LINEAR ALGEBRA AND CALCULUS

UNIT - II

Matrices - 2
NAMALA YUVARAJU
Assistant Professor
Faculty of Mathematics
SNIST
Cell: 9505862392

N YUVARAJU 1
 Topics in UNIT-II

➢ Hermitian, Skew-Hermitian and Unitary matrices

➢ Eigen values, Eigen vectors
➢ Cayley-Hamilton Theorem (without proof) and its
applications in finding higher powers of a matrix and inverse
➢ Diagonalization of a matrix

N YUVARAJU 2
Special Matrices:
1. Hermitian Matrix:
A square matrix A = [𝑎𝑖𝑗 ] is called Hermitian if 𝑎𝑖𝑗 = 𝑎ത𝑗𝑖 for all i & j,
i.e., 𝐴𝜃 = 𝐴. ( where 𝐴𝜃 is transpose of complex conjugate of A, i.e., 𝐴𝜃 = (𝐴)ҧ 𝑇 )
Hence, the diagonal elements of a Hermitian matrix are real.
1 2+𝑖 −3𝑖
2 4+𝑖
Ex: , 2−𝑖 −2 4 + 3𝑖
4−𝑖 3
3𝑖 4 − 3𝑖 3

2. Skew-Hermitian Matrix:
A square matrix A = [𝑎𝑖𝑗 ] is called skew-Hermitian if 𝑎𝑖𝑗 = −𝑎ത𝑗𝑖 for all i & j,
i.e., 𝐴𝜃 = −𝐴.
Hence, the diagonal elements of a skew-Hermitian matrix are either purely
imaginary or zero.
𝑖 1 3+𝑖
0 2+𝑖
Ex: , −1 2𝑖 −2 − 𝑖
−2 + 𝑖 0
−3 + 𝑖 2−𝑖 0
N YUVARAJU 3
Properties of Hermitian and Skew-Hermitian Matrices :

Property 1: Every square matrix can be uniquely expressed as the sum of a Hermitian
matrix and a skew-Hermitian matrix.
1 1
i.e., Let A be a square matrix, then 𝐴 = 𝐴 + 𝐴𝜃 + (𝐴 − 𝐴𝜃 )
2 2
1
where 𝐴 + 𝐴𝜃 = Hermitian matrix
2
1
(𝐴 − 𝐴𝜃 ) = Skew-Hermitian matrix
2

Property 2: Every square matrix can be uniquely expressed as 𝑃 + 𝑖𝑄, where 𝑃 and 𝑄
are Hermitian matrices.
1 1
i.e., Let A be a square matrix, then 𝐴 = 𝐴 + 𝐴𝜃 + 𝑖 𝐴 − 𝐴𝜃
2 2𝑖
= 𝑃 + 𝑖𝑄

1 1
where P = 𝐴 + 𝐴𝜃 and Q = (𝐴 − 𝐴𝜃 ) are Hermitian matrices
2 2𝑖
N YUVARAJU 4
Property 3: Every Hermitian matrix can be written as 𝑃 + 𝑖𝑄, where 𝑃 is a real
symmetric and 𝑄 is a real skew-symmetric matrix.
i.e., Let A be a Hermitian matrix, i.e., 𝐴𝜃 = 𝐴
1 1
then 𝐴 = 𝑃 + 𝑖𝑄 = 𝐴 + 𝐴ҧ + 𝑖 (𝐴 − 𝐴)ҧ
2 2𝑖
1
where P = 𝐴 + 𝐴ҧ = real symmetric matrix
2
1
Q= (𝐴 − 𝐴)ҧ = real skew-symmetric matrix
2𝑖

Property 4: Every skew-Hermitian matrix can be written as 𝑃 + 𝑖𝑄, where 𝑃 is a

real skew-symmetric and 𝑄 is a real symmetric matrix.
i.e., Let A be a skew-Hermitian matrix, i.e., 𝐴𝜃 = −𝐴
1 1
then 𝐴 = 𝑃 + 𝑖𝑄 = 𝐴 + 𝐴ҧ + 𝑖 (𝐴 − 𝐴)ҧ
2 2𝑖
1
where P = 𝐴 + 𝐴ҧ = real skew-symmetric matrix
2
1
Q= (𝐴 − 𝐴)ҧ = real symmetric matrix
2𝑖
N YUVARAJU 5
 𝟐 + 𝟑𝒊 𝟎 𝟒𝒊
Example 1: Express the matrix A= 𝟓 𝒊 𝟖 as the sum of a Hermitian
𝟏 − 𝒊 −𝟑 + 𝒊 𝟔
matrix and a skew-Hermitian matrix.
2 + 3𝑖 0 4𝑖 2 − 3𝑖 0 −4𝑖 2 − 3𝑖 5 1+𝑖
Sol: Given A = 5 𝑖 8 => 𝐴ҧ = 5 −𝑖 8 ⇒ 𝐴𝜃 = (𝐴)ҧ 𝑇 = 0 −𝑖 −3 − 𝑖
1−𝑖 −3 + 𝑖 6 1+𝑖 −3 − 𝑖 6 −4𝑖 8 6

2 + 3𝑖 0 4𝑖 2 − 3𝑖 5 1+𝑖 4 5 1 + 5𝑖
1 1
Now 𝐴+ 𝐴𝜃 = 5 𝑖 8 + 0 −𝑖 −3 − 𝑖 =
1
5 0 5−𝑖
2 2 2
1−𝑖 −3 + 𝑖 6 −4𝑖 8 6 1 − 5𝑖 5+𝑖 12

2 + 3𝑖 0 4𝑖 2 − 3𝑖 5 1+𝑖 6𝑖 −5 −1 + 3𝑖
1 𝜃 1 1
𝐴−𝐴 = 5 𝑖 8 − 0 −𝑖 −3 − 𝑖 = 5 2𝑖 11 + 𝑖
2 2 2
1−𝑖 −3 + 𝑖 6 −4𝑖 8 6 1 + 3𝑖 −11 + 𝑖 0

1 1
Where 𝐴 + 𝐴𝜃 is Hermitian matrix and (𝐴 − 𝐴𝜃 ) is Skew-Hermitian matrix
2 2
4 5 1 + 5𝑖 6𝑖 −5 −1 + 3𝑖
1 1 1 1
∴ 𝐴 = 𝐴 + 𝐴𝜃 + (𝐴 − 𝐴𝜃 ) = 5 0 5−𝑖 + 5 2𝑖 11 + 𝑖
2 2 2 2
1 − 5𝑖 5 + 𝑖 12 1 + 3𝑖 −11 + 𝑖 0
N YUVARAJU 6
 𝟐𝒊 −𝟑 𝟏−𝒊
Example 2: Express the matrix A = 𝟎 𝟐 + 𝟑𝒊 𝟏 + 𝒊 as 𝑷 + 𝒊𝑸, where 𝑷 and 𝑸
−𝟑𝒊 𝟑 + 𝟐𝒊 𝟐 − 𝟓𝒊
are Hermitian matrices.
2𝑖 −3 1−𝑖 −2𝑖 −3 1+𝑖 −2𝑖 0 3𝑖
Sol: Given A = 0 2 + 3𝑖 1+𝑖 ⇒ 𝐴ҧ = 0 2 − 3𝑖 1−𝑖 ⇒ 𝐴𝜃 = (𝐴)ҧ 𝑇 = −3 2 − 3𝑖 3 − 2𝑖
−3𝑖 3 + 2𝑖 2 − 5𝑖 3𝑖 3 − 2𝑖 2 + 5𝑖 1+𝑖 1−𝑖 2 + 5𝑖

2𝑖 −3 1−𝑖 −2𝑖 0 3𝑖
1 1 1 0 −3 1 + 2𝑖
𝜃
Now, P = 𝐴+𝐴 = 0 2 + 3𝑖 1+𝑖 + −3 2 − 3𝑖 3 − 2𝑖 = −3 4 4−𝑖
2 2 2 1 − 2𝑖 4 + 𝑖 4
−3𝑖 3 + 2𝑖 2 − 5𝑖 1+𝑖 1−𝑖 2 + 5𝑖

2𝑖 −3 1−𝑖 −2𝑖 0 3𝑖
1 1 1
Q= 𝐴 − 𝐴𝜃 = 0 2 + 3𝑖 1+𝑖 − −3 2 − 3𝑖 3 − 2𝑖 =
2𝑖 2𝑖 2𝑖
−3𝑖 3 + 2𝑖 2 − 5𝑖 1+𝑖 1−𝑖 2 + 5𝑖
4𝑖 −3 1 − 4𝑖
3 6𝑖 −2 + 3𝑖
−1 − 4𝑖 2 + 3𝑖 −10𝑖
1 1
Where P = 𝐴 + 𝐴𝜃 and Q = (𝐴 − 𝐴𝜃 ) are Hermitian matrices
2 2𝑖

1 0 −3 1 + 2𝑖 1 4𝑖 −3N YUVARAJU
1 − 4𝑖 7
∴ 𝐴 = 𝑃 + 𝑖𝑄 = −3 4 4−𝑖 + i 3 6𝑖 −2 + 3𝑖
2 1 − 2𝑖 4 + 𝑖 4 2𝑖 −1 − 4𝑖 2 + 3𝑖 −10𝑖
Unitary Matrix:
A square matrix 𝐴 is called unitary if 𝐴𝐴𝜃 = 𝐴𝜃 𝐴 = 𝐼 (or) 𝐴−1 = 𝐴𝜃 .
Where 𝐴𝜃 is transpose of complex conjugate of A, i.e., 𝐴𝜃 = (𝐴)ҧ 𝑇
𝟐 −𝒊 𝟐 𝟎
Example 1: If A = 𝟏
𝟐 𝒊 𝟐 − 𝟐 𝟎 , prove that the matrix A is unitary, hence find 𝐴−1 .
𝟎 𝟎 𝟐 2 −𝑖 2 0 2 −𝑖 2 0
𝜃 1 1
∴ 𝐴𝐴 = 𝑖 2 − 2 0 ×2 𝑖 2 − 2 0
2
1
2 −𝑖 2 0 0 0 2 0 0 2
Sol: Given A = 2 𝑖 2 − 2 0 1
2+2+0 −2𝑖 + 2𝑖 + 0 0 + 0 + 0
= 2𝑖 − 2𝑖 + 0 2+2+0 0+0+0
0 0 2 4
0+0+0 0+0+0 0+0+4
4 0 0
1
2 𝑖 2 0 = 0 4 0
1 4
⇒ 𝐴ҧ = −𝑖 2 − 2 0 0 0 4
2 1 0 0
0 0 2 = 0 1 0
0 0 1
2 −𝑖 2 0 = 𝐼
1
and 𝐴𝜃 = (𝐴)ҧ 𝑇 = 𝑖 2 − 2 0 Hence, given A is unitary matrix.
2
0 0 2 2 −𝑖 2 0
And 𝐴−1 = 𝐴𝜃 1
= 𝑖 2 − 2
2 0
N YUVARAJU
0 0 2 8
Practice Problems:

2 2+𝑖 3
1. Express the matrix −2 + 𝑖 0 4 as the sum of a Hermitian and
−𝑖 3−𝑖 1−𝑖
skew-Hermitian matrices.

2 3−𝑖 1 + 2𝑖
2. Express the matrix 𝑖 0 1 as 𝑃 + 𝑖𝑄 , where P and Q are both
1 + 2𝑖 1 3𝑖
Hermitian.

3. Show that the following matrices are unitary, hence, find its inverse:
2+𝑖 2𝑖 𝑖 3
3 3 2 2
𝑖) 2𝑖 2−𝑖
𝑖𝑖)
3 𝑖
3 3 2 2

N YUVARAJU 9
Eigen Values and Eigen Vectors :
Eigen Matrix:
If A is square matrix of order n, 𝜆 is a scalar and I is the unit matrix
of order n then ‘ A − 𝜆𝐼 ’ is called Eigen matrix (or) Characteristic matrix of A, which
is square matrix of order n.
𝑎11 𝑎12 . . . 𝑎1𝑛 1 0 . . . 0
𝑎21 𝑎22 . . . 𝑎
𝑨 − 𝝀𝑰 = …… . …… . …… … 2𝑛
… . −𝜆 …0 …1… . . … …. . …0… .
𝑎𝑛1 𝑎𝑛2 . . . 𝑎𝑛𝑛 0 0 . . . 1
𝑎11 − 𝜆 𝑎12 . . . 𝑎1𝑛
= 𝑎 𝑎 −𝜆 . . . 𝑎2𝑛
…21…. …22
…. …… …….
𝑎𝑛1 𝑎𝑛2 . . . 𝑎𝑛𝑛 − 𝜆
Eigen Polynomial:
The determinant of eigen matrix 𝐴 − 𝜆𝐼, i. e. , 𝐴 − 𝜆𝐼 is called
Eigen polynomial (or) Characteristic polynomial of A.
𝑎11 − 𝜆 𝑎12 . . . 𝑎1𝑛
𝐴 − 𝜆𝐼 = …𝑎21 𝑎 −𝜆 . . . 𝑎2𝑛 = (−1)𝑛 [𝜆𝑛 + 𝑎1 𝜆𝑛−1 + 𝑎2 𝜆𝑛−2 + −− − + 𝑎𝑛 ]
…. …22
…. …… …….
𝑎𝑛1 𝑎𝑛2 . . . 𝑎𝑛𝑛 − 𝜆
N YUVARAJU 10
Eigen Equation:
An equation 𝐴 − 𝜆𝐼 = 0 is called Eigen equation (or) Characteristic
equation of the matrix A of order n.
i.e., 𝐴 − 𝜆𝐼 = 𝜆𝑛 + 𝑎1 𝜆𝑛−1 + 𝑎2 𝜆𝑛−2 + −− − + 𝑎𝑛 = 0

Eigen Values:
The roots of a eigen equation 𝐴 − 𝜆𝐼 = 0 of square matrix A of order n are called
Eigen values (or) Characteristic values (or) Proper values (or) Latent roots of A.
i.e., 𝐴 − 𝜆𝐼 = 0 <=> 𝜆′𝑖 𝑠 are eigen values of A , where 𝑖 = 1,2 … … 𝑛

Eigen Vectors:
A non zero column vector X is said to be Eigen vector (or) Characteristic
Vector of square matrix A of order n, corresponding to eigen value 𝜆 if 𝐴𝑋 = 𝜆X (or)
𝐴 − 𝜆𝐼 𝑋 = 0. where 0 is zero column matrix.

Here 𝐴 − 𝜆𝐼 𝑋 = 0 is called matrix equation. N YUVARAJU 11

Spectrum of a Matrix:

The set of all eigen values of a matrix A is called spectrum of A.

i.e., Spectrum of A = { 𝜆′𝑖 s / 𝜆𝑖 is eigen value of A }

Algebraic Multiplicity of Eigen value of a Matrix:

If 𝜆 is an eigen value of a matrix A repeated ‘m’ times, then the number ‘m’ is
called algebraic multiplicity(A.M) of eigen value 𝜆.

Ex: If A is a square matrix of order 5 and eigen values of A are 𝜆 = 1,3,3, −1, −1

then, A.M of eigen value 1 is 1

A.M of eigen value 3 is 2
A.M of eigen value -1 is 2
N YUVARAJU 12
Geometric Multiplicity of Eigen value of a Matrix:

The required number of arbitrary constants (or) the number of linearly

independent eigen vectors corresponding to the eigen value is called the geometric
multiplicity(G.M) of eigen value.

* If ‘n-r’ is no. of arbitrary constants(geometric multiplicity) values & ‘m’

is algebraic multiplicity of any eigenvalue then 1 ≤ (𝑛 − 𝑟) ≤ 𝑚 .

Trace of a Matrix:

The sum of all principal diagonal elements of a square matrix A is

called trace of a matrix A. It is denoted by tra(A).

2 3 4
Ex: 𝐴 = 4 3 1 => tra(A) = 2 + 3 + 4 = 9
1 2 4 N YUVARAJU 13
Note:

1. The characteristic equation of the matrix A of order 2 can be obtained from

𝜆2 − 𝑆1 𝜆 + 𝑆2 = 0
where 𝑆1 = sum of principal diagonal elements of A = trace of A
𝑆2 = determinant A = 𝐴

2. The characteristic equation of the matrix A of order 3 can be obtained from

𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0
where 𝑆1 = sum of principal diagonal elements of A = trace of A
𝑆2 = sum of minors of principal diagonal elements of A
𝑆3 = determinant A = 𝐴

3. The sum of the eigen values of a matrix is the sum of its principal diagonal elements.

4. The product of the eigen values of a matrix is the determinant of the matrix.
N YUVARAJU 14
Examples:
𝟐 𝟐 𝟏
1. Find the eigen values of the matrix A = 𝟏 𝟑 𝟏
𝟏 𝟐 𝟐
2 2 1
Sol: Given A = 1 3 1
1 2 2
Since, 𝜆 is eigen value of A  eigen equation 𝐴 − 𝜆𝐼 = 0
2−𝜆 2 1
1 3−𝜆 1 =0
1 2 2−𝜆

2−𝜆 2−𝜆 3−𝜆 −2 −2 2−𝜆−1 +1 2−3+𝜆 =0

2 − 𝜆 𝜆2 − 5𝜆 + 4 − 2 + 2𝜆 − 1 + 𝜆 = 0
2𝜆2 − 10𝜆 + 8 − 𝜆3 + 5𝜆2 − 4𝜆 + 3𝜆 − 3 = 0
𝜆3 − 7𝜆2 + 11𝜆 − 5 = 0
clearly, 𝜆 = 1 is root => 𝜆 − 1 𝜆2 − 6𝜆 + 5 = 0
𝜆−1 𝜆−1 𝜆−5 =0
𝜆 = 1,1,5
∴ Eigen values of A are 1, 1, 5
N YUVARAJU 15
 𝟑 −𝟏 𝟏
2. Find the characteristic roots of the matrix A = −𝟏 𝟓 −𝟏
𝟏 −𝟏 𝟑
3 −1 1
Sol: Given A = −1 5 −1
1 −1 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0
(or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 ______ (1)
where 𝑆1 = trace of A = 3 + 5 + 3 = 11
𝑆2 = sum of minors of principal diagonal elements of A
5 −1 3 1 3 −1
= + + = 15 − 1 + 9 − 1 + 15 − 1 = 36
−1 3 1 3 −1 5
𝑆3 = determinant A = 𝐴 = 3 15 − 1 + 1 −3 + 1 + 1 1 − 5 = 42 − 2 − 4 = 36

From (1), 𝜆3 − 11𝜆2 + 36𝜆 − 36 = 0

clearly 𝜆 = 2 is root => 𝜆 − 2 𝜆2 − 9𝜆 + 18 = 0
𝜆−2 𝜆−3 𝜆−6 =0
∴ The characteristic roots of A are 2, 3, 6
N YUVARAJU 16
 𝟖 −𝟔 𝟐
3. Find the eigen values and eigen vectors of the matrix A = −𝟔 𝟕 −𝟒
𝟐 −𝟒 𝟑
8 −6 2
Sol: Given A = −6 7 −4
2 −4 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0
(or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 ______ (1)
where 𝑆1 = trace of A = 8 + 7 +3 = 18
𝑆2 = sum of minors of principal diagonal elements of A
7 −4 8 2 8 −6
= + +
−4 3 2 3 −6 7
= 21 - 16 + 24 – 4 + 56 – 36 = 45
𝑆3 = 𝐴 = 8 21 − 16 + 6 −18 + 8 + 2 24 − 14
= 40 - 60 + 20 = 0
From (1), 𝜆3 − 18𝜆2 + 45𝜆 − 0 = 0
λ 𝜆2 − 18𝜆 + 45 = 0
λ 𝜆 − 3 𝜆 − 15 = 0
∴ The characteristic roots of A are 0, 3, 15 N YUVARAJU 17
Now to find eigen vectors corresponding to each eigen value of the matrix A.
Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix equation
is 𝐴 − 𝜆𝐼 𝑋 = 0 1 1
𝑅1 → 2 𝑅1 , 𝑅2 → 20 𝑅2
𝒊) 𝑰𝒇 𝝀 = 𝟎: 4 −3 1 𝑥 0
AX = 0 ~ 0 1 −1 𝑦 = 0
0 0 0 𝑧 0
8 −6 2 𝑥 0
Rank = r = 2 , n = 3 (no. of unknowns)
−6 7 −4 𝑦 = 0
No. of arbitrary values = n – r = 3 – 2 = 1
2 −4 3 𝑧 0
Corresponding equations
4x – 3y + z = 0
𝑅2 → 8𝑅2 + 6𝑅1 , 𝑅3 → 4𝑅3 − 𝑅1 y-z=0
8 −6 2 𝑥 0
Let 𝑧 = 𝑘1 , 𝑘1 is non zero real number
~ 0 20 −20 𝑦 = 0
Then y - 𝑘1 = 0 => y = 𝑘1
0 −10 10 𝑧 0 1
and 4x - 3𝑘1 + 𝑘1 = 0 => x = 𝑘1
2
1 1
𝑅3 → 2𝑅3 + 𝑅2 2
𝑘1
2
8 −6 2 𝑥 0 ∴ Eigen vector 𝑋1 = 𝑘1 = 1 𝑘1 corresponding to eigen value
~ 0 20 −20 𝑦 = 0 𝑘1 1
0 0 0 𝑧 0 𝜆 = 0. N YUVARAJU 18
 −1
𝒊𝒊) 𝑰𝒇 𝝀 = 𝟑: 𝑅2 →
8 2
𝑅
5 −6 2 𝑥 0
𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 3𝐼 𝑋 = 0 ~ 0 2 1 𝑦 = 0
0 0 0 𝑧 0
8 − 3 −6 2 𝑥 0
−6 7 − 3 −4 𝑦 = 0 Rank = r = 2 , n = 3 (no. of unknowns)
2 −4 3 − 3 𝑧 0 No. of arbitrary values = n – r = 3 – 2 = 1
Corresponding equations
5 −6 2 𝑥 0 5x – 6y + 2z = 0
−6 4 −4 𝑦 = 0 2y + z = 0
2 −4 0 𝑧 0

𝑅2 → 5𝑅2 + 6𝑅1 , 𝑅3 → 5𝑅3 − 2𝑅1

Let 𝑧 = 𝑘2 , 𝑘2 is non zero real number
−1
5 −6 2 𝑥 0 Then 2y + 𝑘2 = 0 => y = 𝑘2
2
~ 0 −16 −8 𝑦 = 0 −1
0 −8 −4 𝑧 0 and 5x - 6. 𝑘2 + 2𝑘2 = 0 => x = -𝑘2
2

𝑅3 → 2𝑅3 − 𝑅2 −𝑘2 −1
5 −6 2 𝑥 0 −1 −1
∴ Eigen vector 𝑋2 = 𝑘 = 𝑘2 corresponding to eigen
~ 0 −16 −8 𝑦 = 0 2 2 2
0 0 0 𝑧 0 𝑘2 1
value 𝜆 = 3. N YUVARAJU 19
 −1
𝒊𝒊𝒊) 𝑰𝒇 𝝀 = 𝟏𝟓: 𝑅2 → 𝑅
20 2
−7 −6 2 𝑥 0
𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 15𝐼 𝑋 = 0 ~ 0 1 2 𝑦 = 0
0 0 0 𝑧 0
8 − 15 −6 2 𝑥 0
−6 7 − 15 −4 𝑦 = 0 Rank = r = 2 , n = 3 (no. of unknowns)
2 −4 3 − 15 𝑧 0 No. of arbitrary values = n – r = 3 – 2 = 1
Corresponding equations
−7 −6 2 𝑥 0 -7x – 6y + 2z = 0
−6 −8 −4 𝑦 = 0 y + 2z = 0
2 −4 −12 𝑧 0

𝑅2 → 7𝑅2 − 6𝑅1 , 𝑅3 → 7𝑅3 + 2𝑅1

Let 𝑧 = 𝑘3 , 𝑘3 is non zero real number
−7 −6 2 𝑥 0 Then y + 2𝑘3 = 0 => y = -2𝑘3
~ 0 −20 −40 𝑦 = 0 and -7x - 6(−2𝑘3 ) + 2𝑘3 = 0 => x = 2𝑘3
0 −40 −80 𝑧 0
2𝑘3 2
𝑅3 → 𝑅3 − 2𝑅2 ∴ Eigen vector 𝑋3 = −2𝑘3 = −2 𝑘3 corresponding to
−7 −6 2 𝑥 0
𝑘3 1
~ 0 −20 −40 𝑦 = 0
0 0 0 𝑧 0 eigen value 𝜆 = 15.
N YUVARAJU 20
Practice Problems:
9 −1 9
1. Find the eigenvalues and eigenvectors for the matrix 3 −1 3
−7 1 −7
( Ans: eigenvalues 𝜆 = −1, 0, 2
1 1 4
eigenvectors 𝑋1 = 1 , 𝑋2 = 0 , 𝑋3 = 1
1 −1 −3
2 1 1
2. Find the eigenvalues and eigenvectors for the matrix 2 3 2
3 3 4
1 0 1
( Ans: eigenvalues 𝜆 = 1, 1, 7 and eigenvectors 𝑋1 = 2 , 𝑋2 = 1 , 𝑋3 = 0
3 −1 −1

−3 −7 −5
3. Find the eigenvalues and eigenvectors for the matrix 2 4 3
1 2 2
−3
( Ans: eigenvalues 𝜆 = 1, 1, 1 and eigenvectors X = 1
1 N YUVARAJU 21
Properties of Eigen values:

1. If 𝝀 is an eigenvalue of the matrix A then 𝝀 is also an eigenvalue of 𝑨𝑻 .

𝟏
2. If 𝝀 ≠ 𝟎 is an eigenvalue of the non-singular matrix A then is an eigenvalue of 𝑨−𝟏 .
𝝀

3. If 𝝀 is an eigenvalue of the matrix A then 𝝀𝒏 is an eigenvalue of 𝑨𝒏 , 𝒏𝝐𝑵.

4. If 𝝀 is an eigenvalue of the matrix A then 𝝀 ± 𝒌 is an eigenvalue of 𝑨 ± 𝒌𝑰 .

5. If 𝝀 is an eigenvalue of the matrix A then 𝐤𝝀 is also an eigenvalue of 𝒌𝑨 , 𝒌 ≠ 𝟎.

6. The eigenvalues of a triangular matrix and diagonal matrix are the diagonal elements
of the matrix.

7. The eigenvalues of a real symmetric matrix are real.

N YUVARAJU 22
8. The sum of the eigen values of a matrix is the sum of its principal diagonal elements.

9. The product of the eigen values of a matrix is the determinant of the matrix.

𝑨
10. If 𝝀 ≠ 𝟎 is an eigenvalue of the matrix A then is eigen value of adjoint of A.
𝝀

11. If A and P are two square matrices of same order & P is non-singular then A and
𝑷−𝟏 𝑨𝑷 have same eigenvalues.

𝟏
12. If 𝝀 ≠ 𝟎 is an eigenvalue of an orthogonal matrix then is also it’s eigenvalue.
𝝀

13. If 𝝀 is an eigenvalue of the matrix A then 𝒂𝟎 + 𝒂𝟏 𝝀 + 𝒂𝟐 𝝀𝟐 + − − + 𝒂𝒏 𝝀𝒏 = 𝒇(𝝀)

is an eigenvalue of the polynomial matrix
𝒂𝟎 𝑰 + 𝒂𝟏 𝑨 + 𝒂𝟐 𝑨𝟐 + − − + 𝒂𝒏 𝑨𝒏 = 𝒇(𝑨)
N YUVARAJU 23
Properties of Eigen vectors:

1. If X is an eigenvector of a matrix A corresponding to the eigenvalue 𝝀 then KX is

also an eigenvector of A corresponding to same eigenvalue 𝝀 , where K≠ 𝟎.

2. If X is an eigenvector of a matrix A then X cannot correspond to more than one

eigenvalue of A.

3. The eigenvectors corresponding to distinct eigenvalues of a matrix are linearly

independent.

4. The eigenvectors corresponding to distinct eigenvalues of a real symmetric matrix

are orthogonal.

Orthogonal vectors:
The vectors 𝑋1 & 𝑋2 are orthogonal if 𝑋1𝑇 𝑋2 = 𝑋1 𝑋2𝑇 = 0
N YUVARAJU 24
Examples:
𝟑 −𝟏 𝟏
1. If 𝐀 = −𝟏 𝟓 −𝟏 , find the eigenvalues of the following matrices:
𝟏 −𝟏 𝟑
𝒊) 𝑨 𝒊𝒊) 𝑨𝑻 𝒊𝒊𝒊) 𝑨−𝟏 𝒊𝒗) 𝑨𝟐 𝒗) 𝟒𝑨 𝒗𝒊) 𝑨𝟐 − 𝟐𝑨 + 𝑰 𝒗𝒊𝒊) 𝒂𝒅𝒋𝒐𝒊𝒏𝒕 𝒐𝒇 𝑨
3 −1 1
Sol: Given A = −1 5 −1
1 −1 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 𝑜𝑟 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 ___ (1)
where 𝑆1 = trace of A = 3 + 5 + 3 = 11
𝑆2 = sum of minors of principal diagonal elements of A
5 −1 3 1 3 −1
= + + = 15 − 1 + 9 − 1 + 15 − 1 = 36
−1 3 1 3 −1 5
𝑆3 = 𝐴 = 3 15 − 1 + 1 −3 + 1 + 1 1 − 5 = 36
From (1), 𝜆3 − 11𝜆2 + 36𝜆 − 36 = 0
(λ − 2) 𝜆2 − 9𝜆 + 18 = 0 ⇒ 𝜆 − 2 𝜆 − 3 𝜆 − 6 = 0
∴ 𝜆 = 2, 3, 6
Hence, the eigenvalues of A are 2, 3, 6 N YUVARAJU 25
i) The eigenvalues of the matrix A = 𝜆 = 2, 3, 6

ii) The eigenvalues of 𝐴𝑇 = The eigenvalues of A = 𝜆 = 2, 3, 6

1 1 1 1
iii) The eigenvalues of 𝐴−1 = = , ,
𝜆 2 3 6

iv) The eigenvalues of 𝐴2 = 𝜆2 = 22 , 32 , 62 = 4, 9, 36

v) The eigenvalues of 4A = 4𝜆 = 4 2 , 4 3 , 4 6 = 8, 12, 24

vi) The eigenvalues of 𝐴2 − 2𝐴 + 𝐼 = 𝜆2 − 2𝜆 + 1

= 22 − 2 2 + 1, 32 − 2 3 + 1, 62 − 2 6 + 1
= 1, 4, 25

𝐴 36 36 36
vii) The eigenvalues of adjoint of A = = , , = 18, 12, 6
𝜆 2 3 6
N YUVARAJU 26
 𝟒 𝟔 𝟔
2. Find the sum and product of the eigenvalues of the matrix 𝑨 = 𝟏 𝟑 𝟐
−𝟏 −𝟒 −𝟑

4 6 6
Sol: 𝐴= 1 3 2
−1 −4 −3
Sum of the eigenvalues of A = Trace of A
= 4+3–3
= 4

Product of the eigenvalues of A = 𝐴

= 4(-9+8) - 6(-3+2) + 6(-4+3)
= -4+6–6
= -4
N YUVARAJU 27
3. Form the matrix whose eigenvalues are 𝜶 − 𝟓, 𝜷 − 𝟓, 𝜸 − 𝟓, 𝒘𝒉𝒆𝒓𝒆 𝜶, 𝜷, 𝜸 are
−𝟏 −𝟐 −𝟑
the eigenvalues of 𝑨 = 𝟒 𝟓 −𝟔
𝟕 −𝟖 𝟗
−1 −2 −3
Sol: Given 𝐴 = 4 5 −6
7 −8 9

We know that, If 𝜆1 , 𝜆2 , 𝑎𝑛𝑑 𝜆3 are eigenvalues of the matrix A then

𝜆1 −𝑘, 𝜆2 −𝑘, 𝑎𝑛𝑑 𝜆3 −𝑘 are the eigenvalues of 𝐴 − 𝑘𝐼.
Here, put 𝑘 = 5
−1 −2 −3 1 0 0
∴ Required matrix = 𝐴 − 5𝐼 = 4 5 −6 − 5 0 1 0
7 −8 9 0 0 1
−6 −2 −3
= 4 0 −6
7 −8 4
N YUVARAJU 28
 𝟔 −𝟐 𝟐
4. Find the eigenvalues of the matrix 𝑨 = 𝟎 𝟑 𝟎 , and also find the
𝟎 𝟎 𝟑
𝟏
eigenvalues of 𝒊) 𝑨−𝟏 𝒊𝒊) 𝒂𝒅𝒋𝑨 𝒊𝒊𝒊) 𝟐𝑨𝟐 − 𝑨 + 𝟑𝑰 .
𝟐

6 −2 2
Sol: Given 𝐴 = 0 3 0 , which is in upper triangular matrix
0 0 3
Since the eigenvalues of the upper triangular matrix are it’s principal diagonal
elements.
∴ The eigenvalues of A = 𝜆 = 6, 3, 3
−1 1 1 1 1
i) The eigenvalues of 𝐴 = = , ,
𝜆 6 3 3
𝐴 54 54 54
ii) The eigenvalues of adjA = = , , = 9, 18, 18
𝜆 6 3 3
2 1 2 1
iii) The eigenvalues of 2𝐴 − 𝐴 + 3𝐼 = 2𝜆 − 𝜆 + 3
2 2
2 1 2 1 2 1
= 2(6) − (6) + 3, 2(3) − (3) + 3, 2(3) − (3) +3
2 2 2
39 39
= 72, , N YUVARAJU 29
2 2
 𝟏 𝟐 𝟑
5. Find the eigenvalues and orthogonal eigenvectors for the matrix 𝟐 𝟒 𝟔
𝟑 𝟔 𝟗
1 2 3
Sol: Let A = 2 4 6
3 6 9
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 (or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 ___ (1)
where 𝑆1 = trace of A = 1 + 4 + 9 = 14
𝑆2 = sum of minors of principal diagonal elements of A
4 6 1 3 1 2
= + +
6 9 3 9 2 4
= 36 – 36 + 9 – 9 + 4 - 4 = 0
𝑆3 = 𝐴 = 1(36 – 36) - 2(18 – 18) + 3(12 – 12) = 0

From (1), 𝜆3 − 14𝜆2 + 0 − 0 = 0

𝜆2 𝜆 − 14 = 0
λ = 0, 0, 14
∴ The characteristic roots of A are 0, 0, 14
N YUVARAJU 30
Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix equation is
𝐴 − 𝜆𝐼 𝑋 = 0
𝑅3 → 𝑅3 + 𝑅2
𝒊) 𝑰𝒇 𝝀 = 𝟏𝟒: −13 2 3 𝑥 0
𝐴 − 14𝐼 𝑋 = 0 ~ 0 −3 2 𝑦 = 0
1 − 14 2 3 𝑥 0 0 0 0 𝑧 0
2 4 − 14 6 𝑦 = 0
3 6 9 − 14 𝑧 0 Rank = r = 2 , n = 3 (no. of unknowns)
No. of arbitrary values = n – r = 3 – 2 = 1
−13 2 3 𝑥 0 The corresponding system of equations is
2 −10 6 𝑦 = 0 -13x + 2y + 3z = 0
3 6 −5 𝑧 0
-3y + 2z = 0
𝑅2 → 13𝑅2 + 2𝑅1 , 𝑅3 → 13𝑅3 + 3𝑅1 Let 𝑧 = 𝑘1 , 𝑘1 is non zero real number
2
−13 2 3 𝑥 0 Then -3y + 2𝑘1 = 0 => y = 𝑘1
3
~ 0 −126 84 𝑦 = 0 2 1
0 84 −56 𝑧 0 and -13x + 2. 𝑘1 +3𝑘1 = 0 => x = 𝑘1
3 3
1 1
𝑥 𝑘
1 1 3 1 3 1
𝑅2 → 𝑅 , 𝑅3 → 𝑅 ∴ Eigen vector 𝑋1 = 𝑦 = 2
42 2 28 3 𝑘 = 2 𝑘1 = 2 𝑡 ,
−13 2 3 𝑥 0 𝑧 3 1 3 3
~ 0 −3 2 𝑦 = 0 𝑘1 1
1
0 3 −2 𝑧 0 (𝑤ℎ𝑒𝑟𝑒 𝑙𝑒𝑡 𝑡 = 3 1
N YUVARAJU
𝑘 ) corresponding to eigen value 𝜆 = 14. 31
𝒊𝒊) 𝑰𝒇 𝝀 = 𝟎: 𝑥 + 2𝑘3 + 3𝑘2 = 0
𝑥 = −2𝑘3 − 3𝑘2
Matrix equation 𝐴𝑋 = 0 𝑥 −2𝑘3 − 3𝑘2
∴ Eigenvectors 𝑋 = 𝑦 = 𝑘3
1 2 3 𝑥 0 𝑧 𝑘2
2 4 6 𝑦 = 0
𝑧
−3 −2
3 6 9 0
= 0 𝑘2 + 1 𝑘3
𝑅2 → 𝑅2 − 2𝑅1 , 𝑅3 → 𝑅3 − 3𝑅1 1 0
1 2 3 𝑥 0
−3 −2
~ 0 0 0 𝑦 = 0
i.e., Eigenvectors 𝑋2 = 0 𝑘2 , 𝑋3 = 1 𝑘3
0 0 0 𝑧 0
1 0
Rank = r = 1 , n = 3 ( no. of unknowns)
corresponding to eigenvalue 𝜆 = 0.
The no. arbitrary values = n – r = 3 – 1 = 2
The corresponding system of equations is
Now verifying orthogonality of these
𝑥 + 2𝑦 + 3𝑧 = 0
vectors.
Let 𝑧 = 𝑘2 , 𝑦 = 𝑘3 , 𝑘2 &𝑘3 are non-zero
real numbers. N YUVARAJU 32
 1 −3 −2
The linearly independent eigenvectors are 𝑋1 = 2 , 𝑋2 = 0 , 𝑋3 = 1
3 1 0

−3
∴ 𝑋1𝑇 𝑋2 = 1 2 3 0 = −3 + 0 + 3 = 0,
1
−2
𝑋1𝑇 𝑋3 = 1 2 3 1 = −2 + 2 + 0 = 0
0
−2
And 𝑋2𝑇 𝑋3 = −3 0 1 1 = 6 + 0 + 0 = 6 ≠ 0
0
We observing, 𝑋1 , 𝑋2 and 𝑋1 , 𝑋3 are orthogonal, but 𝑋2 , 𝑋3 are not orthogonal.

𝑎
Now choose the eigenvector 𝑋3 = 𝑏 such that 𝑋1𝑇 𝑋3 = 0 and 𝑋2𝑇 𝑋3 = 0
𝑐 N YUVARAJU 33
 𝑎
∴ 𝑋1𝑇 𝑋3 = 1 2 3 𝑏 = 0 ⇒ 𝑎 + 2𝑏 + 3𝑐 = 0 ____(1)
𝑐
𝑎
and 𝑋2𝑇 𝑋3 = −3 0 1 𝑏 = 0 ⇒ −3𝑎 + 𝑐 = 0 _____(2)
𝑐
𝑎 −𝑏 𝑐
Solving (1) & (2), 2 3 = 1 3 = 1 2 = 𝑡 (𝑠𝑎𝑦)
0 1 −3 1 −3 0

𝑎 −𝑏 𝑐 𝑎 𝑏 𝑐
= = =𝑡 𝑜𝑟 = = =𝑡
2 10 6 1 −5 3
⇒ 𝑎 = 𝑡, 𝑏 = −5𝑡, 𝑐 = 3𝑡
𝑡 1
Therefore, eigenvector 𝑋3 = −5𝑡 = −5 𝑡 , corresponding to eigenvalue 𝜆 = 0.
3𝑡 3
Hence, the linearly independent orthogonal eigenvectors of A are
1 −3 1
𝑋1 = 2 , 𝑋2 = 0 , 𝑋3 = −5
3 1 3
N YUVARAJU 34
Practice Problems:
1 2 −1
1. If A = 0 2 2 , find the eigenvalues of the following matrices:
0 0 2
𝑖) 𝐴 𝑖𝑖) 𝐴𝑇 𝑖𝑖𝑖) 𝐴−1 𝑖𝑣) 𝐴3 𝑣) − 2𝐴 𝑣𝑖) 𝐴2 − 2𝐴 + 𝐼 𝑣𝑖𝑖) 𝑎𝑑𝑗𝑜𝑖𝑛𝑡 𝑜𝑓 𝐴
1 1
( Ans: 𝑖) 1,2,2 𝑖𝑖) 1,2,2 𝑖𝑖𝑖) 1, , 𝑖𝑣) 1,8,8 𝑣) − 2, −4, −4 𝑣𝑖) 0,1,1 𝑣𝑖𝑖) 4,2,2 )
2 2

2. Find the characteristic roots of 𝑐𝑜𝑠𝛼 −𝑠𝑖𝑛𝛼

𝑠𝑖𝑛𝛼 𝑐𝑜𝑠𝛼
( Ans: 𝑐𝑜𝑠𝛼 ± 𝑠𝑖𝑛𝛼 )
3 10 5
3. Find the eigenvalues and eigenvectors for the matrix −2 −3 −4
3 5 7
1 5
( Ans: eigenvalues 𝜆 = 3, 2, 2 and eigenvectors 𝑋1 = 1 , 𝑋2 = 2 )
−2 −5
3 −1 1
4. Find the eigenvalues and for the matrix −1 3 −1 and check whether the
1 −1 3
eigenvectors are orthogonal.
N YUVARAJU 35
Cayley-Hamilton Theorem:
Statement: Every square matrix satisfies its own characteristic equation.

Let A be a square matrix of order ‘n’. Its characteristic equation is

𝐴 − 𝜆𝐼 = 𝜆𝑛 + 𝑎1 𝜆𝑛−1 + 𝑎2 𝜆𝑛−2 + −− − + 𝑎𝑛 = 0
By Cayley-Hamilton theorem
𝐴𝑛 + 𝑎1 𝐴𝑛−1 + 𝑎2 𝐴𝑛−2 + −− − + 𝑎𝑛 𝐼 = 0
To find inverse of a non-singular matrix by using Cayley-Hamilton theorem:
Let A be a non-singular matrix of order n , then A satisfies its own
characteristic equation.
i.e., 𝐴𝑛 +𝑎1 𝐴𝑛−1 + 𝑎2 𝐴𝑛−2 + −− − + 𝑎𝑛 𝐼 = 0 , (𝑎𝑛 ≠ 0) ____ (1)
premultiplying equation(1) by 𝐴−1 , we have
𝐴𝑛−1 +𝑎1 𝐴𝑛−2 + 𝑎2 𝐴𝑛−3 + −− − + 𝑎𝑛 𝐴−1 = 0
−1 1
∴ 𝐴 =− 𝐴𝑛−1 +𝑎1 𝐴𝑛−2 + 𝑎2 𝐴𝑛−3 + −− − + 𝑎𝑛−1 𝐼
𝑎𝑛
Note: We can also find higher powers of a square matrix by Cayley-Hamilton theorem.
N YUVARAJU 36
 𝟐 −𝟏 𝟏
Example 1: Verify the Cayley-Hamilton theorem for the matrix 𝑨 = −𝟏 𝟐 −𝟏
𝟏 −𝟏 𝟐
and, hence, find 𝑨−𝟏 and 𝑨𝟒 .
2 −1 1
Sol: Given 𝐴 = −1 2 −1
1 −1 2
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 𝑜𝑟 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 ___ (1)

where 𝑆1 = trace of A = 2 + 2 + 2 = 6
𝑆2 = sum of minors of principal diagonal elements of A
2 −1 2 1 2 −1
= + + = 4−1+4−1+4−1=9
−1 2 1 2 −1 2
𝑆3 = 𝐴 = 2 4 − 1 + 1 −2 + 1 + 1 1 − 2 = 6 − 1 − 1 = 4
From (1), the characteristic equation is
𝜆3 − 6𝜆2 + 9𝜆 − 4 = 0
Now, to verify Cayley-Hamilton theorem for given matrix A
i.e., To show that 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = 0
N YUVARAJU 37
 2 −1 1 2 −1 1 6 −5 5 2 −1 1
2
𝐴 = −1 2 −1 −1 2 −1 𝐴3 = 𝐴2 . 𝐴 = −5 6 −5 −1 2 −1
1 −1 2 1 −1 2 5 −5 6 1 −1 2

4+1+1 −2 − 2 − 1 2+1+2 12 + 5 + 5 −6 − 10 − 5 6 + 5 + 10
= −2 − 2 − 1 1+4+1 −1 − 2 − 2 = −10 − 6 − 5 5 + 12 + 5 −5 − 6 − 10
2+1+2 −1 − 2 − 1 1+1+4 10 + 5 + 6 −5 − 10 − 6 5 + 5 + 12

6 −5 5 22 −21 21
= −5 6 −5 = −21 22 −21
5 −5 6 21 −21 22

22 −21 21 6 −5 5 2 −1 1 1 0 0
∴ 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = −21 22 −21 − 6 −5 6 −5 + 9 −1 2 −1 − 4 0 1 0
21 −21 22 5 −5 6 1 −1 2 0 0 1

22 − 36 + 18 − 4 −21 + 30 − 9 − 0 21 − 30 + 9 − 0 0 0 0
= −21 + 30 − 9 − 0 22 − 36 + 18 − 4 −21 + 30 − 9 − 0 = 0 0 0
21 − 30 + 9 − 0 −21 + 30 − 9 − 0 22 − 36 + 18 − 4 0 0 0
= 0
Hence, Cayley-Hamilton theorem is verified. N YUVARAJU 38
Now, to find 𝐴−1 and 𝐴4 by using Cayley-Hamilton theorem.
𝒊) Since 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = 0
Premultiplying by 𝐴−1 on both sides, 𝐴−1 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = 𝐴−1 . 0
𝐴2 − 6𝐴 + 9𝐼 − 4𝐴−1 = 0
4𝐴−1 = 𝐴2 −6𝐴 + 9𝐼
6 −5 5 2 −1 1 1 0 0
= −5 6 −5 − 6 −1 2 −1 + 9 0 1 0
5 −5 6 1 −1 2 0 0 1

6 − 12 + 9 −5 + 6 + 0 5 − 6 + 0
= −5 + 6 + 0 6 − 12 + 9 −5 + 6 + 0
5 − 6 + 0 −5 + 6 + 0 6 − 12 + 9

3 1 −1
= 1 3 1
−1 1 3

3 1 −1
1
∴ 𝐴−1 = 1 3 1
4
−1 1 3 N YUVARAJU 39
𝒊𝒊) Since 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = 0
Premultiplying by A on both sides, A 𝐴3 − 6𝐴2 + 9𝐴 − 4𝐼 = 𝐴. 0

𝐴4 − 6𝐴3 + 9𝐴2 − 4𝐴 = 0

∴ 𝐴4 = 6𝐴3 − 9𝐴2 + 4𝐴
22 −21 21 6 −5 5 2 −1 1
= 6 −21 22 −21 − 9 −5 6 −5 + 4 −1 2 −1
21 −21 22 5 −5 6 1 −1 2

132 − 54 + 8 −126 + 45 − 4 126 − 45 + 4

= −126 + 45 − 4 132 − 54 + 8 −126 + 45 − 4
126 − 45 + 4 −126 + 45 − 4 132 − 54 + 8

86 −85 85
= −85 86 −85
85 −85 86
N YUVARAJU 40
 𝟏 𝟒
Example 2: Find the eigenvalues of the matrix 𝑨 = and verify the
𝟐 𝟑
Cayley-Hamilton theorem for this matrix. Find 𝑨−𝟏 and also, express
𝑨𝟓 − 𝟒𝑨𝟒 − 𝟕𝑨𝟑 + 𝟏𝟏𝑨𝟐 − 𝑨 − 𝟏𝟎𝑰 as a linear polynomial in A.

1 4
Sol: Given 𝐴 =
2 3
1−𝜆 4
Its characteristic equation is 𝐴 − 𝜆𝐼 = 0 ⇒ =0
2 3−𝜆
𝑜𝑟 𝜆2 − 𝑆1 𝜆 + 𝑆2 = 0 ___(1)
Where 𝑆1 = trace of A = 1 + 3 = 4
𝑆2 = 𝐴 = 3 − 8 = −5

From (1), characteristic equation is𝜆2 − 4𝜆 − 5 = 0

𝜆+1 𝜆−5 =0
𝜆 = −1, 5
∴ Eigenvalues are 𝜆 = −1, 5
N YUVARAJU 41
Now, to verify Cayley-Hamilton theorem for given matrix A.
i.e., to show that 𝐴2 −4𝐴 − 5𝐼 = 0

1 4 1 4 1 + 8 4 + 12 9 16
𝐴2 = 𝐴. 𝐴 = = =
2 3 2 3 2 + 6 8 + 9 8 17
9 16 1 4 1 0
∴ 𝐴2 −4𝐴 − 5𝐼 = −4 −5
8 17 2 3 0 1

9 − 4 − 5 16 − 16 − 0
=
8 − 8 − 0 17 − 12 − 5

0 0
=
0 0
= 0
∴ The matrix A satisfies its own characteristic equation.
i.e., The Cayley-Hamilton theorem is verified.
N YUVARAJU 42
𝒊) To find 𝐴−1

Since 𝐴2 −4𝐴 − 5𝐼 = 0
𝐴−1 𝐴2 −4𝐴 − 5𝐼 = 𝐴−1 . 0
𝐴 − 4𝐼 − 5𝐴−1 = 0
5𝐴−1 = 𝐴 − 4𝐼
1 4 1 0
= −4
2 3 0 1
1−4 4−0
=
2−0 3−4
−3 4
=
2 −1
−1 1 −3 4
∴𝐴 =
5 2 −1
N YUVARAJU 43
𝒊𝒊) To express 𝐴5 − 4𝐴4 − 7𝐴3 + 11𝐴2 − 𝐴 − 10𝐼 as a linear polynomial in A.

𝐴5 − 4𝐴4 − 7𝐴3 + 11𝐴2 − 𝐴 − 10𝐼

= 𝐴5 − 4𝐴4 − 5𝐴3 − 2𝐴3 + 11𝐴2 − 𝐴 − 10𝐼

= 𝐴3 𝐴2 − 4𝐴 − 5𝐼 − 2𝐴3 + 8𝐴2 + 3𝐴2 + 10𝐴 − 11𝐴 − 10𝐼

= 𝐴3 0 − 2𝐴 𝐴2 − 4𝐴 − 5𝐼 + 3𝐴2 − 12𝐴 + 𝐴 − 15𝐼 + 5𝐼

= −2𝐴 0 + 3 𝐴2 − 4𝐴 − 5𝐼 + 𝐴 + 5𝐼

= 3 0 + 𝐴 + 5𝐼

= 𝐴 + 5𝐼 , which is in a linear polynomial in A.

N YUVARAJU 44
 𝟏 𝟐
Example 3: Using Cayley-Hamilton theorem find 𝑨𝟖 , if 𝑨 =
𝟐 −𝟏
1 2
Sol: Given 𝐴 =
2 −1
1−𝜆 2
Its eigen equation is 𝐴 − 𝜆𝐼 = 0 ⇒ =0
2 −1 − 𝜆
1 − 𝜆 −1 − 𝜆 − 4 = 0
−1 − 𝜆 + 𝜆 + 𝜆2 − 4 = 0
𝜆2 − 5 = 0
By Cayley-Hamilton theorem, A satisfies its characteristic equation.
Then 𝐴2 − 5𝐼 = 0
⇒ 𝐴2 = 5𝐼
∴ 𝐴8 = 𝐴2 4
= 5𝐼 4
= 625𝐼 N YUVARAJU 45
 𝟐 𝟑
Example 4: If 𝑨 = , express 𝑨𝟒 in powers of 𝑨.
𝟏 𝟐
2 3
Sol: Given 𝐴 =
1 2
2−𝜆 3
Its eigen equation is 𝐴 − 𝜆𝐼 = 0 ⇒ =0
1 2−𝜆
2−𝜆 2−𝜆 −3=0
4 − 2𝜆 − 2𝜆 + 𝜆2 − 3 = 0
𝜆2 − 4𝜆 + 1 = 0
By Cayley-Hamilton theorem, A satisfies its characteristic equation.
Then 𝐴2 − 4𝐴 + 𝐼 = 0
⇒ 𝐴2 = 4𝐴 − 𝐼
∴ 𝐴4 = 𝐴2 2
= 4𝐴 − 𝐼 2
= 16𝐴2 − 8𝐴 + 𝐼 (or)
= 16 4𝐴 − 𝐼 − 8𝐴 + 𝐼
= 56𝐴 − 15𝐼 , which is a linear polynomial in A.
N YUVARAJU 46
Practice Problems:
1 2 3
1. Find the characteristic equation of 𝐴 = 2 −1 4 and hence verify C.H.T.
3 1 −1
1 2 3
2. Verify Cayley-Hamilton theorem for the matrix 𝐴 = 2 4 5 , hence find 𝐴−1 .
3 5 6
1 −3 2
( Ans: 𝐴−1 = −3 3 −1 )
2 −1 0
3. Using Cayley Hamilton theorem find the inverse of A and 𝐴4 of the matrix
7 2 −2 −3 −2 2 241 80 −80
1
𝐴 = −6 −1 2 . ( Ans: 𝐴−1 = 6 4
5 −2 , 𝐴 = −240 −79 80 )
3
6 2 −1 −6 −2 5 240 80 −79
0 𝑐 −𝑏
4. Show that the matrix 𝐴 = −𝑐 0 𝑎 satisfies it’s characteristic equation.
𝑏 −𝑎 0
1 2
5. If 𝐴 = , express 𝐴6 − 4𝐴5 + 8𝐴4 − 12𝐴3 + 14𝐴2 as a linear polynomial
−1 3
in 𝐴. ( Ans: −4𝐴 + 5𝐼 )
N YUVARAJU 47
Diagonalization of a matrix:

Similar Matrices:
If A and B are two square matrices of order n then B is said to be
similar to A if there exists a nonsingular matrix P such that 𝐵 = 𝑃−1 𝐴𝑃.

Note:

1) Similarity of matrices is an equivalence relation.

2) Similar matrices have the determinant.

3) Similar matrices have the same characteristic polynomial and, hence, the same
eigenvalues. If X is an eigenvector of A corresponding to the eigenvalue 𝜆 then
𝑃−1 𝑋 is eigenvector of B corresponding to the eigenvalue 𝜆 , where 𝐵 = 𝑃−1 𝐴𝑃.
N YUVARAJU 48
Diagonalizable matrix:

A matrix A is said to be diagonalizable if it is similar to a diagonal matrix by

similar transformation.
(or)

If A is a square matrix of order n and If there exists a non-singular matrix P of

order n such that 𝑃−1 𝐴𝑃 is a diagonal matrix D (i. e. , 𝑃−1 𝐴𝑃 = 𝐷) whose diagonal
elements are the eigenvalues of A, where 𝑃 = 𝑋1 𝑋2 … … 𝑋𝑛 whose columns
𝑋1 , 𝑋2 … … 𝑋𝑛 are linearly independent eigenvectors corresponding to eigenvalues
of A, then the matrix A is called diagonalizable matrix.
This process is called diagonalization of a matrix.

Here the matrix D is called spectral matrix.

the matrix P is called modal matrix.

N YUVARAJU 49
Imp. Notes:

1) A square matrix A of order n is diagonalizable if and only if it possesses n linearly

independent eigenvectors.

2) If the eigenvalues of the matrix A of order n are all distinct then it is always similar
to a diagonal matrix i.e., A is diagonalizable.

3) If the eigenvalues of the matrix A of order n are not distinct and if the algebraic
multiplicity(A.M) and geometric multiplicity(G.M) of each eigenvalue of A are
same, then the matrix A is diagonalizable i.e., similar to a diagonal matrix.

4) If A is similar to a diagonal matrix D, the diagonal elements of D are the eigenvalues

of A.

N YUVARAJU 50
Working Rule for Diagonalization of a Square Matrix A:

Step 1: Find the eigenvalues of the square matrix A.

Step 2: Find the eigenvectors corresponding to each eigenvalue.

Step 3: Check the condition A.M = G.M, for each eigenvalue of A.

Step 4: Find the modal matrix P having the eigenvectors as its column
vectors.

Step 5: Find the diagonal matrix 𝐷 = 𝑃−1 𝐴𝑃. The diagonal matrix D has
eigenvalues as its diagonal elements.
N YUVARAJU 51
 𝟖 −𝟔 𝟐
Example 1: Show that the matrix A = −𝟔 𝟕 −𝟒 is similar to its diagonal matrix.
𝟐 −𝟒 𝟑
Find the diagonal and modal matrices.
8 −6 2
Sol: Given A = −6 7 −4
2 −4 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 (or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 __ (1)
where 𝑆1 = trace of A = 8 + 7 +3 = 18
𝑆2 = sum of minors of principal diagonal elements of A
7 −4 8 2 8 −6
= + + = 21 - 16 + 24 – 4 + 56 – 36 = 45
−4 3 2 3 −6 7
𝑆3 = 𝐴 = 8 21 − 16 + 6 −18 + 8 + 2 24 − 14 = 40 - 60 + 20 = 0
From (1), 𝜆3 − 18𝜆2 + 45𝜆 − 0 = 0
λ 𝜆2 − 18𝜆 + 45 = 0 => λ 𝜆 − 3 𝜆 − 15 = 0
∴ The characteristic roots of A are 0, 3, 15
And Algebraic multiplicity(A.M) of λ = 0 is 1
Algebraic multiplicity(A.M) of λ = 3 is 1
Algebraic multiplicity(A.M) of λ = 15 is 1 N YUVARAJU 52
Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix equation
is 𝐴 − 𝜆𝐼 𝑋 = 0 1 1
𝑅1 → 𝑅1 , 𝑅2 → 𝑅2
2 20
𝒊) 𝑰𝒇 𝝀 = 𝟎: 4 −3 1 𝑥 0
AX = 0 ~ 0 1 −1 𝑦 = 0
0 0 0 𝑧 0

8 −6 2 𝑥 0 Rank = r = 2 , n = 3 (no. of unknowns)

−6 7 −4 𝑦 = 0 No. of arbitrary values (L.I vectors) = G.M = n – r = 3 – 2 = 1
2 −4 3 𝑧 0 ∴ A.M = G.M for the eigenvalue 𝜆 = 0.
Corresponding equations
𝑅2 → 8𝑅2 + 6𝑅1 , 𝑅3 → 4𝑅3 − 𝑅1 4x – 3y + z = 0
8 −6 2 𝑥 0 y-z=0
~ 0 20 −20 𝑦 = 0 Let 𝑧 = 𝑘1 , 𝑘1 is non zero real number
0 −10 10 𝑧 0 Then y - 𝑘1 = 0 => y = 𝑘1
1
and 4x - 3𝑘1 + 𝑘1 = 0 => x = 𝑘1
𝑅3 → 2𝑅3 + 𝑅2 2
1
8 −6 2 𝑥 0 2
𝑘1 1
1
~ 0 20 −20 𝑦 = 0 ∴ Eigen vector 𝑋1 = 𝑘1 = 2 𝑘1 corresponding to
2
eigen
𝑘1 2
0 0 0 𝑧 0
value 𝜆 = 0. N YUVARAJU 53
𝒊𝒊) 𝑰𝒇 𝝀 = 𝟑: −1
𝑅2 → 𝑅2
8
5 −6 2 𝑥 0
𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 3𝐼 𝑋 = 0 ~ 0 2 1 𝑦 = 0
0 0 0 𝑧 0
8−3 −6 2 𝑥 0
−6 7 − 3 −4 𝑦 = 0 Rank = r = 2 , n = 3 (no. of unknowns)
2 −4 3−3 𝑧 0 No. of arbitrary values (L.I vectors) = G.M = n – r = 3 – 2 = 1
∴ A.M = G.M for eigenvalue 𝜆 = 3.
5 −6 2 𝑥 0 Corresponding equations
−6 4 −4 𝑦 = 0 5x – 6y + 2z = 0
2 −4 0 𝑧 0
2y + z = 0
𝑅2 → 5𝑅2 + 6𝑅1 , 𝑅3 → 5𝑅3 − 2𝑅1 Let 𝑧 = 𝑘2 , 𝑘2 is non zero real number
−1
5 −6 2 𝑥 0 Then 2y + 𝑘2 = 0 => y = 𝑘2
2
~ 0 −16 −8 𝑦 = 0 −1
0 −8 −4 𝑧 0 and 5x - 6. 𝑘2 + 2𝑘2 = 0 => x = -𝑘2
2
−𝑘2 2
−1 −1
𝑅3 → 2𝑅3 − 𝑅2 ∴ Eigen vector 𝑋2 = 2 𝑘2 = 1 ( )𝑘2 corresponding to
5 −6 2 𝑥 0 2
𝑘2 −2
~ 0 −16 −8 𝑦 = 0
0 0 0 𝑧 0 eigen value 𝜆 = 3.
N YUVARAJU 54
 −1
𝒊𝒊𝒊) 𝑰𝒇 𝝀 = 𝟏𝟓: 𝑅2 → 𝑅
20 2
−7 −6 2 𝑥 0
𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 15𝐼 𝑋 = 0 ~ 0 1 2 𝑦 = 0
0 0 0 𝑧 0

8 − 15 −6 2 𝑥 0 Rank = r = 2 , n = 3 (no. of unknowns)

−6 7 − 15 −4 𝑦 = 0
2 −4 3 − 15 𝑧 0
No. of arbitrary values (L.I vectors) = G.M = n – r = 3 – 2 = 1
∴ A.M = G.M for eigenvalue 𝜆 = 15.
−7 −6 2 𝑥 0 Corresponding equations
−6 −8 −4 𝑦 = 0 -7x – 6y + 2z = 0
2 −4 −12 𝑧 0
y + 2z = 0
𝑅2 → 7𝑅2 − 6𝑅1 , 𝑅3 → 7𝑅3 + 2𝑅1
−7 −6 2 𝑥 0 Let 𝑧 = 𝑘3 , 𝑘3 is non zero real number
~ 0 −20 −40 𝑦 = 0 Then y + 2𝑘3 = 0 => y = -2𝑘3
0 −40 −80 𝑧 0 and -7x - 6(−2𝑘3 ) + 2𝑘3 = 0 => x = 2𝑘3
𝑅3 → 𝑅3 − 2𝑅2 2𝑘3
−7 −6 2 𝑥 0 2
~ 0 −20 −40 𝑦 = 0 ∴ Eigen vector 𝑋3 = −2𝑘3 = −2 𝑘3 corresponding to eigen
0 0 0 𝑧 0 𝑘3 1
value 𝜆 = 15. N YUVARAJU 55
 1 2 2
So that the linearly independent eigenvectors are 𝑋1 = 2 , 𝑋2 = 1 , 𝑋3 = −2
2 −2 1
corresponding to eigenvalues λ = 0, 3, 15 respectively.

1 2 2
∴ Modal matrix 𝑃 = 𝑋1 𝑋2 𝑋3 = 2 1 −2
2 −2 1
and 𝑃 = 1 1 − 4 − 2 2 + 4 + 2 −4 − 2 = −3 − 12 − 12 = −27 ≠ 0
Now, to find inverse matrix of the modal matrix P.
−3 −6 −6
Cofactor matrix of 𝑃 = −6 −3 6
−6 6 −3
−3 −6 −6
Adjoint matrix of P = Transpose of cofactor matrix of P = −6 −3 6
−6 6 −3
1 2 2
i.e., Adj(P) = −3 2 1 −2
2 −2 1 N YUVARAJU 56
 1 2 2
−3 2 1 −2 1 2 2
𝐴𝑑𝑗(𝑃) 2 −2 1 1
∴ 𝑃−1 = = = 2 1 −2
𝑃 −27 9
2 −2 1
Now, to find 𝑃−1 𝐴𝑃 = 𝐷

1 2 2 8 −6 2 1 2 2
1
𝑃−1 𝐴𝑃 = 2 1 −2 −6 7 −4 2 1 −2
9
2 −2 1 2 −4 3 2 −2 1

8 − 12 + 4 −6 + 14 − 8 2 − 8 + 6 1 2 2
1
= 16 − 6 − 4 −12 + 7 + 8 4 − 4 − 6 2 1 −2
9
16 + 12 + 2 −12 − 14 − 4 4 + 8 + 3 2 −2 1

0 0 0 1 2 2
1
= 6 3 −6 2 1 −2
9
30 −30 15 2 −2 1
N YUVARAJU 57
 0+0+0 0+0+0 0+0+0
1
= 6 + 6 − 12 12 + 3 + 12 12 − 6 − 6
9
30 − 60 + 30 60 − 30 − 30 60 + 60 + 15

0 0 0
1
= 0 27 0
9
0 0 135

0 0 0
= 0 3 0
0 0 15

= D
0 0 0
∴ 𝑃−1 𝐴𝑃 = D = 0 3 0 ,which is a diagonal matrix (is also called spectral matrix).
0 0 15

The diagonal elements of D are the eigenvalues of the given matrix A.

Hence, the given matrix A is diagonalizable. N YUVARAJU 58
 𝟏 𝟐 𝟐
Example 2: Test for diagonalizable the matrix 𝟎 𝟐 𝟏
−𝟏 𝟐 𝟐
1 2 2
Sol: Let A = 0 2 1
−1 2 2
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 (or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0 __ (1)
where 𝑆1 = trace of A = 1 + 2 + 2 = 5
𝑆2 = sum of minors of principal diagonal elements of A
2 1 1 2 1 2
= + + =4−2+2+2+2−0=8
2 2 −1 2 0 2
𝑆3 = 𝐴 = 1 4 − 2 − 2 0 + 1 + 2 0 + 2 = 2 − 2 + 4 = 4
From (1), 𝜆3 − 5𝜆2 + 8𝜆 − 4 = 0
(λ − 1) 𝜆2 − 4𝜆 + 4 = 0
(λ − 1) 𝜆 − 2 𝜆 − 2 = 0
∴ The characteristic roots of A are 2, 2, 1
And Algebraic multiplicity(A.M) of λ = 2 is 2
Algebraic multiplicity(A.M) of λ = 1 is 1
N YUVARAJU 59
Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix equation
is 𝐴 − 𝜆𝐼 𝑋 = 0
𝒊) 𝑰𝒇 𝝀 = 𝟐: 𝑅3 → 𝑅3 + 2𝑅2
−1 2 2 𝑥 0
~ 0 0 1 𝑦 = 0
𝐴 − 𝜆𝐼 𝑋 = 0 0 0 0 𝑧 0
⇒ 𝐴 − 2𝐼 𝑋 = 0
Rank = r = 2 , No. of unknowns = n = 3
1−2 2 2 𝑥 0 No. of arbitrary values ( L.I vectors ) = G.M = n – r
0 2−2 1 𝑦 = 0
𝑧
=3–2
−1 2 2−2 0
=1
−1 2 2 𝑥 0 But A.M of λ = 2 is 2
0 0 1 𝑦 = 0
−1 2 0 𝑧 0
Then A.M ≠ G.M for eigenvalue λ = 2.
𝑅3 → 𝑅3 − 𝑅1
−1 2 2 𝑥 0
~ 0 0 1 𝑦 = 0
Hence, given matrix A is not diagonalizable.
0 0 −2 𝑧 0 N YUVARAJU 60
Calculation of Powers of a Matrix:

We can find the powers of a square matrix by using diagonalization.

Let A be a square matrix of order n, then there exists an invertible matrix P such that
𝐷 = 𝑃−1 𝐴𝑃
Now, 𝐷2 = 𝐷. 𝐷 = 𝑃−1 𝐴𝑃 𝑃−1 𝐴𝑃
= 𝑃−1 𝐴 𝑃𝑃−1 𝐴𝑃
= 𝑃−1 𝐴 𝐼 𝐴𝑃
∴ 𝐷2 = 𝑃−1 𝐴2 𝑃
Similarly, we get 𝐷3 = 𝑃−1 𝐴3 𝑃, 𝐷4 = 𝑃−1 𝐴4 𝑃………, 𝐷𝑛 = 𝑃−1 𝐴𝑛 𝑃

Now, pre-multiplying 𝐷𝑛 by P and post-multiplying by 𝑃−1 , we get

𝑃𝐷𝑛 𝑃−1 = 𝑃𝑃−1 𝐴𝑛 𝑃𝑃−1
= 𝐼𝐴𝑛 𝐼
= 𝐴𝑛
∴ 𝐴𝑛 = 𝑃𝐷𝑛 𝑃−1 , 𝑛 ∈ 𝑁 N YUVARAJU 61
 𝟏 𝟎 −𝟏
Example: Find the modal matrix P which transform the matrix 𝑨 = 𝟏 𝟐 𝟏 to
𝟐 𝟐 𝟑
diagonal form. Hence calculate 𝑨𝟒 .

1 0 −1
Sol: Given 𝐴 = 1 2 1
2 2 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 (or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0
Eigenvalues of A are 𝜆 = 1, 2, 3

Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix equation
is 𝐴 − 𝜆𝐼 𝑋 = 0

𝒊) 𝑰𝒇 𝝀 = 𝟏: 𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴−𝐼 𝑋 =0
1
∴ L.I eigenvector 𝑋1 = −1 corresponding to eigenvalue 𝜆 = 1.
0 N YUVARAJU 62
𝒊𝒊) 𝑰𝒇 𝝀 = 𝟐: 𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 2𝐼 𝑋 = 0
−2
∴ L.I eigenvector 𝑋2 = 1 corresponding to eigenvalue 𝜆 = 2.
2

𝒊𝒊𝒊) 𝑰𝒇 𝝀 = 𝟑: 𝐴 − 𝜆𝐼 𝑋 = 0 ⇒𝐴 − 3𝐼 𝑋 = 0
−1
∴ L.I eigenvector 𝑋3 = 1 corresponding to eigenvalue 𝜆 = 3.
2
1 −2 −1
∴ Modal matrix 𝑃 = 𝑋1 𝑋2 𝑋3 = −1 1 1
0 2 2
𝐴𝑑𝑗(𝑃)
Now, to find inverse matrix of the modal matrix P. i.e., 𝑃−1 =
𝑃
0 2 −1
−1
∴ 𝑃−1 = 2 2 0
2
−2 −2 −1 N YUVARAJU 63
Now, to find 𝐷 = 𝑃−1 𝐴𝑃
0 2 −1 1 0 −1 1 −2 −1
−1
𝐷 = 𝑃−1 𝐴𝑃 = 2 2 0 1 2 1 −1 1 1
2
−2 −2 −1 2 2 3 0 2 2
1 0 0
= 0 2 0 , which is diagonal matrix ( Spectral matrix )
0 0 3

The diagonal elements of D are the eigenvalues of the given matrix A.

Hence, the given matrix A is diagonalizable.

Now, to find 𝐴4
Since 𝐴𝑛 = 𝑃𝐷𝑛 𝑃−1 , 𝑛 ∈ 𝑁

=> 𝐴4 = 𝑃𝐷4 𝑃−1

N YUVARAJU 64
 4
1 −2 −1 1 0 0 0 2 −1
−1
∴ 𝐴4 = −1 1 1 0 2 0 ( ) 2 2 0
2
0 2 2 0 0 3 −2 −2 −1
1 −2 −1 14 0 0 0 2 −1
−1
= −1 1 1 0 24 0 2
2 2 0
0 2 2 0 0 34 −2 −2 −1

1 −2 −1 1 0 0 0 2 −1
−1
= −1 1 1 0 16 0 ( 2 ) 2 2 0
0 2 2 0 0 81 −2 −2 −1

−49 −50 −40

= 65 66 40
130 130 81

N YUVARAJU 65
Practice Problems:

−2 2 −3
1. Diagonalize the matrix 𝐴 = 2 1 −6
−1 −2 0

2 −2 −2
2. Test for diagonalization for the matrix 𝐴 = −2 3 −1
2 −1 3

8 −8 −2
3. Find modal matrix P which transform the matrix 𝐴 = 4 −3 −2 to diagonal
3 −4 1
4
form. Hence calculate 𝐴 .

N YUVARAJU 66
Orthogonally Similar Matrices:

If A and B are two square matrices of order n then B is said to be

orthogonally similar to A if there exists an orthogonal matrix P such that 𝐵 = 𝑃−1 𝐴𝑃.

Since P is orthogonal, 𝑃−1 = 𝑃𝑇 , then we write 𝐵 = 𝑃−1 𝐴𝑃 = 𝑃𝑇 𝐴𝑃.

Note:

1) To find the orthogonal matrix P , each element of the eigenvector is divided by its
norm (length).

2) A real symmetric matrix of order n has mutually orthogonal real eigenvectors.

3) Any two eigenvectors corresponding to two distinct eigenvalues of a real symmetric

matrix are orthogonal.
N YUVARAJU 67
Example: Determine a diagonal matrix orthogonally similar (orthogonal
𝟔 −𝟐 𝟐
reduction) to the real symmetric matrix −𝟐 𝟑 −𝟏 . Also, find the
𝟐 −𝟏 𝟑
modal matrix.

6 −2 2
Sol: Let A = −2 3 −1
2 −1 3
The characteristic equation is 𝐴 − 𝜆𝐼 = 0 (or) 𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − 𝑆3 = 0
Eigenvalues of A are 𝜆 = 8, 2, 2
Let X be an eigen vector corresponding to eigen value 𝜆 of A, then the matrix
equation is 𝐴 − 𝜆𝐼 𝑋 = 0
𝒊) 𝝀 = 𝟖:
Matrix equation is 𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 8𝐼 𝑋 = 0
2
∴ L.I eigenvector 𝑋1 = −1 corresponding to eigenvalue 𝜆 = 8.
1 N YUVARAJU 68
𝒊𝒊) 𝑰𝒇 𝝀 = 𝟐:
𝐴 − 𝜆𝐼 𝑋 = 0 ⇒ 𝐴 − 2𝐼 𝑋 = 0
1 −1
∴ L.I eigenvectors 𝑋2 = 2 , 𝑋3 = 0 corresponding to eigenvalue 𝜆 = 2.
0 2

2 1 −1
Hence, The linearly independent eigenvectors are 𝑋1 = −1 , 𝑋2 = 2 , 𝑋3 = 0
1 0 2
Since, The orthogonal modal matrix has mutually orthogonal eigenvectors.
Now verifying these three eigenvectors are orthogonal or not.
1
∴ 𝑋1𝑇 𝑋2= 2 −1 1 2 = 2 − 2 + 0 = 0,
0
−1
𝑇
𝑋1 𝑋3 = 2 −1 1 0 = −2 − 0 + 2 = 0
2
−1
And 𝑋2𝑇 𝑋3 = 1 2 0 0 = −1 + 0 + 0 = −1 ≠ 0
2 N YUVARAJU 69
We observing, 𝑋1 , 𝑋2 and 𝑋1 , 𝑋3 are orthogonal, but 𝑋2 , 𝑋3 are not orthogonal.

𝑎
Now choose the eigenvector 𝑋3 = 𝑏 such that 𝑋1𝑇 𝑋3 = 0 and 𝑋2𝑇 𝑋3 = 0
𝑐
𝑎
∴ 𝑋1𝑇 𝑋3 = 2 −1 1 𝑏 = 0 ⇒ 2𝑎 − 𝑏 + 𝑐 = 0 ____(1)
𝑐
𝑎
and 𝑋2𝑇 𝑋3 = 1 2 0 𝑏 = 0 ⇒ 𝑎 + 2𝑏 = 0 _____(2)
𝑐
𝑎 −𝑏 𝑐
Solving (1) & (2), −1 1 = 2 1 = 2 −1 = 𝑡 (≠ 0, 𝑠𝑎𝑦)
2 0 1 0 1 2

𝑎 −𝑏 𝑐 𝑎 𝑏 𝑐
= = =𝑡 𝑜𝑟 = = =𝑡
−2 −1 5 −2 1 5

⇒ 𝑎 = −2𝑡, 𝑏 = 𝑡, 𝑐 = 5𝑡
N YUVARAJU 70
 −2𝑡 −2
Therefore, eigenvector 𝑋3 = 𝑡 = 1 𝑡 , corresponding to eigenvalue 𝜆 = 2.
5𝑡 5
Hence, the linearly independent mutually orthogonal eigenvectors of A are
2 1 −2
𝑋1 = −1 , 𝑋2 = 2 , 𝑋3 = 1
1 0 5

Length of eigenvector 𝑋1 = 𝑋1 = 22 + −1 2 + 12 = 6
Length of eigenvector 𝑋2 = 𝑋2 = 12 + 22 + 0 = 5
Length of eigenvector 𝑋3 = 𝑋3 = (−2)2 +12 + 52 = 30

2 −2
1
6 30
5
−1 1
The normalized eigenvectors are 𝑋1 = 6
, 𝑋2 = 2 , 𝑋3 = 30
5
1 5
6
0 30
N YUVARAJU 71
∴ The modal matrix P has normalized eigenvectors as its column vectors.
2 1 −2
6 5 30
−1 2 1
i.e., 𝑃 = 𝑋1 𝑋2 𝑋3 =
6 5 30
1 5
0
6 30
2 −1 1
6 6 6
1 2
Since P is orthogonal matrix => 𝑃−1 = 𝑃𝑇 = 5 5
0
−2 1 5
30 30 30
2 −1 1 2 1 −2
6 6 6 6 5 30
1 2
6 −2 2 −1 2 1
Now, 𝐷 = 𝑃𝑇 𝐴𝑃 = 5 5
0 −2 3 −1 6 5 30
−2 1 5 2 −1 3 1 5
30 30 30
0
6 30

8 0 0
= 0 2 0 ,which is diagonal matrix ( spectral matrix )
0 0 2
N YUVARAJU 72
Practice Problems:

Use orthogonal reduction the following symmetric matrices reduced to diagonal form.

1 0 0
1) 0 1 1
0 1 1

3 1 1
2) 1 3 −1
1 −1 3

N YUVARAJU 73
N YUVARAJU 74