¼

The characteristic equation of A is given by A − λ I = 0

The general form of characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 5 − 2 + 5 = 8
−2 0 5 1 5 0
C2 = + + =4
0 5 1 5 0 −2

C3 = A = −48

∴ The characteristic equation is λ 3 − 8λ 2 + 4λ + 48 = 0

1 −8 4 48
4
0 4 −16 −48
1 −4 −12 0

λ = 4 is a root and λ 2 − 4λ − 12 = 0
( λ + 2 )( λ − 6 ) = 0
λ = −2, 6
∴ λ = 4, −2, 6 are the eigen values of A.
The eigen vector corresponding to the eigen value λ is given by
ª x1 º
( A − λ I ) X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª5 − λ 0 1 º ª x1 º
i.e., «« 0 −2 − λ 0 »» «« x2 »» = 0
¬« 1 0 5 − λ »¼ «¬ x3 »¼

( 5 − λ ) x1 + 0 x2 + x3 = 0½
°
0 x1 + ( −2 − λ ) x2 + 0 x3 = 0 ¾ (1)
°
x1 + 0 x2 + ( 5 − λ ) x3 = 0 ¿

56
 Sathyabama University

Case 1: λ = −2 , then (1) becomes

7 x1 + 0 x2 + x3 = 0
0 x1 + 0 x2 + 0 x3 = 0
x1 + 0 x2 + 7 x3 = 0

Using the rule of cross multiplication for the first and third distinct
x1 − x2 x
equations we get, = = 3
0 1 7 1 7 0
0 7 1 7 1 0

x1 − x2 x3
= =
0 48 0

ª0 º
∴ The eigen vector corresponding to λ = −2 is X 1 = ««1 »»
«¬0 »¼

Case 2: if λ = 4 , then (1) becomes

x1 + 0 x2 + x3 = 0
0 x1 − 6 x2 + 0 x3 = 0
x1 + 0 x2 + 7 x3 = 0

Using the rule of cross multiplication for the first and second distinct
equations, we get
x1 − x2 x3
= =
0 1 1 1 1 0
−6 0 0 0 0 −6

x1 − x2 x3
= =
6 0 −6
x1 − x2 x3
i.e., = =
1 0 −1

ª1º
∴ The eigen vector corresponding to λ = 4 is X 2 = «« 0 »»
«¬ −1»¼

57
Sathyabama University

Case 3: if λ = 6 , then (1) becomes

− x1 + 0 x2 + x3 = 0
0 x1 − 8 x2 + 0 x3 = 0
x1 + 0 x2 − x3 = 0

Using the rule of cross multiplication for the first and second distinct
equations, we get
x1 − x2 x3
= =
0 1 −1 1 −1 0
−8 0 0 0 0 −8

x1 − x2 x3
= =
8 0 8
x1 − x2 x3
i.e., = =
1 0 1

ª1 º
∴ The eigen vector corresponding to λ = 6 is X 3 = «« 0 »»
«¬1 »¼

The eigen values of A are -2, 4, 6 and its corresponding eigen vectors
§0· § 1 · §1·
¨ ¸ ¨ ¸ ¨ ¸
are ¨ 1 ¸ , ¨ 0 ¸ and ¨ 0 ¸
¨ 0 ¸ ¨ −1 ¸ ¨1¸
© ¹ © ¹ © ¹
Example 9: Find the eigen value and eigen vector of the matrix
§ 6 −2 2 ·
¨ ¸
¨ −2 3 −1¸
¨ ¸
© 2 −1 3 ¹
Solution
§ 6 −2 2 ·
¨ ¸
Let A = ¨ −2 3 −1¸ .
¨ 2 −1 3 ¸
© ¹

The Characteristic equation of A is given by A − λ I = 0

The general form of a characteristic equation is

λ 3 − C1λ 2 + C2 λ − C3 = 0 ,

58
 Sathyabama University

where C1 = 6 + 3 + 3 = 12

3 −1 6 2 6 −2
C2 = + + = 36
−1 3 2 3 −2 3

C3 = A = 32

∴ The characteristic equation is λ 3 − 12λ 2 + 36λ − 32 = 0

1 −12 36 −32
2
0 2 −20 32
1 − 10 16 0

∴ λ = 2 is a root and λ 2 − 10λ + 16 = 0

( λ − 2 )( λ − 8) = 0
λ = 2,8
∴ λ = 2, 2,8 are the eigen values of A.
The eigen vector corresponding to the eigen value λ is given by
ª x1 º
( A − λ I ) X = 0 where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª 6 − λ −2 2 º ª x1 º
i.e., « −2 3 − λ −1 »» «« x2 »» = 0
«
«¬ 2 −1 3 − λ »¼ «¬ x3 »¼

( 6 − λ ) x1 − 2 x2 + 2 x3 = 0 ½
°
−2 x1 + ( 3 − λ ) x2 − x3 = 0 ¾ (1)
°
2 x1 − x2 + ( 3 − λ ) x3 = 0 ¿

Case 1: If λ =8, then (1) becomes

−2 x1 − 2 x2 + 2 x3 = 0

−2 x1 − 5 x2 − x3 = 0

2 x1 − x2 − 5 x3 = 0

59
Sathyabama University

Using rule of cross multiplication for the first and second distinct equations,
we get
x1 − x2 x3
= =
−2 2 −2 2 −2 −2
−5 − 1 −2 −1 −2 −5

x1 − x2 x3
= =
12 6 6
x1 x2 x3
i.e., = =
2 −1 1
§2·
¨ ¸
∴ The Eigen vector corresponding to the eigen value λ = 8 is X 1 = ¨ −1¸
¨1¸
© ¹
Case 2: If λ = 2 , then (1) becomes
4 x1 − 2 x2 + 2 x3 = 0
−2 x1 + x2 − x3 = 0
2 x1 − x2 + x3 =0

All the three equations are equivalent to a single equation.

2 x1 − x2 + x3 =0

x2 x3
Put x1 = 0 Then x2 = x3 Ÿ =
1 1
§0·
¨ ¸
∴ The eigen vector corresponding to the eigen value λ = 2 is X 2 = ¨ 1 ¸
¨1¸
© ¹
Put x2 = 0, then 2x1 = -x3
x1 x3
Ÿ =
1 −2
ª1º
∴ The eigen vector corresponding to the eigen value λ = 2 is X 3 = «« 0 »»
«¬ −2 »¼

60
 Sathyabama University

∴ The eigen values are 2, 2, 8 and its corresponding linearly independent

ª0 º ª 1 º ª 2 º
eigen vectors are ««1 »» , «« 0 »» & «« −1»»
«¬1 »¼ «¬ −2 »¼ «¬ 1 »¼

Example 10: Find the eigen values and eigen vectors of the matrix
ª3 10 5º
A = «« −2 −3 −4 »»
«¬ 3 5 7 »¼

Solution

ª 3 10 5 º
Let A = «« −2 −3 −4 »» be the given matrix
«¬ 3 5 7 »¼

The characteristic equation of A is given by A − AI = 0

The general form of the Characteristic equation is

λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 3 − 3 + 7 = 7

−3 −4 3 5 3 10
C2 = + + = 16
5 7 3 7 −2 −3
C3 = A = 12

∴ The characteristic equation is λ 3 − 7λ 2 + 16λ − 12 = 0

1 − 7 16 − 12
2
0 2 − 10 12
1 −5 6 0

λ = 2 is a root and λ 2 − 5 λ + 6 = 0
(λ − 2) (λ − 3) = 0
λ = 2,3
∴ λ = 2, 2,3are the eigen values of A

61
Sathyabama University

The eigen vector corresponding to the eigen value λ is given by

ª x1 º
( A − λ I ) X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª3 − λ 10 5 º ª x1 º
i.e., «« −2 −3 − λ −4 »» «x » = 0
« 2»
«¬ 3 5 7 − λ »¼ «¬ x3 »¼

i.e., (3 − λ ) x1 + 10 x2 + 5 x3 = 0 ½
°
− 2 x1 + (−3 − λ ) x2 − 4 x3 = 0 ¾ (1)
3 x1 + 5 x2 + (7 − λ ) x3 = 0 ¿°

Case (1): If Ȝ = 3, then equation (1) becomes,

0 x1 + 10 x2 + 5 x3 = 0
−2 x1 − 6 x2 − 4 x3 = 0
3 x1 + 5 x2 + 4 x3 = 0

Using the rule of cross multiplication for the first and second distinct
equations we get,
x1 − x2 x3
= =
10 5 0 5 0 10
−6 −4 −2 −4 −2 −6
x1 − x2 x3
= =
−10 10 20
x x2 x3
i.e., = =
1 1 −2

ª1º
∴ The eigen vector corresponding to the eigen value λ = 3 is , X 1 = «« 1 »»
«¬ −2 »¼

Case (2): If Ȝ = 2, then equation (1) becomes,

x1 + 10 x2 + 5 x3 = 0 ½
°
−2 x1 − 5 x2 − 4 x3 = 0 ¾ (1)
3 x1 + 5 x2 + 5 x3 = 0 °¿

62
 Sathyabama University

Using the rule of cross multiplication for the first and second distinct
equations, we get
x1 − x2 x3
= =
10 5 1 5 1 10
−5 −4 −2 −4 −2 −5
x1 − x2 x3
= =
−15 6 15
x1 x2 x3
i.e., = =
5 2 −5
ª5º
∴ The eigen vector corresponding to λ = 2 is X 2 = «« 2 »»
«¬ −5»¼

Here Ȝ = 2 is a repeated eigen value and the corresponding eigen vector

ª5º
« 2 » is the only possible vector. Hence the eigen vector corresponding to the
« »
«¬ −5»¼
eigen value Ȝ = 2 is linearly independent
∴ The eigen values of A are 2, 2, 3 and its corresponding eigen vectors
ª5 º ª1º
are «« 2 »» & «« 1 »»
«¬ −5 »¼ «¬ −2 »¼

Example 11: Find the eigen values and eigen vectors of the matrix
§ 6 −6 5 ·
¨ ¸
¨ 14 −13 10 ¸
¨ 7 −6 4 ¸
© ¹
Solution

ª6 − 6 5 º
Let A = ««14 − 13 10 »» be the given matrix.
«¬7 − 6 4 »¼

The characteristic equation of A is given by A − λ I = 0 .

The general form of a characteristic equation is

λ 3 − C1λ 2 + C2 λ − C3 = 0

63
Sathyabama University

where C1 = 6-13+4 = –3
−13 10 6 5 6 −6
C2 = + + =3
−6 4 7 4 14 −13

C3 = A = −1

∴ The characteristic equation is Ȝ3+3 Ȝ2+3 Ȝ+1 = 0.

i.e., (Ȝ+1)3 = 0
∴ The eigen values are Ȝ= –1, –1, –1
The eigen vector corresponding to the eigen value Ȝ is given by [A- ȜI]
ª x1 º
X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª6 − λ −6 5 º ª x1 º
i.e., «« 14 −13 − λ 10 »» «x » = 0
« 2»
«¬ 7 −6 4 − λ »¼ «¬ x3 »¼

(6 − λ ) x1 − 6 x2 + 5 x3 = 0 ½
°
14 x1 + (−13 − λ ) x2 + 10 x3 = 0 ¾ (1)
7 x1 − 6 x2 + (4 − λ ) x3 = 0 °
¿
Put λ = − 1 . Then equation (1) reduces to

7 x1 − 6 x2 + 5 x3 = 0
14 x1 − 12 x2 + 10 x3 = 0
7 x1 − 6 x2 + 5 x3 = 0

All the three equations are equivalent to a single equation

7 x1 − 6 x2 + 5 x3 = 0 (2)

In (2), put x1 = 0 then

6 x2 = 5 x3
x2 x3
Ÿ =
5 6
ª0 º
The eigen vector corresponding to λ = −1 is X 1 = ««5 »»
«¬6 »¼

64
 Sathyabama University

x1 x3
In (2), put x2 = 0 then 7 x1 = −5 x3 Ÿ =
5 −7
ª5 º
Another eigen vector corresponding to λ = −1 is X 2 = ««0 »»
«¬ −7 »¼
x x
In (2), put x3 = 0 then 7 x1 = 6 x2 Ÿ 1 = 2
6 7
ª6 º
Another eigen vector corresponding to λ = −1 is X 3 = «« 7 »»
«¬ 0 »¼
Hence the eigen vectors corresponding to the eigen value λ = −1 (three
ª 0 º ª5 º ª6 º
times) are ««5 »» , ««0 »» & ««7 »» which are linearly independent.
«¬ 6 »¼ «¬ −7 »¼ «¬0 »¼

Example 12: Find the eigen values and eigen vectors of the matrix
ª2 1 0º
« »
«0 2 1 »
«¬ 0 0 2 »¼

Solution
ª2 1 0º
Let A = «« 0 2 1 »» be the given matrix. Since A is a triangular matrix,
«¬ 0 0 2 »¼
the eigen values of A are its leading diagonal elements.
∴ Eigen values of A are λ = 2, 2, 2.
The eigen vector corresponding to the eigen value λ is given by
ª x1 º
( A − λ I ) X = 0 where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª2 − λ 1 0 º ª x1 º
« « »
i.e., « 0 2−λ 1 »» « x 2 » = 0
«¬ 0 0 2 − λ »¼ «¬ x3 »¼

65
Sathyabama University

(2 − λ ) x1 + x2 + 0.x3 = 0 ½
°
i.e., 0.x1 + (2 − λ ) x2 + x3 = 0 ¾ → (1)
0.x1 + 0.x2 + (2 − λ ) x3 = 0 °¿

Put λ = 2 in (1). Then we get

0 x1 + x2 + 0 x3 = 0 Ÿ x2 = 0
0 x1 + 0 x2 + 1x3 = 0 Ÿ x3 = 0
0 x1 + 0 x2 + 0 x3 = 0

Let x1 be any arbitrary value say x1 = 1

∴ The only Eigen vector corresponding to the repeated eigen value

ª1 º
λ = 2 is X = ««0 »»
«¬ 0 ¼»
EXERCISE

Part A

ª3 2 º
1. Obtain the eigen value of A3 for A= « »
¬1 2 ¼
ª 3 −1 1 º
2. If two eigen values of the matrix A= «« −1 5 −1»» are 3 and 6, find
«¬ 1 −1 3 »¼
-1
the eigen values of A
ª −1 1 1 º
3. Find the sum and product of the eigen values of «« 1 −1 1 »»
«¬ 1 1 −1»¼
4. If λ1 , λ2 ,...., λn are the latent roots of a matrix A, show that the latent
roots of A3 are λ13 , λ23 ,...., λn3 .
ª 8 −6 2 º
5. If 3 and 15 are the eigen values of the matrix A= «« −6 7 −4 »» ,
«¬ 2 −4 3 »¼
find the eigen values of A − 5 I and A2 .

66
 Sathyabama University

ª4 6 6º
6. If two eigen values of the matrix A= «« 1 3 2 »» are equal and
«¬ −1 −5 −2 »¼
they are double the third, find the eigen values of A2.
7. State any one property of the eigen value of a matrix and verify it for
ª1 1 º
the matrix « »
¬3 2 ¼

ª 7 2 −2 º
8. Find the product of the eigen values of «« −6 −1 2 »»
«¬ 6 2 −1»¼

9. If 1, 3, -1 are the eigen values of a 3×3 matrix A, what are the eigen
values of A3 − 4 I . Is A invertible?
ªa 4º
10. Find the constants a and b such that the matrix « » has 3 and -2
¬1 b ¼
as its eigen values.
ª1 2 º
11. Find the eigen values of « »
¬0 3¼

ª5 4 º
12. Find the eigen values and eigen vectors of A = « »
¬1 2 ¼

Part B

ª1 1 3º
13. Find the eigen values and eigen vectors of the matrix ««1 5 1»»
«¬3 1 1»¼

ª 8 −6 2 º
14. Find the eigen values and eigen vectors of the matrix «« −6 7 −4 »»
«¬ 2 −4 3 »¼

ª11 −4 −7 º
15. Find the eigen values and eigen vectors of the matrix «« 7 −2 −5 »»
«¬10 −4 −6 »¼

67
Sathyabama University

ª 2 −2 3 º
16. Find the eigen values and eigen vectors of the matrix «« 1 1 1 »»
«¬ 1 3 −1»¼

ª −2 5 4 º
17. Find the eigen values and eigen vectors of the matrix «« 5 7 5 »»
«¬ 4 5 −2 »¼

ª 2 −1 1 º
18. Find the eigen values and eigen vectors of the matrix «« −1 2 −1»»
«¬ 1 −1 2 »¼

ª1 1 0 º
19. Find the eigen values and eigen vectors of the matrix «« 0 1 0 »»
«¬ 0 0 1 »¼

ª 2 −2 2 º
20. Find the eigen values and eigen vectors of the matrix «« 1 1 1 »»
«¬ 1 3 −1»¼

ª 2 −1 1 º
21. Find the eigen values and eigen vector of the matrix «« −1 2 −1»»
«¬ 1 −1 2 »¼

ª2 0 1º
22. Find the eigen values and eigen vector of the matrix «« 0 3 0 »»
«¬ 1 0 2 »¼

ª2 2 1º
23. Find the eigen value and eigen vectors of the matrix ««1 3 1 »»
«¬1 2 2 »¼

ª 7 −2 0 º
24. Find the eigen values and eigen vectors of the matrix «« −2 6 −2 »»
«¬ 0 −2 5 »¼

ª1 2 3º
25. Find the eigen values and eigen vectors of the matrix «« 2 4 6 »»
«¬ 3 6 9 »¼

68
 Sathyabama University

ANSWERS

Part-A

1. 1, 64
1 1 1
2. , ,
2 3 6
3. 3, 4
5. –5, -2, 10; 0, 9, 225
6. 1, 2, 2
8 3
9. A is invertible; -3, 23, -5
10. If a = 2 then b= -1
If a = -1 then b = 2
11 1, 3
ª4º ª −1º
12. 1, 6; «1 » and « 1 »
¬ ¼ ¬ ¼

Part-B

ª −1º ª 1 º ª 1 º
13. –2, 3, 6; «« 0 »» , «« −1»» & «« 2 »»
«¬ 1 »¼ «¬ 1 »¼ «¬ 1 »¼

ª1 º ª 2 º ª 2 º
14. 0, 3, 5; «« 2 »» , «« 1 »» & «« −2 »»
«¬ 2 ¼» ¬« −2 ¼» ¬« 1 ¼»

ª1º ª 1 º ª 2 º
15. 0, 1, 2; ««1»» , «« −1»» & «« 1 »»
«¬1»¼ «¬ 2 »¼ «¬ 2 »¼

69
Sathyabama University

ª11º ª −1º ª1º

16. –2, 1, 3; «« 1 »» , «« 1 »» & ««1»»
«¬14 »¼ «¬ 1 »¼ «¬1»¼

ª 1 º ª 1 º ª1 º
17. –3, -6, 12 ; «« −1»» , «« 0 »» & «« 2 »»
«¬ 1 »¼ «¬ −1»¼ «¬ 1 »¼

ª1 º ª −1º ª 1 º
18. 1, 1, 4; ««1 »» , «« 1 »» & «« −1»»
«¬ 0 ¼» ¬« 2 ¼» ¬« 1 ¼»

ª1 º
19. 1, 1, 1; «« 0 »»
«¬1 »¼

ª −4 º ª0 º
20. –2, 2, 2; «« −1»» & ««1 »»
«¬ 7 ¼» ¬«1 ¼»

ª 1 º ª0º ª 1 º
21. 1, 1, 4; «« −1»» , ««1 »» & «« 0 »»
«¬ 1 ¼» ¬«1 ¼» ¬« −1¼»

ª −1º ª1 º ª 1 º
22. 1, 3, 3; «« 0 »» , ««0 »» & «« 2 »»
«¬ 1 »¼ «¬1 »¼ «¬ 1 »¼

ª 0 º ª 1 º ª1º
23. 1, 1, 5; «« 1 »» , «« 0 »» & ««1»»
«¬ −2 »¼ «¬ −1»¼ «¬1»¼

ª 1 º ª −2 º ª 2 º
24. 3, 6, 9; «« 2 »» , «« −1»» & «« −2 »»
«¬ 2 ¼» ¬« 2 ¼» ¬« 1 ¼»

ª 0 º ª −3º ª 1 º
25. 0, 0, 14; «« −3»» , «« 0 »» & «« 2 »»
«¬ 2 »¼ «¬ 1 »¼ «¬ 3 »¼

70
 Sathyabama University

DIAGONALISATION OF A MATRIX

Let A be a square matrix of order n. The process of reducing the matrix A to a

diagonal matrix D is called diagnalisation of a matrix A.
Modal Matrix: The matrix P, P = ª¬ X 1, X 2, …… X n º¼ formed by the
linearly independent eigen vectors of a matrix A is called the modal matrix
and is denoted by M.
Spectral Matrix: The diagonal matrix D having the eigen values of A as
the diagonal elements is called Spectral Matrix.
Similar Matrices: Two matrices A and B are said to be similar, if there
exists a non singular matrix M such that
B = M −1 AM (1)
Thus transforming a matrix A into a matrix B by (1) is called a similarly
transformation.

Diagonalisation by Similarity Transformation:

Let A be a square matrix. Transforming matrix A to a diagonal matrix D by a

transformation D = M −1 AM is called diagonalisation by similarity
transformation.

Remark:

1) For diagonalisation by similarity transformation, matrix A

should have linearly independent eigen vectors.

2) The matrix M which diagonalise A is called the modal matrix of

A.

3) D is called the spectral Matrix of A.

4) The eigen values of D are its diagonal elements.

5) The matrices A and D are similar and hence A and D have same
eigen values.

Calculation of powers of a matrix:

Consider D = M −1 AM
Pre multiply by M and post multiply by M-1 on both sides, we get,
MDM −1 = M ( M −1 A M ) M −1

71
Sathyabama University

i.e., MDM −1 = A
Now A2 = A ⋅ A = ( MDM −1 )( MDM −1 ) = MD 2 M −1

Thus A2 = MD 2 M −1
In general An = MD n M −1

ª λ1n 0 0 0 º!
« »
«0 λ2 n
0 ! 0 »
where D = « #
n
% »
« »
« # % »
«0 n»
0 ! λn ¼
¬ 0

Working rule for diagonalisation by similarity transformation

Step1: Given a square matrix A, find the characteristic equation
Step2: Find the eigen values
Step3: Find the corresponding linearly independent eigen vectors
Step4: Form a Modal Matrix M with columns as eigen vectors of A.
Step5: Find M–1
Step6: Calculate M −1 AM = D

§1 6 1·
¨ ¸
Example 1: Diagonalise the matrix ¨ 1 2 0 ¸ by similarity transformation.
¨ 0 0 3¸
© ¹
Solution
ª1 6 1 º
Let A = ««1 2 0 »»
«¬ 0 0 3 »¼

The characteristic equation of A is given by A − λ I = 0

The general form of a characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 1 + 2 + 3 = 6

2 0 1 1 1 6
C2 = + + = 5
0 3 0 3 1 2

1 6 1
C3 = A = 1 2 0 = 1( 6 − 0 ) – 6 ( 3 − 0 ) + 1( 0 − 0 ) = −12
0 0 3

72
 Sathyabama University

∴The characteristic equation of A is λ3 – 6λ2 + 5 λ + 12 = 0

1 −6 5 12
−1
0 −1 7 −12
1 −7 12 0

λ = -1 is a root and λ 2 − 7λ + 12 = 0
(λ − 4) (λ − 3) = 0

∴λ = -1, 3, 4 are the eigen values of A.

The eigen vector corresponding to the eigen value λ is given by
§ x1 ·
¨ ¸
( A − λ I ) X = 0, where X = ¨ x2 ¸ is a column vector.
¨x ¸
© 3¹

ª1 − λ 6 1 º § x1 ·
« ¨ ¸
« 1 2−λ 0 »» ¨ x2 ¸ = 0
«¬ 0 3 − λ ¼» ¨ ¸
0 © x3 ¹

(1 − λ ) x1 + 6 x2 + x3 = 0 ½
°
x1 + (2 − λ ) x2 + 0 x3 = 0 ¾ (1)
0 x1 + 0 x2 + (3 − λ ) x3 = 0 ¿°

Case 1: If λ = −1 , then (1) becomes

2 x1 + 6 x2 + x3 = 0
x1 + 3x2 + 0 x3 = 0
0 x1 + 0 x2 + 4 x3 = 0

Using the rule of cross multiplication for the first two distinct equations we
get
x1 − x2 x3
= =
6 1 2 1 2 6
3 0 1 0 1 3

x1 x x
= 2 = 3
−3 1 0

73
Sathyabama University

§3·
∴ The eigen vector corresponding to λ = –1 is X 1 = ¨¨ −1¸¸
¨0¸
© ¹
Case 2: If λ = 3, then (1) becomes
−2 x1 + 6 x2 + x3 = 0
x1 − x2 + 0 x3 = 0
0 x1 + 0 x2 + 0 x3 = 0

Using the rule of cross multiplication for the first two distinct equations
we get
x1 − x2 x3
= =
6 1 −2 1 −2 6
−1 0 1 0 1 −1

x1 x2 x3
i.e., = =
1 1 −4

§1·
¨ ¸
∴ The eigen vector corresponding to λ = 3 is X 2 = ¨ 1 ¸
¨ −4 ¸
© ¹
Case 3: If λ = 4, then (1) becomes
−3x1 + 6 x2 + x3 = 0
x1 − 2 x2 + 0 x3 = 0
0 x1 + 0 x2 − x3 = 0

Using the rule of cross multiplication for the first two distinct equations,
we get
x1 − x2 x3
= =
6 1 −3 1 −3 6
−2 0 1 0 1 −2

x1 x2 x3
= =
2 1 0

§ 2·
¨ ¸
∴ The eigen vector corresponding to λ = 4 is X 3 = ¨ 1 ¸
¨0¸
© ¹

74
 Sathyabama University

ª 3 1 2º
Now M = «« −1 1 1 »»
«¬ 0 −4 0 »¼

To find M-1:
3 1 2
M = −1 1 1 = 3 ( 4 ) – 1( 0 ) + 2 ( 4 ) = 20
0 −4 0

1 1
Cofactor of 3 = =4
−4 0

1 2
Cofactor of í1 = − = í8
−4 0

1 2
Cofactor of 0 = = í1
1 1

−1 1
Cofactor of 1 = − =0
0 0

3 2
Cofactor of 1 = =0
0 0
3 2
Cofactor of -4 = − = –5
−1 1
−1 1
Cofactor of 2 = =4
0 −4
3 1
Cofactor of 1 = − = 12
0 −4
3 1
Cofactor of 0 = =4
−1 1
Adj M
M −1 =
M

ª 4 −8 −1º
1 «
M −1
= « 0 0 −5»»
20
«¬ 4 12 4 »¼

75
Sathyabama University

Consider D = M −1 AM
ª 4 −8 −1º ª1 6 1 º ª 3 1 2 º
1 «
= « 0 0 −5»» ««1 2 0 »» «« −1 1 1 »»
20
¬« 4 12 4 »¼ «¬ 0 0 3 »¼ «¬ 0 −4 0 »¼
ª 4 −8 −1º ª −3 3 8 º
1 «
= « 0 0 −5»» «« 1 3 4 »»
20
«¬ 4 12 4 »¼ «¬ 0 −12 0 ¼»

ª −20 0 0 º
1 «
= « 0 60 0 »»
20
«¬ 0 0 80 »¼

ª −1 0 0 º
= «« 0 3 0 »»
«¬ 0 0 4 »¼

Example 2:

ª2 2 1º
Diagonalise the matrix ««1 3 1 »» by similarity transformation.
«¬1 2 2 »¼

Solution:

ª2 2 1º
Let A = ««1 3 1 »»
«¬1 2 2 »¼

The characteristic equation of A is given by A-λ I = 0

The general form of a characteristic equation is

3 2
λ − C1λ + C 2 λ − C3 = 0
where C1 = 2 + 3 + 2 = 7
3 1 2 1 2 2
C2 = + + = 11
2 2 1 2 1 3

76
 Sathyabama University

2 2 1
C3 = A = 1 3 1 = 2 ( 6 − 2 ) – 2 ( 2 − 1) + 1 ( 2 − 3) = 5
1 2 2

∴ The Characteristic equation of A is λ 3 − 7λ 2 + 11λ − 5 = 0

1 −7 11 −5
1
0 1 −6 5
1 1 −6 5

λ = 1 is a root and λ 2 − 6λ + 5 = 0
( λ − 5)( λ − 1) = 0
∴ λ = 1, 1, 5 are the eigen values of A.
The eigen vector corresponding to the eigen value Ȝ is given by (A –ȜI)X = 0,
§ x1 ·
¨ ¸
where X = ¨ x2 ¸ is a column vector.
¨x ¸
© 3¹

ª2 − λ 2 1 º ª x1 º
« 1
« 3−λ 1 »» «« x2 »» = 0
¬« 1 2 2 − λ ¼» «¬ x3 »¼

( 2 − λ ) x1 + 2 x2 + x3 = 0 ½
°
x1 + ( 3 − λ ) x2 + x3 = 0 ¾ (1)
°
x1 + 2 x2 + ( 2 − λ ) x3 = 0 ¿

Case 1: If Ȝ = 5, then (1) becomes

−3 x1 + 2 x2 + x3 = 0
x1 − 2 x2 + x3 = 0
x1 + 2 x2 − 3 x3 = 0

Using the rule of cross multiplication for the first two distinct equations, we
get
x1 − x2 x3
= =
2 1 −3 1 −3 2
−2 1 1 1 1 −2

77
Sathyabama University

x1 x x
= 2 = 3
4 4 4

§1·
¨ ¸
The eigen vector corresponding to λ = 5 is X1 = ¨ 1 ¸
¨1¸
© ¹
Case 2: If Ȝ =1 (twice), then (1) becomes
x1 + 2 x2 + x3 = 0
x1 + 2 x2 + x3 = 0
x1 + 2 x2 + x3 = 0

All the three equations are the same.

Put x3 = 0. Then x1 + 2 x2 = 0

Ÿ x1 = − 2 x2

x1 x2
i.e., =
−2 1
§ −2 ·
∴ The eigen vector corresponding to Ȝ = 1 is X 2 = ¨¨1 ¸¸
¨0 ¸
© ¹

Now put x2 = 0. Then x1 + x3 = 0.

x1 x3
Ÿ x1 = − x3 i.e., =
1 −1
§1·
¨ ¸
∴ The eigen vector corresponding to Ȝ = 1 is X 3 = ¨ 0 ¸
¨ −1 ¸
© ¹
§1 2 1 ·
¨ ¸
Now M = ¨1 −1 0 ¸
¨1 0 −1¸
© ¹
To find M-1:
ª1 2 1 º
M = ««1 −1 0 »» = 1(1) – 2(–1) + 1(1) = 4
¬«1 0 −1»¼

78
 Sathyabama University

Adj M
M −1 =
M

ª1 2 1 º
M -1 = ««1 −2 1 »»
1
4
«¬1 2 −3»¼

Consider D = M −1 AM

ª1 2 1 º ª 2 2 1 º ª1 2 1 º
= «1 −2 1 »» «« 1 3 1 »» ««1 −1 0 »»
1«
4
«¬1 2 −3»¼ «¬1 2 2 »¼ «¬1 0 −1»¼

ª1 2 1 º ª5 2 1 º
= «1 −2 1 »» ««5 −1 0 »»
1«
4
«¬1 2 −3»¼ «¬5 0 −1»¼

ª 20 0 0 º
= « 0 4 0 »»
1«
4
«¬ 0 0 4 »¼

§5 0 0·
¨ ¸
= ¨0 1 0¸
¨0 0 1¸
© ¹

ª1 −3 3 º
Example 3: Diagonalise the given matrix A= «« 3 −5 3 »» and hence find A3
«¬6 −6 4 »¼

Solution:

ª1 −3 3 º
Given A = «« 3 −5 3 »»
«¬6 −6 4 »¼

The characteristic equation of A is given by A – λ I = 0

The general form of a characteristic equation is

λ 3 − C1λ 2 + C2 λ – C3 = 0

79
Sathyabama University

where C1 = 1 − 5 + 4 = 0

−5 3 1 3 1 −3
C2 = + + = − 12
−6 4 6 4 3 −5

C3 = A = 1( −2 ) + 3 ( −6 ) + (12 ) = 16

∴ The characteristics equation is λ 3 − 12λ – 16 = 0

1 0 −12 −16
−2
0 −2 4 16
1 −2 −8 0

Ȝ = –2 is a root and λ 2 − 2 λ − 8 = 0

( λ − 4) ( λ + 2) = 0

∴ λ = 4, − 2, − 2 are the eigen values of A

The eigen vector corresponding to the eigen value Ȝ is given by
§ x1 ·
( A − λ I ) X = 0 where X = ¨¨ x2 ¸¸ is a column vector.
¨x ¸
© 3¹

ª1 − λ −3 3 º § x1 ·
« 3 ¨ ¸
« −5 − λ 3 »» ¨ x2 ¸ = 0
¬« 6 −6 4 − λ »¼ ©¨ x3 ¹¸

(1 − λ ) x1 − 3 x2 + 3 x3 = 0 ½
°
3 x1 − ( 5 + λ ) x2 + 3 x3 = 0 ¾ (1)
6 x1 − 6 x2 + ( 4 − λ ) x3 = 0 °¿

Case 1: If Ȝ = 4, then (1) becomes

−3x1 − 3 x2 + 3 x3 = 0
3 x1 − 9 x2 + 3x3 = 0
6 x1 − 6 x2 + 0 x3 = 0

Using the rule of cross multiplication for the first two distinct equations,
we get

80
 Sathyabama University

x1 − x2 x3
= =
−3 3 −3 3 −3 −3
−9 3 3 3 3 −9

x1 x2 x3
= =
18 18 36
x1 x2 x3
= =
1 1 2
§1·
∴ The eigen vector is corresponding to Ȝ = 4 is X 1 = ¨¨ 1 ¸¸
¨ 2¸
© ¹
Case 2: If Ȝ = -2 (twice), then (1) becomes
3x1 − 3x2 + 3x3 = 0
3x1 − 3x2 + 3x3 = 0
6 x1 − 6 x2 + 6 x3 = 0

All the three equations are equivalent to a single equation

x1 − x2 + x3 = 0
Put x2 = 0 then x1 + x3 = 0 Ÿ x1 = − x3
x1 x3
i.e., =
1 −1
§1·
∴ The eigen vector corresponding to Ȝ = -2 is X 2 = ¨¨ 0 ¸¸
¨ −1 ¸
© ¹
Now put x3 = 0 Then x1 – x2 = 0 Ÿ x1 = x2
x1 x2
i.e., =
−1 −1
§1·
¨ ¸
∴ The eigen vector corresponding to Ȝ= -2 is X 3 = ¨ 1 ¸
¨0¸
© ¹

81
Sathyabama University

§1 1 1 ·
¨ ¸
Now M = ¨ 1 1 0 ¸
¨ 2 0 −1 ¸
© ¹
To find M −1 :
1 1 1
M = 1 1 0 = 1(-1) -1 (-1) + 1(-2) = -2
2 0 −1

ª −1 1 −1º
1 «
1 −3 1 »»
AdjM
M −1 = = «
M −2
¬« −2 −2 0 »¼

ª −1 1 −1º ª1 −3 3 º ª 1 1 1 º
Consider D = M −1 AM =
1 « 1 −3 1 » « 3 − 5 3 » « 1 1 0 »
−2 « »« »« »
«¬ −2 −2 0 »¼ «¬6 −6 4 »¼ «¬ 2 0 −1»¼

ª −1 1 −1º ª 4 −2 2 º
1 «
= « 1 −3 1 »» «« 4 −2 0 »»
−2
«¬ −2 −2 0 »¼ «¬8 0 −2 »¼

ª −8 0 0 º ª 4 0 0 º
1 «
= « 0 4 0 »» = «« 0 −2 0 »»
−2
«¬ 0 0 4 »¼ «¬ 0 0 −2 »¼

To find A3 :
D = M −1 AM
Pre multiplying by M and Post multiply by M −1 , we get MD M −1 = A
i.e., A = MD M −1

Now A 2 = MD M −1 ⋅ MD M −1 = MD 2 M −1

and A3 = MD 3 M −1
ª64 0 0 º
D 3 = «« 0 −8 0 »»
«¬ 0 0 −8»¼

82
 Sathyabama University

§ 1 1 1 ·§ 64 0 0 ·§ −1 1 −1·
1 ¨ ¸¨ ¸¨ ¸
∴ A3 = ¨ 1 1 0 ¸¨ 0 −8 0 ¸¨ 1 −3 1 ¸
−2 ¨ ¸¨ ¸¨ ¸
© 2 0 −1¹© 0 0 −8 ¹© −2 −2 0 ¹

ª 64 −8 −8º ª −1 1 −1º
−1 «
= « 64 −8 0 »» «« 1 −3 1 »»
2
«¬128 0 8 ¼» ¬« −2 −2 0 ¼»

ª −56 104 −72 º

−1 «
= « −72 88 −72 »»
2
«¬ −144 112 −128»¼

ª 28 −52 36 º
∴ A = ««36 −44 36 »»
3

«¬72 −56 64 »¼

Diagonalisation by Orthogonal Transformation

If A is a real symmetric matrix then the eigen vectors of A will be, not only
linearly independent but also pairwise orthogonal. Form a matrix N whose
columns are the normalized eigen vectors of A. Here N is called the
normalized modal matrix. Since N is orthogonal we have N −1 = N T by the
property of orthogonal matrix.
Therefore the similarity transformation M −1 AM = D takes the form
N T AN = D . Thus transforming A into D by means of the transformation
N T AN is known as diagonalisation by orthogonal transformation or
orthogonal reduction.

Note:

To normalize eigen vector X r , divide each element of X r , by the square root

of the sum of the squares of all the elements of X r .

Calculation of Powers of a matrix: (By orthogonal reduction)

Consider D = N T AN

Then, A = NDN T

83
Sathyabama University

A2 = A ⋅ A = ( NDN T )( NDN T ) = ND 2 N T

∴ A2 = ND 2 N T

In general, An = ND n N T

ªλ1n 0 0" 0 º
« »
0 λ2n 0" 0 »
where D = «
n
«# # # # »
« »
«¬ 0 0 0" λnn »¼

Working rule for diagonalisation by orthogonal transformaion

Step 1: Given a real symmetric matrix A, find the characteristic equation.

Step 2: Find the eigen values
Step 3: Find the eigen vectors which are pairwise orthogonal.
Step 4: Form the normalized modal matrix N.
T
Step 5: Find N
Step 6: Calculate N T AN = D

§ 0 0 1·
¨ ¸
Example: 1 Diagonalise the matrix ¨ 0 0 1¸ by orthogonal transformation.
¨ 1 1 1¸
© ¹

Solution:

§ 0 0 1·
¨ ¸
Let A = ¨ 0 0 1¸ be a given matrix.
¨ 1 1 1¸
© ¹

The characteristic equation of A is given by A − λ I = 0

The general form of a characteristic equation is given by

λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 0 + 0 + 1 = 1

84
 Sathyabama University

0 1 0 1 0 0
C2 = + + = −2
1 1 1 1 0 0

0 0 1
C3 = A = 0 0 1
1 1 1

= 0 ( 0 − 1) – 0 ( 0 − 1) + 1 ( 0 − 0 )

= 0

∴ The characteristic equation is λ 3 − λ 2 − 2λ = 0

λ (λ 2 − λ − 2) = 0

λ = 0 and λ 2 − λ − 2 = 0
( λ - 2) ( λ + 1) = 0
λ = 2, − 1
∴ λ = 0, 2, − 1 are the eigen values of A.
The eigen vector corresponding to the eigen value λ is given by
§ x1 ·
¨ ¸
( A − λ I ) X = 0, where X = ¨ x2 ¸ is a column vector.
¨x ¸
© 3¹

§0 − λ 0 1 · § x1 ·
¨ ¸¨ ¸
i.e., ¨ 0 0−λ 1 ¸ ¨ x2 ¸ = 0
¨ 1 1 − λ ¸¹ ¨© x3 ¸¹
© 1

− λ x1 + 0 x2 + x3 = 0 ½
°
0 x1 − λ x2 + x3 = 0 ¾ (1)
x1 + x2 + (1 − λ ) x 3 = 0 °
¿
Case 1: If λ = 0 then (1) becomes
0 x1 + 0 x2 + x3 = 0
0 x1 + 0 x2 + x3 = 0
x1 + x2 + x3 = 0

85
Sathyabama University

Using the cross multiplication rule for the second and third equation, we
get

x1 − x2 x
= = 3
0 1 0 1 0 0
1 1 1 1 1 1
x x x
i.e. 1 = 2 = 3
−1 1 0

§ −1·
∴ The eigen vector corresponding to Ȝ=0 is X1 = ¨ 1 ¸
¨0¸
© ¹
Case 2: If Ȝ = 2 then (1) becomes

2 x1 + 0 x2 + x3 = 0
0 x1 − 2 x2 + x3 = 0
x1 + x2 − x3 = 0

Using cross multiplication rule for the first two distinct equations, we get
x1 − x2 x3
= =
0 1 −2 1 −2 0
−2 1 0 1 0 −2
x1 x2 x3
i.e., = =
2 2 4
x1 x2 x3
i.e., = =
1 1 2

§1·
∴ The eigen vector corresponding to Ȝ = 2 is X 2 = ¨ 1 ¸
¨ ¸
© 2¹
Case 3: If Ȝ = –1, then (1) becomes
x1 + 0 x2 + x3 = 0
0 x1 + x2 + x3 = 0
x1 + x2 + 2 x3 = 0

86
 Sathyabama University

Using the rule of cross multiplication for the first two distinct equations,
we get
x1 − x2 x
= = 3
0 1 1 1 1 0
1 1 0 1 0 1

x1 x2 x3
i.e., = = .
−1 −1 1
§ −1·
¨ ¸
∴ The eigen vector corresponding to Ȝ=-1 is, X 3 = ¨ −1¸
¨1¸
© ¹

§ −1 1 −1·
¨ ¸
Modal Matrix M = ¨ 1 1 −1¸
¨0 2 1¸
© ¹

§ −1 1 −1 ·
¨ ¸
¨ 2 6 3¸
¨ 1 1 −1 ¸
Normalised Modal matrix N = ¨ ¸
¨ 2 6 3¸
¨ 2 1 ¸
¨ 0 ¸
© 6 3¹

§ −1 1 ·
¨ 0 ¸
¨ 2 2 ¸
¨ 1 1 2 ¸
NT = ¨ ¸
¨ 6 6 6¸
¨ −1 −1 1 ¸
¨ ¸
© 3 3 3¹

§ −1 1 · § −1 1 −1 ·
¨ 0 ¸ ¨ ¸
¨ 2 2 ¸ § 0 0 1· ¨ 2 6 3¸
¨ 1 1 2 ¸¨ ¸¨ 1 1 −1 ¸
Consider N T AN = ¨ ¸ 0 0 1 ¸¨ 2 ¸
6 ¸ ¨¨
¨ 1 1 1¸¹ ¨
6 6 6 3¸
¨ −1 −1 1 ¸ © ¨ 2 1 ¸
¨ ¸ ¨ 0 ¸
© 3 3 3¹ © 6 3¹

87
Sathyabama University

§ −1 1 ·§ 2 1 ·
¨ 0 ¸¨ 0 ¸
¨ 2 2 ¸¨ 6 3 ¸ 0 0 0
§ ·
¨ 1 1 2 ¸¨ 2 1 ¸ ¨ ¸
=¨ ¸¨ 0 ¸= ¨0 2 0 ¸ = D
3 ¸ ¨
¨ 0 0 −1¸¹
6 6 6 ¸¨ 6
¨ −1 −1 1 ¸¨ 4 1 ¸ ©
¨ ¸¨ 0 − ¸
© 3 3 3 ¹© 6 3¹

§2 0 1·
¨ ¸
Example 2: Reduce the ¨ 0 2 0 ¸ to a diagonal form by orthogonal
¨1 0 2¸
© ¹
reduction.

Solution

§2 0 1·
¨ ¸
Let A = ¨ 0 2 0 ¸ be the given matrix.
¨ 1 0 2¸
© ¹
The characteristic equation of A is A − λ I = 0

The general form of a characteristic equation is

λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 2 + 2 + 2 = 6

2 0 2 1 2 0
C2 = + + = 11
0 2 1 2 0 2
C3 = A = 6
3 2
∴ The Characteristic equation of A is λ − 6λ + 11λ − 6 = 0

1 −6 11 −6
1
0 1 −5 6
1 −5 6 0
2
λ = 1 is a root and λ − 5λ + 6 = 0
( λ −2) ( λ −3) = 0
λ = 2, 3

88
 Sathyabama University

∴ λ = 1, 2, 3 are the eigen values of A.

The eigen vector corresponding to the eigen value λ is given by
§ X1 ·
( A − λ I ) X = 0, where X = ¨¨ X 2 ¸¸ is a column vector.
¨X ¸
© 3¹

§2−λ 0 1 · § x1 ·
¨ ¸¨ ¸
i.e., ¨ 0 2−λ 0 ¸ ¨ x2 ¸ = 0
¨ 1 2 − λ ¸¹ ¨© x3 ¸¹
© 0

(2 − λ ) x1 + 0 x2 + x3 = 0 ½
°
0 x1 + (2 − λ ) x2 + 0 x3 = 0 ¾ (1)
x1 + 0 x2 + (2 − λ ) x3 = 0 °¿

Case 1: If λ = 1, then (1) becomes,

x1 + 0 x2 + x3 = 0
0 x1 + x2 + 0 x3 = 0
x1 + 0 x2 + x3 = 0

Using cross multiplication rule for the first two distinct equations, we get
x1 x2 x3
= =
0 1 1 1 1 0
1 0 0 0 0 1
x1 x2 x3
i.e., = =
−1 0 1

§ −1·
¨ ¸
∴ The eigen vector corresponding to λ = 1 is X 1 = ¨ 0 ¸
¨1¸
© ¹
Case 2: If λ = 2, then (1) becomes
0 x1 + 0 x2 + x3 = 0 i.e x3 = 0
0 x1 + 0 x2 + 0 x3 = 0
x1 + 0 x2 + 0 x3 = 0 i.e x1 = 0

Let x2 be any arbitrary value say x2 = 1

89
Sathyabama University

§0·
¨ ¸
∴ The eigen vector corresponding to λ = 2 is X 2 = ¨ 1 ¸
¨0¸
© ¹
Case 3: If λ = 3, then (1) becomes
− x1 + 0 x2 + x3 = 0
− 0 x1 − x2 + 0 x3 = 0
x1 + 0 x2 − x3 = 0

Using cross multiplication for the first two distinct equations we get
x1 − x2 x1
= =
0 1 −1 1 −1 0
−1 0 0 0 0 −1

x1 x2 x3
i.e., = =
1 0 1

§1·
¨ ¸
∴ The eigen vector corresponding to λ = 3 is X 3 = ¨ 0 ¸
¨1¸
© ¹

§ −1 0 1 ·
¨ ¸
Modal matrix M = ¨ 0 1 0 ¸
¨ 1 0 1¸
© ¹

§ −1 1 ·
¨ 0 ¸
¨ 2 2¸
Normalized modal matrix N = ¨ 0 1 0 ¸
¨ ¸
¨ 1 1 ¸
¨ 0 ¸
© 2 2¹

§ −1 1 ·
¨ 0 ¸
¨ 2 2¸
NT = ¨ 0 1 0 ¸
¨ ¸
¨ 1 1 ¸
¨ 0 ¸
© 2 2¹

Consider N T AN ,

90
 Sathyabama University

§ −1 1 · § −1 1 ·
¨ 0 ¸§ 2 0 1·¨ 0 ¸
¨ 2 2¸ 2 2¸
¨ ¸¨
N T AN = ¨ 0 1 0 ¸¨ 0 2 0¸¨ 0 1 0 ¸
¨ ¸ ¨ ¸
¨ 1 1 ¸ ¨© 1 0 2 ¸¹ ¨ 1 1 ¸
¨ 0 ¸ ¨ 0 ¸
© 2 2¹ © 2 2¹

§ −1 1 ·§ −1 3 ·
¨ 0 ¸¨ 0 ¸
¨ 2 2¸¨ 2 2 ¸ §1 0 0·
¨ ¸
=¨ 0 1 0 ¸¨ 0 2 0 ¸ = ¨0 2 0¸ = D
¨ ¸¨ ¸
¨ 1 1 ¸¨ 1 3 ¸ ¨© 0 0 3 ¸¹
¨ 0 ¸¨ 0 ¸
© 2 2¹© 2 2¹

§ 1 1 3·
¨ ¸
Example 3: Diagonalise the matrix ¨ 1 5 1 ¸ by orthogonal reduction.
¨ 3 1 1¸
© ¹

Solution:

§ 1 1 3·
¨ ¸
Let A = ¨ 1 5 1 ¸ be the given matrix
¨ 3 1 1¸
© ¹
The characteristic equation of A is given by A − λ I = 0
The general form of a characteristic equation is λ 3 – C1λ 2 + C2 λ – C3 = 0

where C1 = 1 + 5 + 1 = 7

5 1 1 3 1 1
C2 = + + =0
1 1 3 1 1 5

C3 = A = −36

∴ The Characteristic equation of A is λ 3 − 7λ 2 + 36 = 0

1 −7 0 36
−2
0 −2 18 −36
1 −9 18 0

∴ λ = -2 is a root and λ 2 − 9λ + 18 = 0
(λ − 3) (λ − 6) = 0

91
Sathyabama University

λ = 3, 6
∴ λ = −2, 3, 6 are the eigen values of A.
The eigen vector corresponding to the eigen value λ is given by
§ x1 ·
( A − λ I ) X = 0, where X = ¨¨ x2 ¸¸ is a column vector.
¨x ¸
© 3¹

§1 − λ 1 3 · § x1 ·
¨ ¸¨ ¸
i.e., ¨ 1 5−λ 1 ¸ ¨ x2 ¸ = 0
¨ 3 1 − λ ¸¹ ¨© x3 ¸¹
© 1

i.e., (1 − λ ) x1 + x2 + 3 x3 = 0 ½
°
x2 + (5 − λ ) x2 + x3 = 0 ¾ (1)
3 x1 + x2 + (1 − λ ) x3 = 0 ¿°

Case 1: If Ȝ = -2, then (1) becomes,

3x1 + x2 + 3x3 = 0
x1 + 7 x2 + x3 = 0
x1 + x2 + 3x3 = 0

Using cross multiplication rule for the first two distinct equations, we get
x1 − x1 x
= = 3
1 3 3 3 3 1
7 1 1 1 1 7
x1 x x
i.e., = 2 = 3
−20 0 20
x1 x2 x3
i.e., = = .
1 0 −1
§1·
¨ ¸
∴ The eigen vector corresponding to Ȝ = –2 is X 1 = ¨ 0 ¸
¨ −1 ¸
© ¹
Case 2: If λ = 3, then (1) becomes
−2 x1 + x2 + 3 x3 = 0
x1 + 2 x2 + x3 = 0
3 x1 + x2 − 2 x3 = 0

92
 Sathyabama University

Using cross multiplication rule for the first two distinct equations, we get
x1 − x2 x3
= =
1 3 −2 3 −2 1
2 1 1 1 1 2
x1 x2 x3
i.e., = =
−5 5 −5
x1 x2 x3
i.e., = = .
1 −1 1
§1·
∴ The eigen vector corresponding to λ = 3 is X 2 = ¨¨ −1¸¸
¨1¸
© ¹
Case 3: If Ȝ = 6, then (1) becomes
−5 x1 + x2 + 3x3 = 0
x1 − x2 + x3 = 0
3x1 + x2 − 5 x3 = 0

Using the rule of cross multiplication for the first two distinct equations we
get
x1 − x2 x3
= =
1 3 −5 3 −5 1
−1 1 1 1 1 −1
x1 x2 x3
i.e., = =
4 8 4
x1 x2 x3
i.e., = = .
1 2 1
§1·
∴ The eigen vector corresponding to λ = 6 is, X 3 = ¨¨ 2 ¸¸
¨1¸
© ¹
§ 1 1 1·
¨ ¸
M = ¨ 0 −1 2 ¸
¨ −1 1 1 ¸
Modal matrix © ¹

93
Sathyabama University

§ 1 1 1 ·
¨ ¸
¨ 2 3 6¸
¨ −1 2 ¸
Normalized modal matrix N = ¨ 0 ¸
¨ 3 6¸
¨ −1 1 1 ¸
¨ ¸
© 2 3 6¹

§ 1 −1 ·
¨ 0 ¸
¨ 2 2¸
¨ 1 −1 1 ¸
NT = ¨ ¸
¨ 3 3 3¸
¨ 1 2 1 ¸
¨ ¸
© 6 6 6¹

Consider N T AN

§ 1 −1 · § 1 1 1 ·
¨ 0 ¸ ¨ ¸
¨ 2 2¸ 1 1 3 ¨ 2 3 6¸
§ ·
¨ 1 −1 1 ¸¨ ¸¨ −1 2 ¸
N T AN = ¨ ¸¨1 5 1¸¨ 0 ¸
3 ¸¨
¨ 3 1 1 ¸¹ ¨
3 3 3 6¸
¨ 1 2 1 ¸© ¨ −1 1 1 ¸
¨ ¸ ¨ ¸
© 6 6 6¹ © 2 3 6¹

§ 1 −1 ·§ −2 3 6 ·
¨ 0 ¸¨ ¸
¨ 2 2 ¸¨ 2 3 6¸
¨ 1 −1 1 ¸¨ −3 12 ¸
=¨ ¸¨ 0 ¸
¨ 3 3 3 ¸¨ 3 6¸
¨ 1 2 1 ¸¨ 2 3 6 ¸
¨ ¸¨ ¸
© 6 6 6 ¹© 2 3 6¹

§ −2 0 0 ·
¨ ¸
= ¨ 0 3 0¸
¨ 0 0 6¸
© ¹
= D

94
 Sathyabama University

Example 4:

§ 8 −6 2 ·
¨ ¸
Reduce the matrix A = ¨ −6 7 −4 ¸ to a diagonal form by orthogonal
¨ 2 −4 3 ¸
© ¹
4
transformation & hence find A .

Solution

§ 8 −6 2 ·
¨ ¸
Given A = ¨ −6 7 −4 ¸
¨ 2 −4 3 ¸
© ¹
The characteristic equation of A is A − λ I = 0

The general form of a characteristic equation is Ȝ 3 − C1Ȝ 2 + C2 Ȝ − C3 = 0

where C1 = 8 + 7 + 3 = 18

7 −4 8 2 8 −6
C2 = + + = 45
−4 3 2 3 −6 7

C3 = A = 0

∴ The Characteristic equation of A is λ 3 – 18λ 2 + 45λ = 0

2
λ (λ – 18λ + 45) = 0

Ȝ = 0 is a root and λ 2 – 18λ + 45 = 0

( Ȝ − 3)( Ȝ − 15) = 0
Ȝ = 3, 15

∴ Ȝ = 0, 3, 15 are the eigen values of A.

The eigen vector corresponding to the eigen value Ȝ is given by
§ x1 ·
( A-ȜI ) X = 0 , where X = ¨¨ x2 ¸¸ is a column vector.
¨x ¸
© 3¹

95
Sathyabama University

§ 8 − Ȝ −6 2 · § x1 ·
¨ ¸¨ ¸
¨ −6 7 − Ȝ −4 ¸ ¨ x2 ¸ =0
¨ 2 −4 3 − Ȝ ¸¹ ¨© x3 ¸¹
©

i.e., (8 − Ȝ ) x1 − 6 x2 + 2 x3 = 0 ½
°
− 6 x1 + ( 7 − Ȝ ) x2 − 4 x3 = 0 ¾ (1)
°
2 x1 − 4 x2 + ( 3 − Ȝ ) x3 = 0 ¿

Case 1: If Ȝ = 0, then (1) becomes,

8 x1 − 6 x2 + 2 x3 = 0
−6 x1 − 7 x2 − 4 x3 = 0
2 x1 − 4 x2 + 3 x3 = 0

Using cross multiplication rule for the first two distinct equations we get,
x1 − x2 x3
= =
−6 2 8 2 8 −6
7 −4 −6 −4 −6 7
x1 x2 x3
i.e., = =
10 20 20
x1 x2 x3
i.e., = =
1 2 2
§1·
¨ ¸
∴ The eigen vector corresponding to λ = 0 is X 1 = ¨ 2 ¸
¨ 2¸
© ¹
Case 2: It λ = 3 , then (1) becomes
5 x1 − 6 x2 + 2 x3 = 0
−6 x1 + 4 x2 − 4 x3 = 0
2 x1 − 4 x2 + 0 x3 = 0

Using cross multiplication rule for the first two distinct equations we get
x1 − x2 x3
= =
−6 2 5 2 5 −6
4 −4 −6 −4 −6 4
x1 x2 x3
i.e., =
16 8 −16

96
 Sathyabama University

x1 x2 x3
i.e., = =
2 1 −2
§ 2·
∴ The eigen vector corresponding to λ = 3 is X 2 = ¨¨ 1 ¸¸
¨ −2 ¸
© ¹
Case3: If λ = -15, then (1) becomes
−7 x1 − 6 x2 + 2 x3 = 0
−6 x1 + 8 x2 − 4 x3 = 0
2 x1 − 4 x2 − 12 x3 = 0

Using cross multiplication rule for the first two distinct equations we get,
x1 − x2 x3
= =
−6 2 −7 2 −7 −6
−8 −4 −6 −4 −6 −8
x1 x x
i.e., = 2 = 3
40 −40 20
x1 x2 x3
i.e., = =
2 −2 1
§ 2·
¨ ¸
∴ The eigen vector corresponding to λ = 15 is X 3 = ¨ −2 ¸
¨1¸
© ¹
§1 2 2 ·
Modal matrix M = ¨¨ 2 1 −2 ¸¸
¨ 2 −2 1 ¸
© ¹
§1 2 2 ·
¨3 3 3 ¸
¨ ¸
Normalised Modal matrix N = ¨ 2 1 −2 ¸
¨3 3 3 ¸
¨ ¸
¨2 −2 1 ¸
¨ ¸
©3 3 3 ¹
§1 2 2 ·
¨3 3 3 ¸
¨ ¸
N = ¨¨
T 2 1 −2 ¸
3 3 3 ¸
¨ ¸
¨2 −2 1 ¸
¨ ¸
©3 3 3 ¹

97
Sathyabama University

Consider N T AN

§1 2 2 · §1 2 2 ·
¨3 3 ¸
3 § 8 −6 2 · 3 ¨ 3 3 ¸
¨ ¸ ¨ ¸
−2 ¸ ¨ ¸¨ 2
N T AN = ¨¨
2 1 1 −2 ¸
−6 7 −4 ¸ ¨
3 3 3 ¸ ¨¨ 3 3 3 ¸
¨ ¸ © 2 −4 3 ¸¹ ¨ ¸
¨2 −2 1 ¸ ¨2 −2 1 ¸
¨ ¸ ¨ ¸
©3 3 3 ¹ ©3 3 3 ¹

§1 2 2 ·
¨3 3 3 ¸¸
¨ § 0 2 10 ·
¨ ¸
= ¨¨
2 1 −2 ¸
3 3 3 ¸ ¨ 0 1 −10 ¸
¨ ¸ ¨ 0 −2 5 ¸
−2 © ¹
¨2 1 ¸
¨ ¸
©3 3 3 ¹
§0 0 0 ·
¨ ¸
= ¨0 3 0 ¸ = D
¨ 0 0 15 ¸
© ¹

To find A 4 :
We know that, An = ND n N T
∴ A4 = ND 4 N T
§1 2 2 · §1 2 2 ·
¨3 3 ¸
3 §0 0 ¨ 3 ¸
¨ ¸ 0 ·¨ 3 3
¸
−2 ¸ ¨ ¸ 2
= ¨¨ 0 ¸ ¨¨
2 1 1 −2 ¸
¸ ¨ 0 81
3 3 3 ¨ 3 3 3 ¸
¨ ¸ © 0 0 50625 ¸¹ ¨ ¸
¨2 −2 1 ¸ ¨2 −2 1 ¸
¨ ¸ ¨ ¸
©3 3 3 ¹ ©3 3 3 ¹
§1 2 2 ·
¨3 3 3 ¸¸ § 0
¨ 0 0 ·
−2 ¸ ¨ ¸
= ¨¨
2 1
¸ ¨ 54 27 −54 ¸
3 3 3 ¨
¨ ¸ 33750 −33750 16875 ¸¹
¨2 −2 1 ¸©
¨ ¸
©3 3 3 ¹
§ 22536 −22482 11214 ·
¨ ¸
= ¨ −22482 22509 −11268 ¸
¨ 11214 −11268 5661 ¸
© ¹

98
 Sathyabama University

Example 5:
Reduce the following symmetric matrix to a diagonal matrix,
§ 2 −1 1 ·
¨ ¸
A = ¨ −1 2 −1¸
¨ 1 −1 2 ¸
© ¹

Solution:

ª 2 −1 1 º
Given A = «« −1 2 −1»»
«¬ 1 −1 2 »¼

The characteristic equation of A is given by |A – ȜI|=0.

The general form of the characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 2 + 2 + 2 = 6
2 −1 2 1 2 −1
C2 = + + =9
−1 2 1 2 −1 2
C3 =| A |= 4

∴ The characteristic equation is λ 3 − 6λ 2 + 9λ − 4 = 0

1 −6 9 −4
1
0 1 −5 4
1 −5 4 0

∴ λ = 1 is a root and λ 2 − 5λ + 4 = 0

( λ − 1)( λ − 4 ) = 0
λ = 1, 4

∴ λ = 4,1,1 are the eigen values of A.

The eigen vectors corresponding to the eigen value Ȝ is given by
§ x1 ·
¨ ¸
( A − λ I ) X = 0 , where X = ¨ x2 ¸ is a column vector
¨x ¸
© 3¹

99
Sathyabama University

ª2 − λ −1 1 º § x1 ·
« ¨ ¸
« −1 2−λ −1 »» ¨ x2 ¸ = 0
¬« 1 −1 2 − λ »¼ ©¨ x3 ¹¸

( 2 − λ ) x1 − x2 + x3 = 0 ½
°
i.e., − x1 + ( 2 − λ ) x2 − x3 = 0 ¾ (1)
°
x1 − x2 + ( 2 − λ ) x3 = 0 ¿

Case (1): If Ȝ = 4, then (1) becomes

− 2 x1 − x2 + x3 = 0
− x1 − 2 x2 − x3 = 0
x1 − x2 − 2 x3 = 0

Using cross multiplication rule for the first two distinct equation we get
x1 − x2 x3
= =
−1 1 −2 1 −2 −1
−2 −1 −1 −1 −1 −2
x1 x2 x3
= =
3 −3 3
x1 x2 x3
= =
1 −1 1

§1·
∴ The eigen vector corresponding to the eigen value λ = 4 is X 1 = ¨¨ −1¸¸
¨1¸
© ¹
Case 2: If Ȝ = 1 (twice), then (1) becomes
x1 – x2 + x3 = 0
– x1 + x2 – x3 = 0
x1 – x2 + x3 = 0

All the three equations are the same.

Put x3 = 0. Then x2 − x3 = 0
Ÿ x2 = x3
x2 x3
i.e., =
1 1

100
 Sathyabama University

§0·
¨ ¸
The eigen vector corresponding to the eigen value Ȝ = 1 is X 2 = ¨ 1 ¸
¨1¸
© ¹
Since all the three eigen vectors should be orthogonal to each other, we
§a·
¨ ¸
assume the third eigen vector X 3 = ¨ b ¸ to be orthogonal to X1 and X2.
¨c¸
© ¹

i.e., X 1T X 3 = 0 and X 2T X 3 = 0

§a· §a·
(1 − 1 1) ¨¨ b ¸¸ = 0 and ( 0 1 1) ¨¨ b ¸¸ = 0
¨c¸ ¨c¸
© ¹ © ¹
a −b+c = 0
0⋅a + b + c = 0
Using cross multiplication rule we get,
a −b c
= =
−1 1 1 1 1 − 1
1 1 0 1 0 1

a b c
i.e., = =
−2 −1 1
§ −2 ·
¨ ¸
The eigen vector which is orthogonal to both X1 and X2 is X 3 = ¨ −1 ¸
¨1¸
© ¹
§ 1 0 −2 ·
¨ ¸
Modal matrix M = ¨ −1 1 −1 ¸
¨1 1 1¸
© ¹

§ 1 −2 ·
¨ 0 ¸
¨ 3 6 ¸
¨ −1 1 1 ¸
Normalized modal matrix N = ¨ − ¸
¨ 3 2 6¸
¨ 1 1 1 ¸
¨ ¸
© 3 2 6 ¹

101
Sathyabama University

§ 1 −1 1 ·
¨ ¸
¨ 3 3 3¸
¨ 1 1 ¸
NT = ¨ 0 ¸
¨ 2 2¸
¨ −2 −1 1 ¸
¨ ¸
© 6 6 6¹

Consider NTAN

§ 1 −1 1 · § 1 −2 ·
¨ ¸ ¨ 0 ¸
¨ 3 3 3 ¸ 2 −1 1 ¨ 3 6¸
§ ·
¨ 1 1 ¸¨ ¸ ¨ −1 1 −1 ¸
N T AN = ¨ 0 ¸ ¨ −1 2 −1¸ ¨ ¸
2 ¸¨
1 −1 2 ¸¹ ¨
¨ 2 3 2 6¸
¨ −2 −1 1 ¸ © ¨ 1 1 1 ¸
¨ ¸ ¨ ¸
© 6 6 6¹ © 3 2 6¹
§ 1 −1 1 ·§ 4 −2 ·
¨ ¸¨ 0 ¸
¨ 3 3 3 ¸¨ 3 6¸
¨ 1 1 ¸ ¨ −4 1 −1 ¸
=¨ 0 ¸¨ ¸
¨ 2 2 ¸¨ 3 2 6¸
¨ −2 −1 1 ¸¨ 4 1 1 ¸
¨ ¸¨ ¸
© 6 6 6 ¹© 3 2 6¹
§ 4 0 0·
¨ ¸
= ¨ 0 1 0¸
¨0 0 1¸
© ¹
=D
Example 6:
§ 2 0 4·
¨ ¸
Diagonalise the matrix A = ¨ 0 6 0 ¸ by means of an orthogonal
¨ 4 0 2¸
© ¹
transformation.

Solution

§ 2 0 4·
¨ ¸
Given A = ¨ 0 6 0 ¸
¨ 4 0 2¸
© ¹

102
 Sathyabama University

The characteristic equation of A is given by |A – ȜI| = 0

The general form of a characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 2 + 6 + 2 = 10

6 0 2 4 2 0
C2 = + + = 12
0 2 4 2 0 6
C3 = | A | = −72

∴ The characteristic equation is λ 3 − 10λ 2 + 12λ + 72 = 0

1 −10 12 72
−2
0 −2 24 −72
1 − 12 36 0

∴ Ȝ = -2 is a root and λ 2 − 12λ + 36 = 0

( λ − 6 )( λ − 6 ) = 0
λ = 6, 6
∴ Ȝ = -2, 6, 6 are the eigen values of A.
The eigen vectors corresponding to the eigen value Ȝ is given by (A – ȜI)X = 0
ª x1 º
where X = «« x2 »» is a column vector.
«¬ x3 »¼

§2−λ 0 4 ·§ x1 ·
¨ ¸¨ ¸
¨ 0 6−λ 0 ¸¨ x2 ¸ = 0
¨ 4 2 − λ ¸¨ ¸
© 0 ¹© x3 ¹

i.e. ( 2 − λ ) x1 + 0 x2 + 4 x3 = 0½
°
0 x1 + ( 6 − λ ) x2 + 0 x3 = 0 ¾ (1)
°
4 x1 + 0 x2 + ( 2 − λ ) x3 = 0 ¿

Case 1: If Ȝ = -2, then (1) becomes,

4 x1 + 0 x2 + 4 x3 = 0
0 x1 + 8 x2 + 0 x3 = 0
4 x1 + 0 x2 + 4 x3 = 0

Using cross multiplication rule for the first two distinct equations we get,

103
Sathyabama University

x1 − x2 x3
= =
0 4 4 4 4 0
8 0 0 0 0 8
x1 x x
i.e., = 2 = 3
−32 0 32
x x x
i.e., 1 = 2 = 3
−1 0 1

§ −1·
¨ ¸
The eigen vector corresponding to the eigen value. λ = −2 is X 1 = ¨ 0 ¸
¨1¸
© ¹
Case 2: If Ȝ = 6, then (1) becomes,
−4 x1 + 0 x2 + 4 x3 = 0
0 x1 + 0 x2 + 0 x3 = 0
4 x1 + 0 x2 − 4 x3 = 0

The above equations reduce to a single equation

4 x1 − 4 x3 = 0
i.e., x1 = x3

Let x2 can be any arbitrary value, say x2 = 0

§1·
¨ ¸
∴ The eigen vector corresponding to Ȝ = 6 is X2 = ¨ 0 ¸
¨1¸
© ¹
Since all the three eigen vectors should be orthogonal to each other, we
§a·
¨ ¸
assume the third eigen vector X 3 = ¨ b ¸ to be orthogonal to X1 and X2
¨c¸
© ¹

∴ X 1T X 3 = 0 & X 2T X 3 = 0
§a· §a·
( −1 0 1) ¨¨ b ¸¸ = 0 & (1 0 1) ¨¨ b ¸¸ = 0
¨c¸ ¨c¸
© ¹ © ¹
a + 0b − c = 0
a + 0b + c = 0

104
 Sathyabama University

Using cross multiplication rule, we get

a −b c
= =
0 −1 1 −1 1 0
0 1 1 1 1 0

a b c
i.e., = =
0 −2 0
a b c
i.e., = =
0 1 0

§0·
¨ ¸
∴ The eigen vector corresponding to the eigen value Ȝ = 6 is X 3 = ¨ 1 ¸
¨0¸
© ¹

§ −1 1 0 ·
¨ ¸
Modal matrix M = ¨ 0 0 1 ¸
¨ 1 1 0¸
© ¹
§ −1 1 ·
¨ 0¸
¨ 2 2 ¸
Normalised modal matrix N = ¨ 0 0 1¸
¨ ¸
¨ 1 1
¨ 0 ¸¸
© 2 2 ¹
§ −1 1 ·
¨ 0 ¸
¨ 2 2¸
¨ 1 1 ¸
NT = ¨ 0 ¸
¨ 2 2¸
¨ 0 1 0 ¸
¨ ¸
© ¹
Consider N T AN ,
§ −1 1 ·
¨ 0 ¸ § −1 1 ·
2 2¸ 2 0 4 ¨ 0¸
¨ § ·¨ 2 2
T ¨ 1 1 ¸¨ ¸ ¸
N AN = ¨ 0 ¸¨0 6 0¸¨ 0 0 1¸
2 ¸¨ ¨ ¸
4 0 2 ¸¹ ¨ 1
¨ 2
¨ 0 ¸ © 1
0 ¸¸
0 1 ¨
¨ ¸ © 2 2 ¹
© ¹

105
Sathyabama University

§ −1 1 ·
¨ 0 ¸§ 2 6 ·
0¸
¨ 2 2 ¸¨
2 2
¨ 1 1 ¸¨ ¸
=¨ 0 ¸¨ 0 0 6¸
¨ 2 2 ¸¨ ¸
−2 6
¨ 0 1 0 ¸ ¨¨ ¸
0¸
¨ ¸© 2 2 ¹
© ¹
§ −2 0 0 ·
¨ ¸
= ¨ 0 6 0¸
¨ 0 0 6¸
© ¹
= D.

Example 7
§0 1 1·
¨ ¸
Diagonalise ¨ 1 0 1 ¸ by orthogonal transformation
¨1 1 0¸
© ¹
Solution
§0 1 1·
¨ ¸
Let A = ¨ 1 0 1 ¸ be the matrix
¨1 1 0¸
© ¹

The characteristic equation of A is given by A − λ I = 0 |

The general form of the characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 0 + 0 + 0 = 0

0 1 0 1 0 1
C2 = + + = −3
1 0 1 0 1 0
C3 = A = 2

∴ The characterstic equation is Ȝ3 – 3Ȝ – 2 = 0

1 0 −3 −2
−1
0 −1 1 2
1 −1 − 2 0

Ȝ = -1 is a root and λ 2 − λ − 2 = 0

106
 Sathyabama University

( λ + 1)( λ − 2 ) = 0
λ = −1, 2
∴ λ = −1, − 1, 2 are the eigen values of A.

The eigen vector corresponding to the eigen value Ȝ is given by

ª x1 º
( A − λ I ) X = 0, ere X = «« x2 »» is a column vector.
«¬ x3 »¼

§0 − λ 1 1 ·§ x1 ·
¨ ¸¨ ¸
¨ 1 0−λ 1 ¸¨ x2 ¸ = 0
¨ 1 0 − λ ¸¨ ¸
© 1 ¹© x3 ¹

i.e., − λ x1 + x2 + x3 = 0 ½
°
x1 − λ x2 + x3 = 0 ¾ (1)
x1 + x2 − λ x3 = 0 ¿°

Case 1: If Ȝ = 2, then (1) becomes

−2 x1 + x2 + x3 = 0
x1 − 2 x2 + x3 = 0
x1 + x2 − 2 x3 = 0

Using cross multiplication rule for the first two distinct equations we get
x1 − x2 x3
= =
1 1 −2 1 −2 1
−2 1 1 1 1 −2
x1 x2 x3
= =
3 3 3
x x2 x
i.e., 1 = = 3
1 1 1

§ 1·
¨ ¸
The eigen vector corresponding to the eigen value λ = 2 is X 1 = ¨1¸ .
¨ 1¸
© ¹
Case 2: If Ȝ = –1, then (1) becomes,

107
Sathyabama University

x1 + x2 + x3 = 0
x1 + x2 + x3 = 0
x1 + x2 + x3 = 0

All the three equations are the same.

Put x2 = 0. Then we get, x1 + x3 = 0
Ÿ x1 = –x3
x1 x3
i.e. =
1 −1
∴ The eigen vector corresponding to the eigen value λ = −1 is
§1·
¨ ¸
X2 = ¨ 0 ¸
¨ −1¸
© ¹
Since all the three eigen vectors should be orthogonal to each other we
§a·
¨ ¸
assume the third eigen vector X 3 = ¨ b ¸ to be orthogonal to X1 & X2.
¨c¸
© ¹
∴ X 1T X 2 = 0 & X 1T X 3 = 0
§a· §a·
(1 1 1) ¨¨ b ¸¸ = 0 & (1 0 −1) ¨¨ b ¸¸ = 0
¨ ¸ ¨c¸
©c¹ © ¹
a+b+c = 0
a + 0b + c = 0
Using cross multiplication rule we get
a −b c
= =
1 1 1 1 1 1
0 −1 1 −1 1 0
a b c
i.e., = =
−1 2 −1
a b c
i.e., = =
1 −2 1
§1·
¨ ¸
∴ The eigen vector which is orthogonal to both X1 & X2 is X 3 = ¨ −2 ¸
¨1¸
© ¹

108
 Sathyabama University

§1 1 1 ·
¨ ¸
Modal matrix M = ¨ 1 0 −2 ¸
¨ 1 −1 1 ¸
© ¹
§ 1 1 1 ·
¨ ¸
¨ 3 2 6¸
¨ 1 −2 ¸
Normalised modal matrix N = ¨ 0 ¸
¨ 3 6¸
¨ 1 −1 1 ¸
¨ ¸
© 3 2 6¹
§ 1 1 1 ·
¨ ¸
¨ 3 2 6¸
¨ 1 −1 ¸
NT = ¨ 0 ¸
¨ 2 2¸
¨ 1 −1 1 ¸
¨ ¸
© 6 6 6¹

Consider N T AN ,

§ 1 1 1 · § 1 1 1 ·
¨ ¸ ¨ ¸
¨ 3 3 3¸ 0 1 1 ¨ 3 2 6¸
§ ·
¨ 1 −1 ¸ ¨ ¸¨ 1 −2 ¸
N T AN = ¨ 0 ¸¨1 0 1¸¨ 0 ¸
2 ¸¨
1 1 0 ¸¹ ¨
¨ 2 3 6¸
¨ 1 −2 1 ¸© ¨ 1 −1 1 ¸
¨ ¸ ¨ ¸
© 6 6 6¹ © 3 2 6¹
§ 1 1 1 ·§ 2 −1 1 ·
¨ ¸¨ − ¸
¨ 3 3 3 ¸¨ 3 2 6¸
¨ 1 −1 ¸ ¨ 2 2 ¸
=¨ 0 ¸¨ 0 ¸
¨ 2 2 ¸¨ 3 6 ¸
¨ 1 −2 1 ¸¨ 2 1 −1 ¸
¨ ¸¨ ¸
© 6 6 6 ¹© 3 2 6 ¹
§2 0 0 ·
¨ ¸
= ¨ 0 −1 0 ¸ = D
¨ 0 0 −1¸
© ¹

109
Sathyabama University

Example 8:

§ 3 −1 · 3
Reduce the matrix A = ¨ ¸ to a diagonal form. Hence find A .
© − 1 3 ¹
Solution
§ 3 −1 ·
Let A = ¨ ¸ be a given matrix
© −1 3 ¹
The characteristic equation of A is given by A − λ I = 0

The general form of a characteristic equation is λ 2 − C1λ + C2 = 0

where C1 = 3 + 3 = 6
C2 = A = 9 − 1 = 8

∴ The characteristic equation is λ 2 − 6λ + 8 = 0

( λ − 2 )( λ − 4 ) = 0
The eigen values are λ = 2, 4

The eigen vector corresponding to the eigen value Ȝ is given by

§x ·
( A − λ I ) X = 0 , where X = ¨ 1 ¸ is a column vector
© x2 ¹

§3− λ 1 · § x1 ·
¨ ¸¨ ¸ = 0
© −1 3 − λ ¹ © x2 ¹

i.e., ( 3 − λ ) x1 − x2 = 0½°
¾ (1)
− x1 + ( 3 − λ ) x2 = 0 °¿

Case 1: If Ȝ = 2 then (1) becomes

x1 − x2 = 0
− x1 + x2 = 0

Both the equations are the same.

x1 x2
∴ x1 − x2 = 0 Ÿ x1 = x2 i.e., =
1 1
§1·
∴ The eigen vector corresponding to the eigen value Ȝ = 2 is X 1 = ¨ ¸
©1¹

110
 Sathyabama University

Case 2: If Ȝ = 4 then (1) becomes

− x1 − x2 = 0
− x1 − x2 = 0
Both the equations are the same.
∴ − x1 + x2 = 0 Ÿ x1 = − x2
x1 x2
i.e. =
1 −1
§1·
∴ The eigen vector corresponding to the eigen value Ȝ = 4 is X 2 = ¨ ¸
© −1¹
§1 1 ·
Modal matrix M = ¨ ¸
© 1 −1¹
§ 1 1 ·
¨ ¸
2 2¸
Normalised Modal Matrix N = ¨
¨ 1 −1 ¸
¨ ¸
© 2 2¹
§ 1 1 ·
¨ ¸
T ¨ 2 2¸
N =
¨ 1 −1 ¸
¨ ¸
© 2 2¹
Consider N T AN

§ 1 1 · § 1 1 ·
¨ ¸ § 3 −1· ¨ ¸
2 2 ¸ 2 2 ¸
N T AN = ¨ ¨ ¸¨
¨ 1 1 ¸ © −1 3 ¹ ¨ 1 1 ¸
¨ − ¸ ¨ − ¸
© 2 2¹ © 2 2¹
§ 2 2 ·§ 1 1 ·
¨ ¸¨ ¸
¨ 2 2 ¸¨ 2 2 ¸
=
¨ 4 −4 ¸¨ 1 1 ¸
¨ ¸¨ − ¸
© 2 2 ¹© 2 2¹
§ 2 0·
=¨ ¸=D
© 0 4¹
To find A3 :
We know N T AN = D

Ÿ A = NDN T
A3 = ND3 N T

111
Sathyabama University

§ 1 1 · § 1 1 ·
¨ ¸§8 0 ·¨ ¸
2 2¸ 2 2 ¸
=¨ ¨ ¸¨
¨ 1 −1 ¸ © 0 64 ¹ ¨ 1 1 ¸
¨ ¸ ¨ − ¸
© 2 2¹ © 2 2¹

§ 8 64 · § 1 1 ·
¨ ¸¨ ¸
2 2 ¸¨ 2 2 ¸
=¨
¨ 8 −64 ¸ ¨ 1 1 ¸
¨ ¸¨ − ¸
© 2 2 ¹© 2 2¹
§ 36 −28 ·
=¨ ¸
© −28 36 ¹

EXERCISE

1. Reduce the following real symmetric matrix to a diagonal form and

§ −1 0 2 ·
4 ¨ ¸
hence find A , A = ¨ 0 1 2 ¸
¨ 2 2 0¸
© ¹

§ 3 2 2·
¨ ¸
2. Diagonalise the matrix A = ¨ 2 2 0 ¸ by orthogonal
¨ 2 0 4¸
© ¹
transformation.

§2 0 1·
¨ ¸
3. Diagonalise the Matrix A = ¨ 0 3 0 ¸
¨1 0 2¸
© ¹

§ 2 2 −7 ·
¨ ¸
4. Find the eigen value and eigen vector of the matrix A = ¨ 2 1 2 ¸
¨ 0 1 −3 ¸
© ¹
and hence diagonalise by similarity transformation.

§5 0 1·
¨ ¸
5. Diagonalise the matrix A = ¨ 0 −2 0 ¸
¨1 0 5¸
© ¹

112
 Sathyabama University

6. Diagonalise the given matrix by orthogonal transformation and

§ 6 −2 2 ·
¨ ¸
hence find A3 where A = ¨ −2 3 −1¸
¨ ¸
© 2 −1 3 ¹

§ −1 2 −2 ·
¨ ¸
7. Diagonalise A = ¨ 1 2 1 ¸
¨ −1 −1 0 ¸
© ¹

§0 1 1·
¨ ¸
8. Diagonalise A = ¨ 1 0 1 ¸
¨1 1 0¸
© ¹

§ 10 −2 −5 ·
¨ ¸
9. Diagonalise A = ¨ −2 2 3 ¸
¨ −5 3 5 ¸
© ¹

§ 3 −1 0 ·
¨ ¸
10. Diagonalise A = ¨ −1 2 −1¸
¨ 0 −1 3 ¸
© ¹

§3 1 1 ·
¨ ¸
11. Diagonalise the matrix A = ¨ 1 3 −1¸ by means of orthogonal
¨ 1 −1 3 ¸
© ¹
transformation.
§ 2 1 −1 ·
¨ ¸
12. Diagonalise the matrix A = ¨ 1 1 −2 ¸ by means of orthogonal
¨ −1 −2 1 ¸
© ¹
transformation

§ 11 −4 −7 ·
¨ ¸
13. Diagonalise the matrix A = ¨ 7 −2 −5 ¸ and hence find A5
¨ 10 −4 −6 ¸
© ¹

§ 1 1 1·
¨ ¸
14. Diagonalise the matrix A = ¨ 0 2 1 ¸
¨ −4 4 3 ¸
© ¹

113
Sathyabama University

§ −3 2 2 ·
¨ ¸
15. Diagonalise the matrix A = ¨ −6 5 2 ¸ by similarity
¨ −7 4 4 ¸
© ¹
transformation

ANSWERS

§ 2· §1· § 2·
¨ ¸ ¨ ¸ ¨ ¸
1. λ = 0, − 3, 3 X 1 = ¨ −2 ¸ X 2 = ¨ 2 ¸ X 3 = ¨ 1 ¸
¨1¸ ¨ 2¸ ¨ −2 ¸
© ¹ © ¹ © ¹

§ −2 · § 1· § 2·
2. λ = 0,3, 6 X 1 = ¨¨ 2 ¸¸ X 2 = ¨¨ 2 ¸¸ X 3 = ¨¨ 1 ¸¸
¨1¸ ¨ −2 ¸ ¨ 2¸
© ¹ © ¹ © ¹

§1· §1· §0·

3. λ = 1,3,3 X 1 = ¨ 0 ¸ X 2 = ¨ 0 ¸ X 3 = ¨¨ 1 ¸¸
¨ ¸ ¨ ¸
¨ −1¸ ¨1¸ ¨0¸
© ¹ © ¹ © ¹

§1· § 5· § 3·
4. λ = 1,3, −4 X 1 = ¨ −4 ¸ X 2 = ¨ 6 ¸ X 3 = ¨¨ −2 ¸¸
¨ ¸ ¨ ¸
¨ −1 ¸ ¨1¸ ¨ 2¸
© ¹ © ¹ © ¹

§0· §1· §1·

5. λ = −2, 4, 6 X 1 = ¨ 1 ¸ X 2 = ¨ 0 ¸ X 3 = ¨¨ 0 ¸¸
¨ ¸ ¨ ¸
¨0¸ ¨ −1 ¸ ¨1¸
© ¹ © ¹ © ¹

§2· §0· §1·

6. λ = 2, 2,8 X 1 = ¨ −1¸ X 2 = ¨ 1 ¸ X 3 = ¨¨ 0 ¸¸
¨ ¸ ¨ ¸
¨1¸ ¨1¸ ¨ −2 ¸
© ¹ © ¹ © ¹

§1· § 5 − 1· § 5 + 1·
¨ ¸ ¨ ¸
7. λ = 1, ± 5 X 1 = ¨¨ 0 ¸¸ X 2 = ¨ 1 ¸ X 3 = ¨ −1 ¸
¨ −1¸ ¨ −1 ¸ ¨ 1 ¸
© ¹ © ¹ © ¹

114
 Sathyabama University

§ 1· §1· §1·
8. λ = 2, −1, −1 X 1 = ¨¨1¸¸ X 2 = ¨¨ 0 ¸¸ X 3 = ¨¨ −2 ¸¸
¨ 1¸ ¨ −1¸ ¨1¸
© ¹ © ¹ © ¹

§1· § 1· § −3 ·
9. λ = 0,3,14 X 1 = ¨ −5 ¸ X 2 = ¨1¸ X 3 = ¨¨ 1 ¸¸
¨ ¸ ¨ ¸
¨ 4¸ ¨ 1¸ ¨2¸
© ¹ © ¹ © ¹

§1· §1· §1·

¨ ¸ ¨ ¸ ¨ ¸
10. λ = 1,3, 4 X 1 = ¨ 2 ¸ X 2 = ¨ 0 ¸ X 3 = ¨ −1¸
¨1¸ ¨ −1¸ ¨1¸
© ¹ © ¹ © ¹

§ −1· §1· § −1 ·
¨ ¸ ¨ ¸ ¨ ¸
11. λ = 1, 4, 4 X 1 = ¨ 1 ¸ X 2 = ¨ 1 ¸ X 3 = ¨ 1 ¸
¨1¸ ¨0¸ ¨ −2 ¸
© ¹ © ¹ © ¹

§0· §2· §1·

¨ ¸ ¨ ¸ ¨ ¸
12. λ = −1,1, 4 X 1 = ¨ 1 ¸ X 2 = ¨ −1¸ X 3 = ¨ 1 ¸
¨1¸ ¨1¸ ¨ −1¸
© ¹ © ¹ © ¹

§ 1· §1· § 2·
¨ ¸ ¨ ¸ ¨ ¸
13. λ = 0,1, 2 X 1 = ¨ 1¸ X 2 = ¨ −1¸ X 3 = ¨ 1 ¸
¨ 1¸ ¨2¸ ¨ 2¸
© ¹ © ¹ © ¹

§ −1 · §1· § 1·
¨ ¸ ¨ ¸ ¨ ¸
14. λ = 1, 2,3 X 1 = ¨ −2 ¸ X 2 = ¨ 1 ¸ X 3 = ¨ 1¸
¨ 2¸ ¨0¸ ¨ 1¸
© ¹ © ¹ © ¹

§5· § 2· § 0·
¨ ¸ ¨ ¸ ¨ ¸
15. λ = 0, −1, 2 X 1 = ¨ 1 ¸ X 2 = ¨ 0 ¸ X 3 = ¨ 1 ¸
¨0¸ ¨1¸ ¨ −2 ¸
© ¹ © ¹ © ¹

115
Sathyabama University

QUADRATIC FORMS

Definition: A homogeneous polynomial of the second degree in any number

of variables is called a quadratic form (Q.F).
Example: (i) x 2 + 2 y 2 + 5 xy is a quadratic form in two variables x and y.
(ii) x12 + x22 + x32 + 2 x1 x2 − 3 x1 x3 + x2 x3 is a quadratic form in three
variables x1, x2 and x3

Matrix of a quadratic form:

The general form of a quadratic form in n variables x1, x2…….xn denoted by

Q is given by
Q = f ( x1 , x2 ,! xn ) = c11 x12 + c12 x1 x2 + " + c1n x1 xn
+ c21 x2 x1 + c22 x22 + " + c2 n x2 xn
+ ....................................
+ cn1 xn x1 + cn 2 xn x2 + " + cnn xn2
n n
= ¦¦ cij xi x j
i =1 j −1

In general, cij ≠ c ji . The coefficient of xi x j = cij + c ji

1
Now we define aij =
2
( cij + c ji ) , for all i, j.
Then aii = cii , aij = a ji and aij + a ji = 2aij = cij + c ji
n n
∴ The QF becomes Q = ¦¦ aij xi x j .
i =1 j =1

It can be written in the matrix form as Q = X T A X ,

§ a11 a12 ... a1n ·

¨ ¸
a a22 ... a2 n ¸
where A = ¨ 21 is a symmetric matrix,
¨ # ¸
¨ ¸
© an1 an 2 ... ann ¹

116
 Sathyabama University

§ x1 ·
¨ ¸
x
X = ¨ 2 ¸ and X T = ( x1 x2 … xn )
¨# ¸
¨ ¸
© xn ¹
Here A is called the matrix of the quadratic form.
Note: aii = coefficient of xi2 in the quadratic form.

1
aij = a ji = ª coefficient of xi x j º¼ in the QF
2¬

Example 1

Write down the matrix of the QF: 10 x12 + 2 x22 + 5 x32 + 6 x2 x3 − 10 x3 x1 − 4 x1 x2 .

Solution

This quadratic form contains 3 variables x1, x2 and x3.

The matrix to be obtained is A which has 3 rows and 3 columns.

§ x1 · § a11 a12 a13 ·

¨ ¸ ¨ ¸
X = ¨ x2 ¸ , X T = ( x1 x2 x3 ) , A = ¨ a21 a22 a23 ¸
¨x ¸ ¨a a33 ¸¹
© 3¹ © 31 a32

Here a11 = 10 a22 = 2 a33 = 5

1
a12 = a21 = ( −4 ) = −2
2
1
a23 = a32 = ( 6 ) = 3
2
1
a13 = a31 = ( −10 ) = −5
2
ª10 −2 −5º
∴ A = «« −2 2 3 »»
«¬ −5 3 5 »¼

Example 2

Write the matrix of the QF 3 x12 + 3 x22 + 3 x32 + 2 x1 x2 + 2 x1 x3 − 2 x3 x2

117
Sathyabama University

Solution

a11 = 3 a22 = 3 a33 = 3

1
a12 = a21 = ( 2) = 1
2
1
a23 = a32 = ( −2 ) = −1
2
1
a13 = a31 = ( 2 ) = 1
2

ª3 1 1 º
A = ««1 3 −1»»
«¬1 −1 3 »¼

Example 3
Write the quadratic form corresponding to the symmetric matrix
ª 0 −1 2 º
A = «« −1 1 4 »»
«¬ 2 4 3 »¼

Solution: The QF corresponding to the symmetric matrix A is given by

ª 0 −1 2 º ª x1 º
Q = X AX = [ x1 x2 x3 ] «« −1 1 4 »» «« x2 »»
T

«¬ 2 4 3 »¼ «¬ x3 »¼
= 0 x12 + x22 + 4 x32 − x1 x2 + 2 x3 x1 − x2 x1 + 4 x3 x2 + 2 x1 x3 + 4 x2 x3
= 0 x12 + x22 + 4 x32 − 2 x1 x2 + 4 x1 x3 + 8 x2 x3

Example 4

ª1 0 0 º
Write the quadratic form to the matrix A = ««0 3 −1»» .
«¬0 −1 3 »¼

Solution
The QF corresponding to the symmetric matrix A is given by Q = X T AX .

118
 Sathyabama University

ª1 0 0 º x1
i.e., Q = [ x1 x2 x3 ] «« 0 3 −1»» x2
«¬ 0 −1 3 »¼ x3
= x12 + 3 x22 + 3 x32 − 2 x2 x3

Canonical form (CF)

Let P be a matrix chosen in such way that P T AP = D , a diagonal matrix
§ λ1 0 ! 0 ·
¨ ¸
¨ 0 λ2 ! 0 ¸
of the form Then the linear transformation
¨# # # ¸
¨ ¸
© 0 0 ! λn ¹
X = PY reduces the quadratic form Q = X T AX to the form
Q = Y T DY .

§ λ1 0 ... 0 · ª y1 º
¨ ¸
0 λ2 ... 0 ¸ «« y2 »»
i.e., Q = [ y1 y2 " yn ] ¨
¨# ¸« # »
¨ ¸« »
©0 0 ... λn ¹ ¬ yn ¼

i.e., Q = λ1 y12 + λ2 y22 + .... + λn yn2 (1)

Equation (1) is called the canonical form of the quadratic form.

Reduction of a quadratic form to canonical form through orthogonal

transformation

A quadratic form Q = X T AX is reduced to a canonical form Y T DY by an

orthogonal transformation, if in the linear transformation X = PY , P is an
orthogonal matrix. Let P = N , the normalized modal matrix of A, which is an
orthogonal matrix. Thus the orthogonal transformation X = NY , where N is a
modal matrix of A, reduces a quadratic form Q = X T AX into the canonical
form Y T DY where D is a diagonal matrix with the eigen values of A as its
diagonal elements.

119
Sathyabama University

Nature of quadratic form

The reduction of quadratic form Q = X T AX to canonical form will contain r

terms if the rank of A is r. The terms in the CF may be positive, negative or
zero.

Index:

The number of positive square terms in the canonical form is called the index
of the quadratic form and is denoted by p.

Signature:

The excess number of positive square terms over the negative square terms of
the canonical form is called the signature of the quadratic form and is denoted
by s.
i.e., s = p − ( r − p ) = 2 p − r

Nature of the quadratic form:

A quadratic form Q = X T AX is said to be

(i) Positive definite: if r = n and p = n (or) if all the eigen values of A are
positive.
(ii) Positive semi definite: If r < n and p = r (or) if all the eigen values of
A • 0 and atleast one eigen value is zero.
(iii) Negative definite: If r = n and p = 0 (or) all the eigen values of A are
negative.
(iv) Negative semi definite: If r < n and p = r (or) if all the eigen values
of A ≤ 0 and atleast one eigen value is zero.
(v) Indefinite: For all other cases (or) if A has positive as well as negative
eigen values.

Working rule to reduce a quadratic form to canonical form:

Step 1: Write the matrix of the QF ‘A’.

Step2: Find the characteristic equation.
Step3: Find the eigen values of A.
Step4: Find the eigen vectors which are pairwise orthogonal.

120
 Sathyabama University

Step5: Form normalized modal matrix N.

Step6: Find NT
Step7: Calculate N T AN = D
ª y1 º
Step8: Canonical form of the QF is given by Y DY , where Y = «« y2 »»
T

«¬ y3 ¼»

Example 1

Reduce the quadratic form Q : 3 x12 + 2 x22 + 3 x32 − 2 x1 x2 − 2 x2 x3 into canonical

form using orthogonal transformation. Also find its rank, index, signature and
nature.

Solution

The matrix of the quadratic form is given by

ª 3 −1 0 º
A = «« −1 2 −1»»
«¬ 0 −1 3 »¼

The characteristic equation of A is given by A – λ I = 0

The general form of a characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0

where C1 = 3 + 2 + 3 = 8

2 −1 3 0 3 −1
C2 = + + = 19
−1 3 0 3 −1 2
C3 = | A | = 12

∴ The characteristic equation is λ 3 − 8λ 2 + 19λ − 12 = 0

1 −8 19 −12
1
0 1 −7 12
1 − 7 12 0

Ȝ = 1 is a root and λ 2 − 7λ + 12 = 0

i.e. ( λ − 3)( λ − 4 ) = 0

121
Sathyabama University

i.e. λ = 3, 4

∴ Ȝ = 1, 3, 4 are the eigen values of A.

The eigen vector corresponding to the eigen value Ȝ is given by
ª x1 º
[ A − λ I ] X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª3 − λ −1 0 º ª x1 º ª 0 º
i.e. «« −1 2 − λ −1 »» «« x2 »» = «« 0 »»
«¬ 0 −1 3 − λ »¼ «¬ x3 »¼ «¬ 0 »¼

( 3 − λ ) x1 − x2 + 0 x3 = 0 ½
°
i.e. − x1 + ( 2 − λ ) x2 − x3 = 0 ¾ (1)
°
0 x1 − x2 + ( 3 − λ ) x3 = 0 ¿

Case 1: If Ȝ=1 then (1) becomes

2 x1 − x2 + 0 x3 = 0
− x1 + x2 − x3 = 0
0 x1 − x2 + 2 x3 = 0

Using the rule of cross multiplication for first two distinct equations we get
x1 − x2 x3
= =
−1 0 2 0 2 −1
−1 − 1 −1 −1 −1 −1

x1 x2 x3
i.e., = =
1 2 1

ª1 º
∴ Eigen vector corresponding to Ȝ = 1 is X 1 = «« 2 »»
«¬ 1 »¼

Case 2: If Ȝ = 3, then (1) becomes

0 x1 − x2 + 0 x3 = 0
− x1 − x2 − x3 = 0
0 x1 − x2 + 0 x3 = 0

122
 Sathyabama University

Using the rule of cross multiplication for first two distinct rows we get
x1 − x2 x3
= =
−1 0 0 0 0 −1
−1 −1 −1 −1 −1 −1
x2 x2 x3
i.e., = =
1 0 −1
ª1º
∴ Eigen vector corresponding to Ȝ = 3 is X 2 = «« 0 »»
«¬ −1»¼
Case 3: If Ȝ = 4 then (1) becomes
− x1 − x2 + 0 x3 = 0
− x1 − 2 x2 − x3 = 0
0 x1 − x2 − x3 = 0

Using the rule of cross multiplication for first two district rows we get,
x1 − x2 x3
= =
−1 0 −1 0 −1 −1
−2 −1 −1 −1 −1 −2

x1 x2 x3
i.e., = =
1 −1 1

ª1º
∴ Eigen vector corresponding to Ȝ = 4 is X 3 = «« −1»»
«¬ 1 »¼

ª1 1 1 º
Modal matrix M = «« 2 0 −1»»
«¬1 −1 1 »¼

ª 1 1 1 º
« »
« 6 2 3 »
« 2 1 »
Normalised modal matrix N = « 0 − »
« 6 3»
« 1 −1 1 »
« »
¬ 6 2 3 ¼

123
Sathyabama University

ª 1 2 1 º
« »
« 6 6 6 »
« 1 1 »
NT = « 0 − »
« 2 2»
« 1 −1 1 »
« »
¬ 3 3 3 ¼

Consider N T AN

ª 1 2 1 º ª 1 1 1 º
« » « »
« 6 6 6 » 3 −1 0 « 6 2 3»
ª º
«
=«
1 −1 » « »« 2 −1 »
0 » « −1 2 −1» « 0 »
2 2» 6 3»
« « 0 −1 3 ¼» «
¬
« 1 −1 1 » « 1 −1 1 »
« » « »
¬ 3 3 3¼ ¬ 6 2 3¼
ª 1 2 1 ºª 1 3 4 º
« »« »
« 6 6 6»« 6 2 3»
« 1 −1 » « 2 −4 »
=« 0 »« 0 »
« 2 2»« 6 3»
« 1 −1 1 »« 1 −3 4 »
« »« »
¬ 3 3 3¼¬ 6 2 3¼
ª1 0 0 º
= ««0 3 0 »» = D ( say )
¬«0 0 4 ¼»

ª y1 º
Consider the linear transformation X = NY where Y = «« y2 »» is a column
«¬ y3 »¼
vector.

ª 1 1 1 º
« »
« 6 2 3 » y
ª 1º
x ª 1º
« 1 »« »
i.e., « x2 »» = «
« 2
0 − » y2
6 3»« »
«¬ x3 »¼ « «y »
« 1 −1 1 »¬ 3¼
« »
¬ 6 2 3 ¼

124
 Sathyabama University

y1 y2 y3
i.e., x1 = + +
6 2 3
2 y1 y3
x2 = + 0 y2 −
6 3
y1 y2 y3
x3 = − +
6 2 3
The linear transformation X = NY reduces the quadratic form Q = XTAX into
the canonical form YTDY.
ª1 0 0 º ª y1 º
Consider Y T DY = [ y1 y2 y3 ] ««0 3 0 »» «« y2 »»
«¬0 0 4 »¼ «¬ y3 »¼

= y12 + 3 y22 + 4 y32 is the canonical form of the given QF.

As the number of terms in the canonical form is three rank of the

quadratic form ‘r’ = 3. Since all the three terms of the CF is positive, index of
the QF ‘p’ = 3
Hence signature of the quadratic form ‘s’ =2p – r =2(3)-3=3
Nature of the QF is positive definite.

Example 2

Reduce the Quadratic form x12 + 2 x22 + x32 − 2 x1 x2 + 2 x2 x3 to the canonical form
through an orthogonal reduction and hence show that it is positive semi
definite. Also give a non zero set of values of (x1, x2, x3) which makes this
quadratic form zero.

Solution

Let the matrix of the quadratic form be

ª 1 −1 0 º
A = «« −1 2 1 »»
«¬ 0 1 1 »¼

The characteristic equation of A is given by A − λ I = 0.

The general form of a characteristic equation is
λ 3 − C1λ 2 + C2 λ − C3 = 0

125
Sathyabama University

where C1 =1 + 2 + 1 = 4
2 1 1 0 1 −1
C2 = + + =3
1 1 0 1 −1 2
C3 = | A |= 0.

∴ The characteristic equation of A is λ 3 − 4λ 2 + 3λ = 0

i.e., λ ª¬λ 2 − 4λ + 3º¼ = 0

i.e., Ȝ [ Ȝ -1][ Ȝ - 3] = 0

i.e., λ = 0,1,3 are the eigen values of A.

The eigen vector corresponding to the eigen value Ȝ is given

ª x1 º
by [ A − λ I ] X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª1 − λ −1 0 º ª x1 º ª 0 º
i.e., «« −1 2−λ 1 »» «« x2 »» = «« 0 »»
¬« 0 1 1 − λ »¼ «¬ x3 »¼ «¬ 0 »¼

i.e., (1 − λ ) x1 − x2 + 0 x3 = 0 ½
°
− x1 + ( 2 − λ ) x2 + x3 = 0 ¾ (1).
°
0 x1 + x2 + (1 − λ ) x3 = 0 ¿

case1: If Ȝ = 0, then (1) becomes

x1 − x2 + 0 x3 = 0
− x1 + 2 x2 + x3 = 0
0 x1 + x2 + x3 = 0

Using the rule of cross multiplication for the first two distinct equations,
we get
x1 − x2 x3
= =
−1 0 1 0 1 −1
2 1 −1 1 −1 2

x1 x2 x3
i.e., = =
−1 −1 1

126
 Sathyabama University

ª1º
∴ Eigen vector corresponding to Ȝ = 0 is X 1 = «« 1 »»
«¬ −1»¼

Case 2: If Ȝ = 1, then (1) becomes

0 x1 − x2 + 0 x3 = 0
− x1 + x2 + x3 = 0
0 x1 + x2 + 0 x3 = 0

Using the rule of cross multiplication, for the first two distinct equations,
we get
x1 − x2 x3
= =
−1 0 0 0 0 −1
1 1 −1 1 −1 1

x1 x2 x3
i.e., = =
−1 0 −1

ª1 º
∴ The eigen vector corresponding to the eigen value λ = 1 is X 2 = «« 0 »»
«¬1 »¼

Case 3: If Ȝ = 3, then (1) becomes

−2 x1 − x2 + 0 x3 = 0
− x1 − x2 + x3 = 0
0 x1 + x2 − 2 x3 = 0

Using the rule of cross multiplication for the first two distinct equations, we
get
x1 − x2 x3
= =
−1 0 −2 0 −2 −1
−1 1 −1 1 −1 −1

x1 x2 x3
i.e., = =
−1 2 1

ª −1º
∴ The eigen vector corresponding to the eigen value Ȝ = 3 is X 3 = «« 2 »»
«¬ 1 »¼

127
Sathyabama University

ª 1 1 −1º
Modal Matrix M = «« 1 0 2 »»
«¬ −1 1 1 »¼

ª 1 1 −1 º
« »
« 3 2 6»
« 1 2 »
Normalised Modal Matrix N = « 0 »
« 3 6»
« −1 1 1 »
« »
¬ 3 2 6¼

ª 1 1 −1 º
« »
« 3 3 3»
« 1 1 »
NT = « 0 »
« 2 2»
« −1 2 1 »
« »
¬ 6 6 6¼

Consider N T AN

ª 1 1 −1 º ª 1 1 −1 º
« » « »
« 3 3 3»
ª 1 −1 0 º«
3 2 6»
« 1
N T AN = «
1 »« »« 1 2 »
0 » « −1 2 1 » « 0 »
« 2 2»
«¬ 0 1 1 »¼ «
3 6»
« −1 2 1 » « −1 1 1 »
« » « »
¬ 6 6 6¼ ¬ 3 2 6¼

ª 1 1 −1 º ª 1 −3 º
« » «0 »
« 3 3 3»« 2 6»
« 1 1 »« 6 »
=« 0 » «0 0 »
« 2 2»« 6»
« −1 2 1 »« 1 3 »
« » «0 »
¬ 6 6 6¼¬ 2 6¼
ª0 0 0º
= ««0 1 0 »» = D ( say )
«¬0 0 3 »¼

128
 Sathyabama University

ª y1 º
Consider the linear transformation X = NY where Y = «« y2 »»
«¬ y3 »¼

ª 1 1 −1 º
« »
3 2 6» y
ª x1 º « ª 1º
« » « 1
i.e., « x2 » = «
2 »« »
0 » y2
3 6»« »
«¬ x3 »¼ « «y »
« −1 1 1 »¬ 3¼
« »
¬ 3 2 6¼

y1 y2 y3 ½
i.e., x1 = + − °
3 2 6 °
y1 2 y3 °
x2 = + ¾ (2)
3 6 °
− y1 y2 y °
x3 = + + 3°
3 2 6¿
The linear transformation X = NY, reduces the given quadratic form
Q = XTAX to the canonical form YTDY.

ª0 0 0 º ª y1 º
Consider Y T DY = [ y1 y2 y3 ] ««0 1 0 »» «« y2 »»
«¬0 0 3 »¼ «¬ y3 »¼

= 0. y12 + y22 + 3 y32 is the canonical form of the given QF.

Nature of the quadratic form is positive semi definite.

Taking y1 = 3, y2 = 0 and y3 = 0 in (2), we get

x1 = 1, x2 = 1, x3 = -1.

These values of x1, x2, x3 make the Q.F zero.

Example 3

Reduce the quadratic form 2x1x2 + 2x1x3 - 2x2x3 to the canonical form by
orthogonal transformation. Also find the rank, index, signature and nature of
the quadratic form.

129
Sathyabama University

Solution

The matrix of the quadratic form is given by

ª0 1 1 º
A = ««1 0 −1»»
«¬1 −1 0 »¼
The characteristic equation of A is given by |A – ȜI| = 0
The general form of a characteristic equation is λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 0 + 0 + 0 = 0

0 −1 0 1 0 1
C2 = + + = −3
−1 0 1 0 1 0
C3 =| A |= −2

The characteristic equation of A is given by

λ 3 − 3λ + 2 = 0
1 0 −3 2
1
0 1 1 −2
1 1 −2 0

∴ λ = 1 is a root and λ 2 + λ − 2 = 0

i.e., ( λ − 1)( λ + 2 ) = 0

i.e.. λ = 1, −2
∴ The eigen values of A are 1, 1, -2.
The eigen vector corresponding to the eigen value Ȝ is given by
ª x1 º
( A − λ I ) X = 0, where X = «« x2 »» is a column vector.
«¬ x3 »¼

ª0 − λ 1 1 º ª x1 º ª 0 º
i.e., «« 1 0−λ −1 »» «« x2 »» = «« 0 »»
«¬ 1 −1 0 − λ »¼ «¬ x3 »¼ «¬ 0 »¼

i.e., − λ x1 + x2 + x3 = 0 ½
°
x1 − λ x2 − x3 = 0 ¾ (1)
x1 − x2 − λ x3 = 0 °¿

130
 Sathyabama University

Case 1: If Ȝ = –2, then (1) becomes

2 x1 + x2 + x3 = 0
x1 + 2 x2 − x3 = 0
x1 − x2 + 2 x3 = 0

Using the rule of cross multiplication for the first two distinct equations,
we get
x1 − x2 x3
= =
1 1 2 1 2 1
2 −1 1 −1 1 2

x1 x2 x3
i.e. = =
−3 3 3
x1 x2 x3
i.e. = =
−1 1 1
ª −1º
∴ The eigen vector corresponding to the eigen value Ȝ = –2 is X 1 = «« 1 »»
«¬ 1 »¼

Case 2: If Ȝ = 1, then (1) becomes

− x1 + x2 + x3 = 0
x1 − x2 − x3 = 0
x1 − x2 − x3 = 0

All the three equations are the same.

Put x2 = 0. Then –x1 + x3 = 0 Ÿ x1 = x3
x1 x3
i.e. = .
1 1
ª1 º
∴ Eigen vector corresponding to Ȝ = 1 is X 2 = «« 0 »» .
«¬1 »¼

Since all the three eigen vectors should be orthogonal to each other, we
assume the third eigen vector
§a·
¨ ¸
X 3 = ¨ b ¸ to be orthogonal to X1 and X2.
¨c¸
© ¹

131
Sathyabama University

X 1T X 3 = 0 and X 2 T X 3 = 0

§a· §a·
¨ ¸ ¨ ¸
i.e., (–1 1 1) ¨ b ¸ = 0 and (1 0 1) ¨b¸ = 0
¨c¸ ¨c¸
© ¹ © ¹
i.e., − a + b + c = 0
a + 0b + c = 0
Using the rule of cross multiplication, we get,
a −b c
= =
1 1 −1 1 −1 1
0 1 1 1 1 0

a b c
i.e. = =
1 2 −1
∴ The eigen vector which is orthogonal to both X1 & X2 is given by
§1·
¨ ¸
X3 = ¨ 2 ¸
¨ −1¸
© ¹

§ −1 1 1 ·
¨ ¸
Modal matrix M = ¨ 1 0 2 ¸
¨ 1 1 −1¸
© ¹

§ −1 1 1 ·
¨ ¸
¨ 3 2 6¸
¨ 1 2 ¸
Normalised modal matrix N = ¨ 0 ¸
¨ 3 6¸
¨ 1 1 −1 ¸
¨ ¸
© 3 2 6¹

§ −1 1 1 ·
¨ ¸
¨ 3 3 3¸
¨ 1 1 ¸
NT = ¨ 0 ¸
¨ 2 2¸
¨ 1 2 −1 ¸
¨ ¸
© 6 6 6¹

132
 Sathyabama University

Consider N T AN

§ −1 1 1 · § −1 1 1 ·
¨ ¸ ¨ ¸
¨ 3 3 3¸ 0 1 1 ¨ 3 2 6¸
ª º
¨
N T AN = ¨
1 1 ¸«
− »¨ 1 2 ¸
0 ¸
2 ¸«
1 0 1»¨ 3 0 ¸
«¬1 −1 0 »¼ ¨
¨ 2 6¸
¨ 1 2 −1 ¸ ¨ 1 1 −1 ¸
¨ ¸ ¨ ¸
© 6 6 6¹ © 3 2 6¹
§ −1 1 1 ·§ 2 1 1 ·
¨ ¸¨ ¸
¨ 3 3 3 ¸¨ 3 2 6¸
¨ 1 1 ¸ ¨ −2 2 ¸
=¨ 0 ¸¨ 0 ¸
¨ 2 2 ¸¨ 3 6¸
¨ 1 2 −1 ¸ ¨ −2 1 −1 ¸
¨ ¸¨ ¸
© 6 6 6 ¹© 3 2 6¹
ª −2 0 0 º
= «« 0 1 0 »» = D( say )
«¬ 0 0 1 »¼

ª y1 º
Consider the linear transformation X = NY, where Y = «« y2 »» .
«¬ y3 »¼

y1 y2 y3
i.e., x1 = − + +
3 2 6
y1 2 y3
x2 = +
3 6
y1 y2 y3
x3 = + −
3 2 6
This linear transformation X = NY reduces the QF: Q = X T AX into the
canonical form YTDY.

ª −2 0 0 º ª y1 º
Consider Y DY = [ y1
T
y2 y3 ] «« 0 1 0 »» «« y2 »»
«¬ 0 0 1 »¼ «¬ y3 »¼

= −2 y12 + y22 + y32 is the canonical form of the given QF.

133
Sathyabama University

Rank of the QF = 3
Index of the QF = 2
Signature of the QF = 2(2) – 3 = 1
Nature of the QF = Indefinite

Example 4
Reduce the quadratic form Q = 6 x 2 + 3 y 2 + 3 z 2 − 4 xy − 2 yz + 4 xz into
canonical form by an orthogonal reduction and also find rank, index,
signature and nature of the quadratic form

Solution
The matrix of the quadratic form is given by

ª 6 −2 2 º
A = «« −2 3 −1»»
«¬ 2 −1 3 »¼

The characteristic equation of A is given by |A – ȜI| = 0.

The general form of a characteristic equation is
λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 6 + 3 + 3 = 12

3 −1 6 2 6 −2
C2 = + + = 36
−1 3 2 3 −2 3
C3 = | A | = 32

∴ The characteristic equation of A is given by

λ 3 − 12λ 2 + 36λ − 32 = 0
1 −12 36 −32
2
0 2 −20 32
1 − 10 16 0

Ȝ = 2 is a root and λ2 − 10λ + 16 = 0

i.e. ( λ − 2 )( λ − 8) = 0

i.e. λ = 2,8
∴ The eigen values of A are 2, 2, 8.

134
 Sathyabama University

The eigen vector corresponding to the eigen value Ȝ is given by (A – ȜI)X = 0,

§ x1 ·
¨ ¸
where X = ¨ x2 ¸ is a column vector.
¨x ¸
© 3¹

ª6 − λ −2 2 º ª x1 º ª0º
« »« » « »
i.e., « −2 3 − λ −1 » « x2 » = «0»
«¬ 2 −1 3 − λ »¼ «¬ x3 »¼ «¬0»¼

i.e., (6 − λ ) x1 − 2 x2 + 2 x3 = 0½
°
− 2 x1 + (3 − λ ) x2 + x3 = 0¾ (1)
2 x1 − x2 + (3 − λ ) x3 = 0 °
¿
Case 1: If Ȝ = 8, then (1) becomes

2 x1 − 2 x2 + 2 x3 = 0 ½
°
−2 x1 − 5 x2 − x3 = 0 ¾
°
2 x1 − x2 − 5 x3 = 0 ¿
Using the rule of cross multiplication for first two distinct equations, we get
x1 − x2 x3
= =
−2 2 −2 2 −2 −2
−5 −1 −2 −1 −2 −5

x1 x2 x3
i.e., = =
12 −6 6
x1 x2 x3
i.e., = =
2 −1 1

§2·
¨ ¸
∴ Eigen vector corresponding to the eigen value Ȝ = 8 is X 1 = ¨ −1¸
¨1¸
© ¹
Case 2: If Ȝ = 2, then (1) becomes
4 x1 − 2 x2 + 2 x3 = 0
− 2 x1 + x2 − x3 = 0
2 x1 − x2 + x3 = 0

135
Sathyabama University

All the three equations are equivalent to a single equation

2 x1 − x2 + x3 = 0

Put x1 = 0. Then − x2 + x3 Ÿ x2 = x3

x2 x3
i.e., =
1 1

§0·
¨ ¸
∴ Eigen vector corresponding to λ = 2 is X 2 = ¨ 1 ¸
¨1¸
© ¹
Since all the three eigen vectors should be orthogonal to each other, we
§a·
¨ ¸
assume the third eigen vector X 3 = ¨ b ¸ to be orthogonal to X1 and X2.
¨c¸
© ¹

∴ X 1T X 3 = 0 and X 2T X 3 = 0

§a· §a·
¨ ¸ ¨ ¸
i.e., ( 2 −1 1) ¨ b ¸ = 0 and ( 0 1 1) ¨ b ¸ = 0
¨c¸ ¨c¸
© ¹ © ¹
i.e., 2 a − b + c = 0
0a + b + c = 0

Using the rule of cross multiplication, we get

a −b c
= =
−1 1 2 1 2 −1
1 1 0 1 0 1

a b c
i.e., = =
−2 −2 2
a b c
i.e., = =
1 1 −1
∴ The third eigen vector which is orthogonal to X1, and X2 is given by
§1·
¨ ¸
X3 = ¨ 1 ¸
¨ ¸
© −1¹

136
 Sathyabama University

ª2 0 1º
« »
Modal matrix M = « −1 1 1 »
¬« 1 1 −1¼»

ª 2 1 º
« 0 »
« 6 3»
« −1 1 1 »
Normalised Modal Matrix N = « »
« 6 2 3»
« 1 1 −1 »
« »
¬ 6 2 3¼

ª 2 −1 1 º
« »
« 6 6 6»
T « 1 1 »
N =« 0 »
« 2 2»
« 1 1 −1 »
« »
¬ 3 3 3¼

Consider N T AN

ª 2 −1 1 º ª 2 1 º
« » « 0 »
« 6 6 6 » 6 −2 2 « 6 3»
ª º
«
N T AN = «
1 1 »« » « −1 1 1 »
0 » « −2 3 −1» « »
2 2» 6 2 3»
« «¬ 2 −1 3 »¼ «
« 1 1 −1 » « 1 1 −1 »
« » « »
¬ 3 3 3¼ ¬ 6 2 3¼
ª 2 −1 1 ºª 8 2 º
« »« 0 »
« 6 6 6»« 6 3»
« 1 1 » « −8 2 2 »
=« 0 »« »
« 2 2»« 6 2 3»
« 1 1 −1 » « 8 2 −2 »
« »« »
¬ 3 3 3¼¬ 6 2 3¼
ª8 0 0º
= ««0 2 0 »» = D ( say )
«¬0 0 2 »¼

137
Sathyabama University

ª y1 º
« »
Let Y = « y2 » . Then the linear transformation X = NY transforms the
«¬ y3 »¼
quadratic form Q = XTAX into canonical form YTDY.
§ 8 0 0 · § y1 ·
¨ ¸¨ ¸
Consider Y T DY = ( y1 y2 y3 ) ¨ 0 2 0 ¸ ¨ y2 ¸ = 8 y12 + 2 y22 + 2 y32
¨ 0 0 2¸ ¨ y ¸
© ¹© 3¹

The canonical form of the QF is 8 y12 + 2 y2 2 + 2 y32 of the given QF

Rank = 3
Index = 3
Signature = 3
Nature of the quadratic form = Positive definite.

Example 5

Find the canonical form of the quadratic form

2 2 2
2 x + y + z + 2 xy − 2 xz − 4 yz. Also find the nature of the quadratic form.

Solution

The matrix of the quadratic form is given by

ª 2 1 −1º
A = «« 1 1 −2»»
«¬ −1 −2 1 »¼

The characteristic equation of A is given by |A – ȜI| = 0.

The general form of a characteristic equation is

λ 3 − C1λ 2 + C2 λ − C3 = 0
where C1 = 2 + 1 + 1 = 4

1 −2 2 −1 2 1
C2 = + + = −1
−2 1 −1 1 1 1
C3 = | A | = −4

∴ The characteristic equation of A is

138
 Sathyabama University

λ 3 − 4λ 2 − λ + 4 = 0
1 −4 −1 4
1
0 1 −3 −4
1 −3 −4 0

∴ Ȝ = 1 is a root and λ 2 − 3λ − 4 = 0
i.e. ( λ − 4 )( λ + 1) = 0

i.e. λ = −1, 4
∴ Eigen values of A are 1, -1 and 4.
The eigen vector corresponding to the eigen value Ȝ is given by (A – ȜI)X = 0
§ x1 ·
¨ ¸
where X = ¨ x2 ¸ is a column vector.
¨x ¸
© 3¹

ª2 − λ 1 −1 º ª x1 º ª0º
« 1 − λ −2 »» «« x2 »» = ««0»»
i.e., « 1
«¬ −1 −2 1 − λ ¼» ¬« x3 ¼» ¬«0¼»

i.e., (2 − λ ) x1 + x2 − x3 = 0 ½
°
x1 + (1 − λ ) x2 − 2 x3 = 0 ¾ (1)
− x1 − 2 x2 + (1 − λ ) x3 = 0 °
¿
Case 1: If Ȝ =1, then (1) becomes
x1 + x2 − x3 = 0
x1 + 0 x2 − 2 x3 = 0
− x1 − 2 x2 + 0 x3 = 0

Using the rule of cross multiplication for the first two distinct equations, we
get
x1 − x2 x
= = 3
1 −1 1 −1 1 1
0 −2 1 −2 1 0

x1 x2 x3
i.e.. = =
−2 1 −1

139
Sathyabama University

§2·
¨ ¸
∴ The eigen vector corresponding to Ȝ = 1 is X 1 = ¨ −1¸
¨1¸
© ¹
Case 2: If Ȝ = -1, the (1) becomes
3x1 + x2 − x3 = 0
x1 + 2 x2 − 2 x3 = 0
− x1 − 2 x2 + 2 x3 = 0

Using the rule of cross multiplication for the first two distinct equations, we
get
x1 − x2 x3
= =
1 −1 3 −1 3 1
2 −2 1 −2 1 2

x1 x2 x3
i.e., = =
0 5 5

§0·
¨ ¸
∴ The eigen vector corresponding to Ȝ = -1 is X 2 = ¨ 1 ¸
¨1¸
© ¹
Case 3: If Ȝ = 4, then (1) becomes
− 2 x1 + x2 − x3 = 0
x1 − 3x2 − 2 x3 = 0
− x1 − 2 x2 − 3x3 = 0

Using the rule of cross multiplication for the first two distinct equations, we
get
x1 − x2 x3
= =
1 −1 −2 −1 −2 1
−3 −2 1 −2 1 −3

x1 x x
i.e., = 2 = 3
−5 −5 5

§1·
¨ ¸
∴ The eigen vector corresponding to Ȝ = 4 is X 3 = ¨ 1 ¸
¨ −1¸
© ¹

140
 Sathyabama University

ª2 0 1º
« »
Modal Matrix M = « −1 1 1 »
«¬ 1 1 −1»¼

ª 2 1 º
« 0 »
« 6 3»
« −1 1 1 »
Normalised Modal Matrix N = « »
« 6 2 3»
« 1 1 −1 »
« »
¬ 6 2 3¼

ª 2 −1 1 º
« »
« 6 6 6»
T « 1 1 »
N =« 0 »
« 2 2»
« 1 1 −1 »
« »
¬ 3 3 3¼

Consider N T AN

ª 2 −1 1 º ª 2 1 º
« » « 0 »
« 6 6 6 » 2 1 −1 « 6 3»
ª º
«
N T AN = «
1 1 »« » « −1 1 1 »
0 » « 1 1 −2 » « »
2 2» 6 2 3»
« «¬ −1 −2 1 ¼» «
« 1 1 −1 » « 1 1 −1 »
« » « »
¬ 3 3 3¼ ¬ 6 2 3¼
ª 2 −1 1 ºª 2 4 º
« »« 0 »
« 6 6 6»« 6 3»
« 1 1 » « −1 1 4 »
=« 0 »« »
« 2 2»« 6 2 3»
« 1 1 −1 » « 1 −1 −4 »
« »« »
¬ 3 3 3¼¬ 6 2 3¼
ª1 0 0 º
= ««0 −1 0 »» = D ( say )
«¬0 0 4 »¼

141
Sathyabama University

ª y1 º
« »
Let Y = « y2 » . Then the linear transformation X = NY transforms the
«¬ y3 »¼
quadratic form Q = XTAX into canonical form YTDY.
§ ·
0 · ¨ y1 ¸
§1 0
¨¸
Consider Y T DY = y y ( −1 0 ¸ ¨ y ¸ = y12 − y 2 + 4 y 2 is
)
y ¨0
1 2 3 ¨ 2¸ 2 3
¨0
0 4 ¸¹ ¨¨ y ¸¸
©
© 3¹
the canonical form of the given QF. Nature of the quadratic form is
indefinite.

Nature of the Quadratic form using Determinants:

ª a11 a12 ..... a1n º

«a a22 ..... a2 n »»
Let A = « 21 be the matrix of quadratic form.
«...... ..... ..... ..... »
« »
¬ an1 an 2 ..... ann ¼
a a12
Let D1 =| a11 |= a11 D2 = 11
a21 a22
a11 a12 a13
D3 = a21 a22 a23 etc.
a31 a32 a32
and Dn = |A|

D1, D2, …. Dn are called the principal sub-determinants (or) principal minors
of A.

The quadratic form is

(i) positive definite if D1, D2,….Dn are all positive i.e., Dn > 0 for all ‘n’
(ii) negative definite if D1, D3, D5 …. are all negative and D2, D4, D6 …. are
all positive i.e., (-1)n Dn > 0 for all ‘n’.
(iii) positive semidefinite if Dn • 0 and at least one Di = 0
n
(iv) negative semidefinite if ( –1) Dn ≥ 0 and atleast one Di = 0

(v) Indefinite in all other cases.

142
 Sathyabama University

Example 1

Determine the nature of the quadratic form x 2 + 3 y 2 + 6 z 2 + 2 xy + 2 yz + 4 xz

Solution

ª1 1 2 º
«
The matrix of the quadratic form is A = «1 3 1 »
»
«¬ 2 1 6 »¼
1 1
Now D1 = 1 =1 D2 = =2 D3 = | A | = 3
1 3
Here D1, D2, D3 are all positive
∴ The quadratic form is positive definite.

Example 2

Determine the nature of quadratic form

2 2 2
10 x1 + 2 x2 + 5 x3 + 6 x2 x3 − 10 x3 x1 − 4 x1 x2

Solution

Let the matrix of the quadratic form is

ª10 −2 −5º
A = «« −2 2 3 »»
«¬ −5 3 5 »¼

10 −2
Now D1 =|10 |= 10 D2 = = 36 D3 = A = 0
−2 2
Here D1, D2 are positive and D3 = 0
∴ The quadratic form is positive semi definite

Example 3

Determine the nature of the quadratic form

2 2 2
x1 + 2 x2 + 3 x3 + 2 x1 x2 + 2 x2 x3 − 2 x3 x1

143
Sathyabama University

Solution

ª 1 1 −1º
«
Let the matrix of the quadratic from is A = « 1 2 1 »
»
«¬ −1 1 3 »¼

1 1
Now D1 =| 1|= 1 D2 = =1 D3 = A = − 2
1 2

Here D1, D2 are positive and D3 is negative

∴ The quadratic form is indefinite.

EXERCISE

Part – A

1. Write the matrix of the quadratic form

x12 + 3 x2 2 + 6 x32 − 2 x1 x3 + 6 x1 x2 + 5 x2 x3

2. Write the matrix of the QF = 2 x12 − 3 x1 x2 + 4 x22 .

3. Write the matrix of the quadratic form

2 x 2 + y 2 + z 2 + 2 xy − 2 xz − 4 yz

4. Write the matrix of the QF = 2 x 2 + 6 y 2 + 2 z 2 + 8 xz

5. Write the matrix of the QF = x 2 + 2 y 2 + z 2 − 2 xy + 2 yz

6. Find the matrix of the quadratic form

QF = x12 + 3 x2 2 + 6 x3 2 + 2 x1 x2 + 2 x2 x3 + 4 x1 x3

§ 3 1 0·
¨ 2 ¸
7. ¨
Write the Q.F whose matrix is A = 1 0 6 ¸.
¨ 2 ¸
¨¨ 0 6 −7 ¸¸
© ¹

144
 Sathyabama University

§1 2·
8. Write the Q.F whose matrix is A = ¨ ¸
© 2 3¹
9. Find the nature of the Q.F without reducing them to canonical form.
(i) 6 x 2 + 3 y 2 + 14 z 2 + 4 yz + 18 xz + 4 xy

(ii) x12 + 2 x2 2 + 3 x3 2 + 2 x1 x2 + 2 x2 x3 − 2 x1 x3

(iii) x 2 + y 2 + z 2 − 2 xy

10. Find the nature of the QF

(i) 5 x12 + 5 x2 2 + 14 x32 + 2 x1 x2 − 16 x2 x3 − 8 x1 x3

(ii) 2 x 2 + y 2 − 3 z 2 + 12 xy − 8 yz − 4 xz

ANSWERS

§ 1 −1 3 ·¸
¨
1 A = ¨ −1 3 5 ¸
¨ 2¸
¨¨ ¸
3 5 6 ¸
1. © 2 ¹

§ 2 −3 ·
2. A= ¨ 2¸
¨ −3 4 ¸
© 2 ¹

§ 2 1 −1 ·
¨ ¸
3. A = ¨ 1 1 −2 ¸
¨ −1 −2 1 ¸
© ¹

§ 2 0 4·
¨ ¸
4. A = ¨ 0 6 0¸
¨ 4 0 2¸
© ¹

§ 1 −1 0 ·
¨ ¸
5. A = ¨ −1 2 1 ¸
¨ 0 1 1¸
© ¹

145
Sathyabama University

§1 1 2·
¨ ¸
6. A = ¨1 3 1¸
¨ 2 1 6¸
© ¹

§ 1 ·
¨3 2
0¸
¨ ¸
A = ¨¨ 6 ¸¸
1
7. 0
2
¨ ¸
¨0 6 −7 ¸
¨ ¸
© ¹

§1 2·
8. A=¨ ¸
© 2 3¹
9. (i) Positive definite
(ii) Indefinite
(iii) Positive semidefinite
10. (i) Positive semidefinite
(ii) Indefinite

Part – B

1. C.F: − y12 + y22 + 4 y32

Indefinite in nature, Rank = 3, Index = 2, Signature = 1.

2. C.F: −2 y12 + 6 y22 + 6 y32

Nature: Indefinite in nature.

3. C. F: y22 + 3 y32

Nature: Positive semidefinite

4. C.F: 3 y2 2 + 14 y32

x1 = 1, x2 = −5, x3 = 4

4 2
5. (i) C. F: 3 y12 − y2 − y32 , Nature: Indefinite
3

(ii) C.F: 3 y2 2 + 15 y32 , Nature: Positive semidefinite

146
 Sathyabama University

3 2 1 2
(iii) C.F: 3 y12 + y2 + y3 , Nature: Positive definite
2 3

(iv) C.F: y12 + y2 2 + y32 , Nature: Positive definite

(v) C.F: 2 y12 + 2 y2 2 + 6 y32 , Nature: Positive definite

(vi) C.F: 4 y12 + y2 2 + y32 , Nature: Positive definite

(vii) C.F: 3 y12 + 6 y2 2 − 9 y3 2 , Nature: Indefinite

147