eigen values and eigen vector class note Dr. Motilal Panigrahi 18.01.

2022

So this means to solve the problem we need to solve

 det(𝐴 − 𝜆𝐼 ) = 0 … (1)
𝑎11 − 𝜆 𝑎12
| | = (𝑎11 − 𝜆)(𝑎22 − 𝜆) − 𝑎21 𝑎12
𝑎21 𝑎22 − 𝜆
= 𝜆2 − 𝜆(𝑎11 + 𝑎22 ) + (𝑎11 𝑎22 − 𝑎21 𝑎12 ) = 0
Degree of this polynomial is 2
(𝑎11 + 𝑎22 ) = 𝑇𝑟𝑎𝑐𝑒(𝐴)
(𝑎11 𝑎22 − 𝑎21 𝑎12 ) = det(𝐴)
Euation (1) is known as the Characteristic equation of matrix
A.
Characteristic equation has two roots and the roots are
known as characteristic values or eigen values.

If A is n x n square matrix then you get a Characteristic

polynomial of degree n.
Characteristic equation has n roots and the roots are known
as characteristic values or eigen values.
Suppose the roots are 𝜆1 and 𝜆2 .
When 𝜆 = 𝜆1 , we have got the system of equations
(𝑎11 − 𝜆1 )𝑥1 + 𝑎12 𝑥2 = 0
𝑎21 𝑥1 + (𝑎22 − 𝜆1 )𝑥2 = 0
What is the rank of this system? That is, to say, What is the
rank of the coefficient matrix?
Note that we desire that the solution has to be non-zero,
means non-trivial solution is desired.
So
Rank <=1

If rank is 1, you get one non-trivial solution.

Take an example.

1 3
𝐴=[ ]
4 20
Characteristic equation is
𝜆2 − 21𝜆 + 8 = 0
−(−21) ± √441 − 4(1)(8)
𝜆=
2
21 ± √409
=
2
= 𝛼, 𝛽
are the eigen values of A.
Take 𝜆 = 𝛼
(1 − 𝛼 )𝑥1 + 3𝑥2 = 0
4𝑥1 + (20 − 𝛼 )𝑥2 = 0

3𝑥2
𝑥1 = −
(1 − 𝛼 )
 −3
𝑥1 −3
𝑢
⃗ = [𝑥 ] = [ ]=[ 21 + √409 ]
2 (1 − 𝛼 ) (1 − )
2
−3
= [ −19 − √409 ]
( )
2
is an eigen vector corresponding to the eigen value 𝛼.
Take 𝜆 = 𝛽
(1 − 𝛽 )𝑥1 + 3𝑥2 = 0
4𝑥1 + (20 − 𝛽 )𝑥2 = 0

3𝑥2
𝑥1 = −
(1 − 𝛽 )
−3
𝑥1 −3
𝑣 = [ 𝑥 ] = [( )] = [ −19 + √409 ]
2 1 − 𝛽 ( )
2

is an eigen vector corresponding to the eigen value 𝛽.

If the eigen values are distinct, then we have got two eigen
vectors.

An eigen space is formed by taking the span of the eigen

vectors.
Example: Find the eigen values and eigen vectors of the
matrix.

1 2 3
𝐴 = [1 0 1]
2 1 3
1−𝜆 2 3
𝐴 − 𝜆𝐼 = [ 1 0−𝜆 1 ]
2 1 3−𝜆
Characteristic equation is
𝜆3 − 𝑆1 𝜆2 + 𝑆2 𝜆 − |𝐴| = 0
𝑆1 = 𝑇𝑟𝑎𝑐𝑒(𝐴) = 1 + 0 + 3 = 4
0 1 1 3 1 2
𝑆2 = | |+| |+| |
1 3 2 3 1 0
= −1 − 3 − 2 = −6
|𝐴| = 1(−1) − 2(1) + 3(1) = 0
𝜆3 − (4)𝜆2 + (−6)𝜆 − 0 = 0
𝜆(𝜆2 − 4𝜆 − 6) = 0
4 ± √16 + 24
𝜆 = 0,
2
= 0, 𝛼, 𝛽
𝛼 = 2 + √10
𝛽 = 2 − √10
To find eigen vectors:
i) Take 𝜆 = 0,
 1 2 3 𝑥1 0
(𝐴 − 𝜆𝐼 )𝑥 = [1 0 1] [𝑥2 ] = [0]
2 1 3 𝑥3 0
1 2 3
𝑅2 − 𝑅1 [0 −2 −2]
𝑅3 − 2𝑅1 0 −3 −3
1 2 3
[0 −2 −2]
2𝑅3 − 3𝑅2 0 0 0
1 2 3
𝑅2/−2 [0 1 1]
0 0 0
𝑥2 = −𝑥3
𝑥1 = −2𝑥2 − 3𝑥3 = −2(−𝑥3 ) − 3𝑥3 = −𝑥3
𝑥1 −𝑡 1
[𝑥2 ] = [−𝑡] = −𝑡 [ 1 ]
𝑥3 𝑡 −1
1
[ 1 ] is an eigen vector corresponding to the eigen value 0.
−1
ii) To find eigen vectors: Consider 𝜆 = 𝛼 = 2 + √10,
1−𝛼 2 3 𝑥1 0
(𝐴 − 𝛼𝐼 )𝑥 = [ 1 0−𝛼 1 ] [𝑥2 ] = [0]
2 1 3 − 𝛼 𝑥3 0
−1 − √10 2 3
[ 1 −2 − √10 1 ]
2 1 1 − √10
As we know that the rank is not 3, and the second row is not
a multiple of 1st row, so we discard the 3rd row, and apply
Cramer’s rule, to get general solution
𝑥1 −𝑥2
=
(2)(1) − (3)(−2 − √10) (−1 − √10)(1) − (3)(1)
𝑥3
= = 𝑡 (𝑠𝑎𝑦)
(−1 − √10) (−2 − 10) − (2)(1)
√

𝑥1 2 + 6 + 3√10
[𝑥2 ] = 𝑡 [ 1 + √10 + 3 ]
𝑥3 3√10 + 10
Thus

8 + 3√10
[ 4 + √10 ] is an eigen vector corresponding to the eigen
10 + 3√10
value 2 + √10.
iii) To find eigen vectors: Consider 𝜆 = 𝛽 = 2 − √10,
1−𝛽 2 3 𝑥1 0
(𝐴 − 𝛽𝐼 )𝑥 = [ 1 0−𝛽 1 ] [𝑥2 ] = [0]
2 1 3 − 𝛽 𝑥3 0
−1 + √10 2 3
[ 1 −2 + √10 1 ]
2 1 1 + √10
As we know that the rank is not 3, and the second row is not
a multiple of 1st row, so we discard the 3rd row, and apply
Cramer’s rule, to get general solution
 𝑥1 −𝑥2
=
(2)(1) − (3)(−2 + √10) (−1 + √10)(1) − (3)(1)
𝑥3
= = 𝑡 (𝑠𝑎𝑦)
(−1 + √10) (−2 + √10) − (2)(1)

𝑥1 2 + 6 − 3√10
[𝑥2 ] = 𝑡 [ 1 − √10 + 3 ]
𝑥3 −3√10 + 10
Thus

8 − 3√10
[ 4 − √10 ] is an eigen vector corresponding to the eigen
10 − 3√10
value 2 − √10.
Note: From the above you can observe that if you have found
a solution for 𝛼, then a solution for 𝛽 can be written directly,
just by changing the sign of the surd.
Thus
Eigen values 0 2 + √10 2 − √10
Eigen vectors 1 8 + 3√10 8 − 3√10
[1] [ 4 + √10 ] [ 4 − √10 ]
−1 10 + 3√10 10 − 3√10

Example: Find the eigen values and eigen vectors of the

matrix.
2 1 3
A= 0 0 -6
1 1 4
 1 1 -4
B= 2 1 6
1 0 4

Type equation here.

2 2 -6
C= 2 1 7
1 0 4

2 -2 2
D= 2 1 5
1 0 4
Note that multiplicity(algebraic) depends on the number of
eigen value is repeated, and geometric multiplicity 9depends
on number of free variables), which will be explained in next
class.
 2 −12
𝐴=[ ]
1 −5
Characteristic equation is
𝜆2 − (−3)𝜆 + 2 = 0
−3 ± √9 − (8)
𝜆=
2
−3 ± 1
= = −2, −1
2
are the eigen values of A.
Take 𝜆 = −2
(2 − (−2))𝑥1 − 12𝑥2 = 0
𝑥1 + (−5 − (−2))𝑥2 = 0
That is,
4𝑥1 − 12𝑥2 = 0
𝑥1 − 3𝑥2 = 0
Thus
𝑥1 = 3𝑥2
𝑥1 3
𝑢
⃗ = [𝑥 ] = 𝑡 [ ]
2 1
3
Thus [ ] is an eigen vector corresponding to the eigen value
1
−2.
Take 𝜆 = −1
(2 − (−1))𝑥1 − 12𝑥2 = 0
 𝑥1 + (−5 − (−1))𝑥2 = 0
That is,
3𝑥1 − 12𝑥2 = 0
𝑥1 − 4𝑥2 = 0
Thus
𝑥1 = 4𝑥2
𝑥1 4
𝑢
⃗ = [𝑥 ] = 𝑡 [ ]
2 1
4
Thus [ ] is an eigen vector corresponding to the eigen value
1
−1.