Chapter-8

Matrix Eigenvalue Problems

8.1 Eigenvalues and Eigenvectors
Eigen values:
Let 𝐴 = [𝑎𝑗𝑘 ] be an 𝑛 × 𝑛 matrix and 𝜆 be a scalar. Then 𝐷(𝜆) = det(𝐴 − 𝜆 𝐼) = |𝐴 − 𝜆𝐼| = 0
is called the characteristic equation of 𝐴.
A root of the characteristic equation, |𝐴 − 𝜆𝐼| = 0 is called an eigenvalue of characteristic value
of A.
𝑎11 − 𝜆 𝑎12 ⋯ 𝑎1𝑛
𝑎 𝑎22 − 𝜆 ⋯ 𝑎2𝑛
Here, |𝐴 − 𝜆𝐼| = | 21 |.
⋮ ⋮ ⋮ ⋮
𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛 − 𝜆
Note: An 𝑛 × 𝑛 matrix 𝐴 has at least one eigenvalue and at most 𝑛 numerically different
eigenvalues.
Examples:
1 3
Find the eigenvalues of the matrix𝐴 = [ ].
4 5
Solution: The characteristic equation of A is |𝐴 − 𝜆𝐼| = 0
1−𝜆 3
⇒| |=0
4 5−𝜆
⇒ (1 − 𝜆)(5 − 𝜆) − 12 = 0
⇒ 𝜆2 − 6𝜆 − 7 = 0
⇒ (𝜆 − 7)(𝜆 + 1) = 0
⇒ 𝜆 = −1, 7
Hence the eigenvalues of 𝐴 are −1, 7.
Example:
1 −1 0
Find the eigen values of the matrix𝐴 = [1 2 −1].
3 2 −2
Solution: The characteristic equation of A is |𝐴 − 𝜆𝐼| = 0
1−𝜆 −1 0
⇒| 1 2−𝜆 −1 | = 0
3 2 −2 − 𝜆
⇒ (1 − 𝜆){−(2 − 𝜆)(2 + 𝜆) + 2} + 1(−2 − 𝜆 + 3) = 0
⇒ (1 − 𝜆)(𝜆2 − 2) + (1 − 𝜆) = 0
⇒ (1 − 𝜆)(𝜆2 − 1) = 0
⇒ 𝜆 = 1, 1, −1
Hence the eigenvalues of 𝐴 are1, 1 − 1.
Eigen Vectors:
Let 𝐴 be an 𝑛 × 𝑛 matrix. A non-null vector 𝑋 which is a solution of the vector equation
𝐴𝑋 = 𝜆𝑋, is called an eigenvector or characteristic vector of 𝐴. Here 𝜆 is a scalar.
Example:
−5 2
Find the eigenvalues and eigenvectors of 𝐴 = [ ]
2 −2
Solution: The characteristic equation of A is |𝐴 − 𝜆𝐼| = 0
−5 − 𝜆 2
⇒| |=0
2 −2 − 𝜆
⇒ 𝜆2 + 7𝜆 + 6 = 0
⇒ 𝜆 = −1, −6
𝑥1
Let 𝑋 = [𝑥 ] be an eigen vector corresponding to𝜆 = −1. Then
2
𝐴𝑋 = 𝜆𝑋
−5 2 𝑥1 𝑥1
⇒[ ] [𝑥 ] = −1 [𝑥 ]
2 −2 2 2
⇒ −4𝑥1 + 2𝑥2 = 0, 2𝑥1 − 𝑥2 = 0
⇒ 2𝑥1 − 𝑥2 = 0
Let 𝑥1 = 𝑘, then 𝑥2 = 2𝑘 where, 𝑘 ≠ 0.
𝑥1 𝑘 1
So, [𝑥 ] = [ ] = 𝑘 [ ]
2 2𝑘 2
1
Hence the eigenvector corresponding to 𝜆 = −1 is[ ].
2
𝑥1
Similarly, let 𝑋 = [𝑥 ] be an eigenvector corresponding to𝜆 = −6. Then
2
𝐴𝑋 = 𝜆𝑋
−5 2 𝑥1 𝑥1
⇒[ ] [𝑥 ] = −6 [𝑥 ]
2 −2 2 2
⇒ 𝑥1 + 2𝑥2 = 0, 2𝑥1 + 4𝑥2 = 0
⇒ 𝑥1 + 2𝑥2 = 0
Let 𝑥1 = 𝑘, then 𝑥2 = −𝑘/2 where, 𝑘 ≠ 0.
𝑥1 𝑘 2
So, [𝑥 ] = [ ] = 𝑘/2 [ ]
2 −𝑘/2 −1
2
Hence the eigenvector corresponding to 𝜆 = −6 is[ ].
−1
Example:
1 1 −2
Find the eigenvalues and eigenvectors of 𝐴 = [−1 2 1 ]
0 1 −1
Solution: The characteristic equation of A is |𝐴 − 𝜆𝐼| = 0
1−𝜆 1 −2
⇒ | −1 2 − 𝜆 1 |=0
0 1 −1 − 𝜆
⇒ (1 − 𝜆){(2 − 𝜆)(−1 − 𝜆) − 1} − 1(𝜆 + 1 − 0) − 2(−1 − 0) = 0
⇒ (1 − 𝜆){(2 − 𝜆)(−1 − 𝜆) − 1} + (1 − 𝜆)=0
⇒ (1 − 𝜆)(𝜆2 − 𝜆 − 2) = 0
⇒ (1 − 𝜆)(𝜆 − 2)(𝜆 + 1) = 0
⇒ 𝜆 = 1, 2, −1
𝑥1
Let 𝑋 = [𝑥2 ] be an eigenvector corresponding to𝜆 = 1. Then
𝑥3
𝐴𝑋 = 𝜆𝑋
1 1 −2 𝑥1 𝑥1
⇒ [−1 2 1 ] [𝑥2 ] = 1 [𝑥2 ]
0 1 −1 𝑥3 𝑥3
⇒ 0𝑥1 + 𝑥2 − 2𝑥3 = 0, …………………(1)
−𝑥1 + 𝑥2 + 𝑥3 = 0, ………………….(2)
0𝑥1 + 𝑥2 − 2𝑥3 = 0 ………………….(3)
Solving Eqs.(1) and (2) by cross multiplication method, we get
𝑥1 𝑥2 𝑥3
= =
1+2 2−0 0+1
𝑥 𝑥 𝑥
⇒ 31 = 22 = 13 = 𝑘 (say)
Then 𝑥1 = 3𝑘, 𝑥2 = 2𝑘, 𝑥3 = 𝑘.
You can easily verify that this solution is also satisfied Eq. (3).
𝑥1 3𝑘 3
𝑥
So, [ 2 ] = [2𝑘] = 𝑘 [2]
𝑥3 𝑘 1
3
Hence the eigenvector corresponding to 𝜆 = 1 is[2].
1
𝑥1
Let 𝑋 = [𝑥2 ] be an eigen vector corresponding to𝜆 = 2. Then
𝑥3
𝐴𝑋 = 𝜆𝑋
1 1 −2 𝑥1 𝑥1
𝑥
⇒ [−1 2 1 ] [ 2 ] = 2 [ 2 ] 𝑥
0 1 −1 𝑥3 𝑥3
⇒ −𝑥1 + 𝑥2 − 2𝑥3 = 0, …………………(1)
−𝑥1 + 0𝑥2 + 𝑥3 = 0, ………………….(2)
0𝑥1 + 𝑥2 − 3𝑥3 = 0 ………………….(3)
Solving Eqs.(1) and (2) by cross multiplication method, we get
𝑥1 𝑥2 𝑥3
= =
1−0 2+1 0+1
𝑥 𝑥 𝑥
⇒ 11 = 32 = 13 = 𝑘 (say)
Then 𝑥1 = 𝑘, 𝑥2 = 3𝑘, 𝑥3 = 𝑘.
You can easily verify that this solution is also satisfied Eq. (3).
𝑥1 𝑘 1
So, [𝑥2 ] = [3𝑘] = 𝑘 [3]
𝑥3 𝑘 1
1
Hence the eigenvector corresponding to 𝜆 = 2 is[3].
1
Similarly, you can obtain the eigen value corresponding to 𝜆 = −1.
1
The eigenvalue corresponding to 𝜆 = −1 will be [0]
1
Eigen Basis (Basis of eigenvectors):
Let 𝐴 be an 𝑛 × 𝑛 matrix. Then an eigen basis is a basis of ℝ𝑛 consisting of eigenvectors of A.
If 𝐴 has 𝑛 distinct eigenvalues, then 𝐴 has a basis of eigenvectors 𝑋1 , 𝑋2 , … , 𝑋𝑛 for ℝ𝑛 .
Eigen space:
The eigenvectors corresponding to the same eigenvalue 𝜆 of 𝐴 together with the null vector 𝟎,
form a vector space, called the Eigen space of A, corresponding to that 𝜆. It is usually denoted
by𝐸𝜆 .
Algebraic Multiplicity and Geometric Multiplicity:
The order of an eigenvalue 𝜆 is called the algebraic multiplicity of 𝜆 and it is denoted by 𝑀𝜆 .
The number of linearly independent eigenvectors corresponding to an eigenvalue 𝜆 is called the
geometric multiplicity of 𝜆 and it is denoted by 𝑚𝜆 . Thus, the geometric multiplicity is the
dimension of the Eigen space corresponding to this 𝜆.
An eigenvalue 𝜆 is regular if 𝑀𝜆 = 𝑚𝜆 .
Example: Find the algebraic multiplicity and geometric multiplicity of the eigenvalues of the
3 2
matrix 𝐴 = [ ].
0 3
Solution: The characteristic equation of A is |𝐴 − 𝜆𝐼| = 0
3−𝜆 2
⇒| |=0
0 3−𝜆
2
⇒ (𝜆 − 3) = 0
⇒ 𝜆 = 3, 3
Hence the algebraic multiplicity of 𝜆 = 3 is 𝑀3 = 2.
𝑥1
Let 𝑋 = [𝑥 ] be an eigen vector corresponding to 𝜆 = 3. Then
2
𝐴𝑋 = 𝜆𝑋
3 2 𝑥1 𝑥1
⇒[ ] [𝑥 ] = 3 [𝑥 ]
0 3 2 2
⇒ 2𝑥2 = 0, 3𝑥2 = 3𝑥2
⇒ 𝑥2 = 0
𝑥1 𝑥 1
So, 𝑋 = [𝑥 ] = [ 1 ] = 𝑥1 [ ]
2 0 0
1
The eigen vector corresponding to 𝜆 = 3 is [ ].
0
Hence the geometric multiplicity of 𝜆 = 3 is 𝑚3 = 1.
Results:
1: If 𝐴 is a 𝑛 × 𝑛 triangular matrix -upper triangular, lower triangular or diagonal, the
eigenvalues of 𝐴 are the diagonal entries of 𝐴.
2: 𝜆 = 0 is an eigenvalue of 𝐴 if 𝐴 is a singular (noninvertible) matrix.
3: 𝐴 and 𝐴𝑇 have the same eigenvalues.
4: det 𝐴 is the product of the eigenvalues of 𝐴.
5: If 𝜆 be an eigenvalue of 𝐴, then
a) 𝑘𝜆 is an eigenvalue of 𝑘𝐴, 𝑘 =scalar.
b) 𝜆𝑚 is an eigenvalue of 𝐴𝑚 .
1
c) 𝜆 is the eigenvalue of 𝐴−1 .
d) 𝜆 − 𝑘 is the eigenvalue of 𝐴 − 𝑘𝐼.
e) 𝑘𝑚 𝜆𝑚 + 𝑘𝑚−1 𝜆𝑚−1 + ⋯ + 𝑘1 𝜆 + 𝑘0 is the eigenvalue of
k m Am + k m −1 Am −1 +  + k1 A + k0 I .
 8.3. Symmetric, Skew-Symmetric, and Orthogonal Matrices

Symmetric, Skew-Symmetric, and Orthogonal Matrices.

A real square matrix A = [ajk] is called

1. symmetric if transposition leaves it unchanged,

AT = A, thus akj = ajk,

2. skew-symmetric if transposition gives the negative of A,

AT = −A, thus akj = -ajk,

3. orthogonal if transposition gives the inverse of A,

AT = A−1.
Example: The matrices

− 3 1 5 0 9 − 12  2 1 2 
     3 3 3 
     
− 9 20 , and  − 2
0 − 2 , 0 2 1 
1
     3 3 3 
 5 − 2 4   12 − 20 0   1 2 2
 − 
 3 3 3
are symmetric, skew-symmetric, and orthogonal, respectively.

Note:
1. The determinant of an orthogonal matrix has the value +1 or −1.
Proof: Let A be an orthogonal matrix  AT = A−1 or AAT = I .
We have det( I ) = 1 .
 1 = det( I ) = det( AA−1 ) = det( AAT ) = det( A) det( AT ) = det( A) det( A) = (det( A) )2
 det( A) = 1

2. Every square matrix can be expressed as the sum of a symmetric and a skew-symmetric
matrix.
A + AT A − AT
A= +
2 2
2 1 5
 
Example: Express A = 4 3 2 as the sum of a symmetric and a skew-symmetric matrix.
 
1 − 1 0
2 1 5 2 4 1 

Solution: Given matrix is A = 4 3 2
  A = 1 3 − 1
T
 
1 − 1 0 5 2 0 
 2 1 5 2 4 1    4 5 6
A + AT 1   1
  
= 4 3 2 + 1 3 − 1  = 5 6 1 ,
2 2  2 6 1 0
1 − 1 0 5 2 0    
It is a symmetric matrix because a jk = akj , j  k .
 2 1 5  2 4 1    0 − 3 4
A − AT 1      1
= 4 3 2 − 1 3 − 1  =  3 0 3 ,
2 2  2 − 4 − 3 0
1 − 1 0 5 2 0    
It is a skew-symmetric matrix because a jk = −akj , j  k .
2 1 5   4 5 6  0 − 3 4 
1  
 
 A = 4 3 2 = 5 6 1 +  3 0 3 
2 
1 − 1 0 6 1 0 − 4 − 3 0 

Eigenvalues of Symmetric and Skew-Symmetric and Orthogonal Matrices:

(a) The eigenvalues of a symmetric matrix are real.

(b) The eigenvalues of a skew-symmetric matrix are pure imaginary or zero.

(c) The eigenvalues of an orthogonal matrix A are real or complex conjugates in pairs
and have absolute value 1.

Example:

5 3
The matrix   is symmetric and has eigenvalues 2 and 8.
3 5
 0 1
The matrix   is skew-symmetric and has eigenvalues -ⅈ and ⅈ.
 − 1 0
Example: The orthogonal matrix

 2 1 2 
 3 3 3 
 
− 2 2 1 
 3 3 3 
 1 2 2
 − 
 3 3 3

has eigenvalues -1, (5+ⅈ√11)/6 and (5-ⅈ√11)/6, which have absolute value 1.
 8.5. Complex Matrices and Different Forms.

Complex Conjugate Matrix

Ā = [ājk] is obtained from A = [ajk] by replacing each entry ajk = α + iβ (α, β real) with its complex
conjugate ājk = α − iβ. Also, ĀT = [ākj] is the transpose of Ā, hence the conjugate transpose of A.

Hermitian, Skew-Hermitian, and Unitary Matrices.

A square matrix A = [akj] is called

1. Hermitian: if ( A )T = A, akj = a jk .
2. Skew-Hermitian: if ( A )T = − A, akj = −a jk .
3. Unitary: if ĀT = A−1.
Note:
1. Let A be a unitary matrix. Then its determinant has absolute value one, that is,
|det A| = 1.
Proof: Let A be a unitary matrix.
 ( A )T = A−1 or A( A )T = I .
We have det( I ) = 1.
 1 = det( I ) = det( AA−1 ) = det( A( A )T ) = det( A) det(( A )T )
2
= det( A) det( A ) = det( A)det( A) = det( A)
 det( A) = 1

2. Every complex square matrix can be expressed as the sum of Hermitian and Skew-
Hermitian matrices.
A + ( A )T A − ( A )T
A= +
2 2
 2 + 3i 5+i 
Example: Express A =   as the addition of Hermitian and Skew-Hermitian matrices.
7 + 2i − 3 + 5i 
 2 + 3i 5+i   2 − 3i 5−i 
Solution: Given matrix is A =   A= 
7 + 2i − 3 + 5i  7 − 2i − 3 − 5i 
2 − 3i 7 − 2i 
 ( A )T =  
 5 − i − 3 − 5i 
A + ( A )T 1  2 + 3i 5 + i  2 − 3i 7 − 2i   1  4 12 − i 
=  + = 
2  
2 7 + 2i − 3 + 5i   5 − i − 3 − 5i   2 12 + i − 6 
It is a Hermitian matrix because the diagonal elements are real and akj = a jk , j  k .
A − ( A )T 1  2 + 3i 5 + i  2 − 3i 7 − 2i   1  6i − 2 + 3i 
=  − = 
2 2 7 + 2i − 3 + 5i   5 − i − 3 − 5i   2 2 + 3i − 10i 
 
It is a Skew-Hermitian matrix because the diagonal elements are purely imaginary and
akj = −a jk , j  k .
 2 + 3i 5 + i  1  4 12 − i   6i − 2 + 3i  
 A=  =   + 
7 + 2i − 3 + 5i  2 12 + i − 6  2 + 3i − 10i  

Eigenvalues of Hermitian, Skew-Hermitian and Unitary Matrices.

(a) The eigenvalues of a Hermitian matrix (and thus of a symmetric matrix) are real.
(b) The eigenvalues of a skew-Hermitian matrix (and thus of a skew-symmetric matrix) are pure
imaginary or zero.
(c) The eigenvalues of a unitary matrix (and thus of an orthogonal matrix) have absolute value 1.

Example:

 4 1 − 3i 
The matrix   is Hermitian and has eigenvalues 9 and 2.
 
1 + 3i 7 

 3i 2 + i
The matrix   is skew-Hermitian and has eigenvalues 4ⅈ and -2ⅈ.
 
− 2 + i − i 
 1 1 
 2i 3
2  is unitary and has eigenvalues (√3+ⅈ)/2 and (-√3+ⅈ)/2, which have
The matrix  
1 3 1 
i
 2 2 
absolute value 1.

Different Forms
1. Quadratic Form
Quadratic form is an expression of the form
𝑄 = 𝑥 𝑇 𝐴𝑥
𝑎11 𝑎12 ⋯ 𝑎1𝑛 𝑥1
𝑎21 𝑎22 ⋯ 𝑎2𝑛 𝑥2
Where𝐴 = [ ⋮ ⋮ ⋮ ⋮ ]and 𝑥 = [ ⋮ ]
𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛 𝑥𝑛
⇒ 𝑄 = 𝑥 𝑇 𝐴𝑥 = 𝑎11 𝑥12 + 𝑎12 𝑥1 𝑥2 + ⋯ + 𝑎1𝑛 𝑥1 𝑥𝑛 + 𝑎21 𝑥1 𝑥2 + 𝑎22 𝑥22 + ⋯ + 𝑎𝑛1 𝑥1 𝑥𝑛
+ ⋯ + 𝑎𝑛𝑛 𝑥𝑛2
⇒ 𝑄 = 𝑥 𝐴𝑥 = 𝑎11 𝑥12 + 𝑎22 𝑥22 + ⋯ + 𝑎𝑛𝑛 𝑥𝑛2 + (𝑎12 + 𝑎21 )𝑥1 𝑥2 + (𝑎13 + 𝑎31 )𝑥1 𝑥3
𝑇

+ ⋯ + (𝑎1𝑛 + 𝑎𝑛1 )𝑥1 𝑥𝑛

In particular
𝑎11 𝑎12 𝑎13 𝑥1
Let𝐴 = [𝑎21 𝑎22 𝑎23 ]and𝑥 = [𝑥2 ]
𝑎31 𝑎32 𝑎33 𝑥3
⇒ 𝑄 = 𝑥 𝐴𝑥 = 𝑎11 𝑥1 + 𝑎22 𝑥2 + 𝑎33 𝑥32 + (𝑎12 + 𝑎21 )𝑥1 𝑥2 + (𝑎13 + 𝑎31 )𝑥1 𝑥3 + (𝑎23
𝑇 2 2

+ 𝑎32 )𝑥2 𝑥3
Determination of Symmetric matrix from Quadratic form
From the given Quadratic form𝑄 = 𝑥 𝑇 𝐴𝑥we can obtain symmetric matrix as follows:
I. Place coefficients of 𝑥𝑖2 diagonally.
II. Divide the coefficients of 𝑥𝑖 𝑥𝑗 , 𝑖 ≠ 𝑗 equally and place equal values at 𝑎𝑖𝑗 and
𝑎𝑗𝑖 places respectively.
Example: Find the symmetric matrix from the following quadratic form
𝑄 = 2𝑥12 + 4𝑥22 + 3𝑥32 + 10𝑥1 𝑥2 + 12𝑥2 𝑥3 + 8𝑥3 𝑥1
Solution: In case of symmetric matrix𝑎𝑖𝑗 = 𝑎𝑗𝑖
⇒ 𝑄 = 2𝑥12 + 4𝑥22 + 3𝑥32 + (5 + 5)𝑥1 𝑥2 + (6 + 6)𝑥2 𝑥3 + (4 + 4)𝑥3 𝑥1
The symmetric matrix is
2 5 4
𝐴 = [5 4 6]
4 6 3
2. Hermitian Form
Hermitian form 𝐻 is given by
𝐻 = 𝑥̄ 𝑇 𝐴𝑥
Where 𝐴 is a Hermitian matrix and 𝑥is a column matrix.
Note: Hermitian form is always real.
Example: Is the following matrix Hermitian or Skew-Hermitian matrix? Find the Hermitian
form or Skew-Hermitian form.
4 3 − 2𝑖 −2𝑖
𝐴=[ ], 𝑥 = [ ]
3 + 2𝑖 −4 1+𝑖
4 3 + 2𝑖 ̄ 𝑇 4 3 − 2𝑖
Solution: 𝐴̄ = [ ], 𝐴 = [ ]=𝐴
3 − 2𝑖 −4 3 + 2𝑖 −4
Since 𝐴̄𝑇 = 𝐴, so 𝐴 is Hermitian.
Hermitian form is 𝐻 = 𝑥̄ 𝑇 𝐴𝑥
−2𝑖 2𝑖
Here 𝑥 = [ ] ⇒ 𝑥̄ = [ ]
1+𝑖 1−𝑖
4 3 − 2𝑖 −2𝑖 −2𝑖
⇒ 𝐻 = 𝑥̄ 𝑇 𝐴𝑥 = [2𝑖 1 − 𝑖 ] [ ][ ] = [7𝑖 + 5 10𝑖 ] [ ]=4
3 + 2𝑖 −4 1+𝑖 1+𝑖
3. Skew-Hermitian Form
Skew-Hermitian form S is given by
𝑆 = 𝑥̄ 𝑇 𝐴𝑥
Where A is a Skew-Hermitian matrix and x is a column matrix.
Note: Skew-Hermitian form is always 0 or purely imaginary.
Example: Is the following matrix Hermitian or Skew-Hermitian matrix? Find the Hermitian
form or Skew-Hermitian form.
𝑖 −2 + 3𝑖 𝑖
𝐴=[ ], 𝑥 = [ ]
2 + 3𝑖 0 4
 −i − 2 − 3i  T  − i 2 − 3i   i − 2 + 3i 
Solution: A =   , A =  = − = −A
2 − 3i 0  − 2 − 3i 0  2 + 3i 0 
Since A T = − A , so A is Skew-Hermitian.
Skew-Hermitian form is S = x T Ax
i  − i 
Here x =    x =  
 4 4
  i − 2 + 3i   i  i 
 S = x T Ax = − i 4    = 9 + 12i 2i + 3  = 17i
2 + 3i 0   4  4

8.4 Eigen basis. Diagonalization

Similarity of Matrices
A square matrix  of order n is said to be similar to a n  n square matrix A if there exists a non-
singular matrix P such that
Aˆ = P −1 AP
This transformation of A to  is known as a Similarity transformation.
Note:
1. Eigenvalues of A = Eigenvalues of Â
2. If X is an eigenvector of A corresponding to the eigenvalue  i.e. AX = X , X  0 then
Y = P −1 X is an eigenvector of  corresponding to the same eigenvalue  , means
Aˆ ( P −1 X ) =  ( P −1 X ), P −1 X = Y  0

Basis of Eigenvectors
If A is a square matrix of order n and it has n distinct eigenvalues then the corresponding n
eigenvectors are L. I. and they will form a basis which is known as basis of eigenvectors or
Eigen basis.

Note:
1. A symmetric matrix has an orthonormal basis of eigenvectors.
2. A Hermitian, Skew-Hermitian or unitary matrix has a basis of eigenvectors which is a
unitary system or unitary Eigen basis.
3. A basis of eigenvectors is possible if algebraic multiplicity of  = geometric multiplicity.
0 1 
Example: Let A =  
0 0 
Eigenvalues
The characteristic equation is A − I = 0.
− 1
 =0  2 = 0   = 0, 0
0 −
Hence for  = 0 , A. M. =2
Eigenvectors
The homogeneous system is ( A − I ) X = 0, X  0.
0 1  x1  0  0 x1 + x2 = 0
    = 
0 0  x2  0 0 x1 + 0 x2 = 0
Here x2 = 0 but x1 is arbitrary.
x  1
 X =  1  = x1   , This is the only one eigenvector.
0 0 
Hence G. M. =1
Therefore A cannot provide a basis of eigenvectors.
Diagonalization
Let A be any square matrix and X be the Modal matrix whose column vectors are eigenvectors of
A . Then the Diagonalization of the matrix A is denoted by D and defined by
D = X −1 AX
Note:
1. Only those matrices can be diagonalized if there exists a basis of eigenvectors.
2. A square matrix A of order n with n L. I. eigenvectors is similar to a diagonal matrix D
whose diagonal elements are eigenvalues of A .
3. Powers of a matrix A
We have D = X −1 AX
 D 2 = DD = ( X −1 AX )( X −1 AX ) = X −1 A2 X
Similarly D 3 = X −1 A3 X
In general, D n = X −1 An X
To obtain An we have to pre-multiply by X and post multiply by X −1
 XDn X −1 = XX −1 An XX −1 = An
Hence An = XDX −1 .
Procedure of Diagonalization
Any square matrix which has a basis of eigenvectors can be diagonalized as follows:
I. Find the eigenvalues of the given matrix A
II. Find eigenvectors.
III. Form a Modal matrix X by taking eigenvectors as its columns.
IV. Calculate X −1 .
V. The diagonal matrix is D = X −1 AX (Main diagonal elements are eigenvalues of A ).
Example: Diagonalize the following matrix
2 1
A= 
2 1
Solution: Eigenvalues
The characteristic equation is A − I = 0.
2− 1
 =0  (2 −  )(1 −  ) − 2 = 0
2 1− 
 2 − 3 = 0
The eigenvalues are 1 = 0 and 2 = 3.
Eigenvector for 1 = 0
The homogeneous system is ( A − 1I ) X1 = 0, X1  0.
2 1  x1  0  2 x1 + x2 = 0
 =
   
2 1  x2  0 2 x1 + x2 = 0
Both the equations are identical.
 x2 = −2x1 , Here x1 is arbitrary.
 x   x  1
X1 =  1  =  1  = x1  
 x2  − 2 x1   − 2
The eigenvector for 1 = 0 is
1
X1 =  
 − 2

Eigenvector for 2 = 3
The homogeneous system is ( A − 2 I ) X 2 = 0, X 2  0.
− 1 1   x1  0  − x1 + x2 = 0
   =  
 2 − 2  x2  0 2 x1 − 2 x2 = 0
Both equations are identical.
 x2 = x1 , Here x1 is arbitrary.
 x  x  1
X 2 =  1  =  1  = x1  
 x2   x1  1
The eigenvector for 2 = 3 is
1
X2 =  
1
 1 1
The modal matrix is X = X 1 X2=  
− 2 1
 X =1+ 2 = 3
1 1 − 1 1 / 3 − 1 / 3
 X −1 =  =
3 2 1  2 / 3 1 / 3 
1 / 3 − 1 / 3 2 1  1 1 1 / 3 − 1 / 3 0 3 0 0
 D = X −1 AX =    =  = 
2 / 3 1 / 3  2 1 − 2 1 2 / 3 1 / 3  0 3 0 3
This is the required diagonal matrix whose diagonal elements are the eigenvalues of the matrix A
.

************************************END*************************************