Professional Documents
Culture Documents
Numerical Methods Lecture #11
Numerical Methods Lecture #11
Methods
Lecture-11
[MA-200]
ARISHA ALI
Numerical Computation of
Eigen Value
ARISHA ALI 2
Power Method
ARISHA ALI 3
Eigen Value and Eigen Vector
Let 𝐴 be the square matrix of order 𝑛 × 𝑛, then a number (real or complex) λ is
said to be eigen value of matrix 𝐴, if there exist a relation column matrix 𝑋 of
order 𝑛 × 1 such that
𝐴𝑋 − λ𝑋 = 0
𝐴 − λ𝐼 𝑋 = 0
Eigen value is also called proper value, latent value or characteristic value.
Eigen vector is also called proper vector, latent vector or characteristic vector.
ARISHA ALI 4
How to find Eigen value
Characteristic equation of matrix A
𝐴 − λ𝐼 = 0
Solution of characteristic equation is called eigen value.
For example:
0 1
𝐴=
1 0
−λ 1
𝐴 − λ𝐼 = = λ2 − 1 = λ = −1, 1
1 −λ
ARISHA ALI 5
How to find Eigen Vector
λ = λ1 , λ2 , … , λ𝑛
𝐴 − λ𝐼 𝑋 = 0 ------- (1)
Solution of equation (1) for any eigen value λ i.e., 𝑋 is called eigen vector of
matrix 𝐴 corresponding to λ = λ𝑖
For Example: For λ = −1
−λ 1 1 1
=
1 −λ 1 1
A𝑥 = 0
ARISHA ALI 6
How to find Eigen Vector
1 1 𝑥1 0
=
1 1 𝑥2 0
1 1 0
(𝐶𝑜𝑛𝑣𝑒𝑟𝑡 𝑖𝑛 𝑟𝑟𝑒𝑓)
1 1 0
1 1 0
0 0 0
𝑥1 = −1
𝑥2 = 1
𝑥1 −1
𝑥2 =
1
ARISHA ALI 7
How to find Eigen Vector
For λ = 1
−λ 1 −1 1
=
1 −λ 1 −1
A𝑥 = 0
−1 1 𝑥1 0
=
1 −1 𝑥2 0
−1 1 0
(𝐶𝑜𝑛𝑣𝑒𝑟𝑡 𝑖𝑛 𝑟𝑟𝑒𝑓)
1 −1 0
ARISHA ALI 8
How to find Eigen Vector
1 −1 0
0 0 0
𝑥1 = 1
𝑥2 = 1
𝑥1 1
𝑥2 =
1
−1
Eigen Value = −1, Eigen Vector =
1
1
Eigen Value = 1, Eigen Vector =
1
ARISHA ALI 9
Power Method
The power method is an iterative technique used to determine the dominant eigen
value of a matrix.
Dominant Eigen Value:
λ1 ≥ λ2 ≥ λ3 ≥ ⋯ ≥ λ𝑛
Then λ1 is known as dominant eigen value
ARISHA ALI 10
Working Rule
𝑇
Take an initial iterative vector 1 0 0…0 0 = 𝑦 (0)
Then the first iteration is 𝑦 (1) = 𝐴𝑦 (0) = 𝑘1 𝑧1 whereas 𝑘1 is the largest Eigen
value.
Then for second value we have to put 𝑧1 in the next step to get the second
iteration value 𝑦 (2) = 𝐴𝑧1 = 𝑘2 𝑧2
By continuing in this way, the iteration will stop if 𝑦 (𝑝) and 𝑦 (𝑝+1) are nearly
equal, then 𝑘 (𝑝+1) is called dominant eigen value and 𝑧 (𝑘+1) is called the
dominant eigen vector.
ARISHA ALI 11
Example
Apply power method to evaluate the dominant eigen value and eigen vector
2 2 1
1 3 1
1 2 2
Solution: Let
2 2 1
𝐴= 1 3 1
1 2 2
and take 𝑦 (0) = 1 0 0 𝑇 be an initial vector
ARISHA ALI 12
1st Iteration:
2 2 1 1 2+0+0 2 1
𝑦 (1) = 𝐴𝑦 (0) = 1 3 1 0 = 1 + 0 + 0 = 1 = 2 0.5
1 2 2 0 1+0+0 1 0.5
𝑘1 = 2, 𝑧1 = 1 0.5 0.5 𝑇
2nd Iteration:
ARISHA ALI 13
3rd Iteration:
2 2 1 1 4.58 1
𝑦 (3) = 𝐴𝑧2 = 1 3 1 0.86 = 4.44 = 4.58 0.97
1 2 2 0.86 4.44 0.97
𝑘3 = 4.58, 𝑧3 = 1 0.97 0.97 𝑇
4th Iteration:
2 2 1 1 4.91 1
𝑦 (4) = 𝐴𝑧3 = 1 3 1 0.97 = 4.88 = 4.91 0.99
1 2 2 0.97 4.88 0.99
𝑇
𝑘4 = 4.91, 𝑧4 = 1 0.99 0.99
ARISHA ALI 14
5th Iteration:
2 2 1 1 4.97 1 1
𝑦 (5) = 𝐴𝑧4 = 1 3 1 0.99 = 4.96 = 4.97 0.99 = 4.97 1
1 2 2 0.99 4.96 0.99 1
𝑘5 = 4.97, 𝑧5 = 1 1 𝑇
1
6th Iteration:
2 2 1 1 5 1
𝑦 (6) = 𝐴𝑧5 = 1 3 1 1 = 5 =5 1
1 2 2 1 5 1
𝑇
𝑘6 = 5, 𝑧6 = 1 1 1
ARISHA ALI 15
So dominant Eigen value is λ = 5
And dominant Eigen vector is 1 1 𝑇
1
ARISHA ALI 16
Practice Problems
Exercise 9.3
Q1, Q2
ARISHA ALI 17
Quiz # 2
ARISHA ALI 18