Chapter 7 (Eigenvalues and Eigenvectors)

Exercise 7.1: 5-8,17,18,30 ; Exercise 7.2: 1-3,5-20.

Objective of Lecture:-
(1) To find eigenvalues and eigenvectors of a matrix.
(2) Cayley-Hamilton Theorem.
(3) Diagonalization of Matrix.
Skills Needed:- Determinant (Ch-3), Homogeneous system (Ex-4.7), Inverse of matrix [Ch-
2 (Row operations) and Ch-3 (Adjoint method)], matrix product. (I believe, you have hands
on experience on these skills already)
Let 𝐴 be a square matrix, if 𝜆 ∈ 𝑅 𝑎𝑛𝑑 𝑿 ≠ 𝟎 such that 𝐴𝑿 = 𝜆𝑿, then 𝜆 is called eigenvalue
(proper, characteristic, latent) and 𝑿 is called eigenvector of matrix 𝐴.

Question:- How to find eigenvalue 𝜆 and corresponding eigenvector 𝑿?

Consider 𝐴𝑿 = 𝜆𝑿 − − − (1)
We can write it as 𝐴𝑿 − 𝜆𝑿 = 𝟎
Or 𝐴𝑿 − 𝜆𝐼𝑿 = 𝟎

Or (𝐴 − 𝜆𝐼)𝑿 = 𝟎 − − − (2)(𝑯𝒐𝒎𝒐𝒈𝒆𝒏𝒆𝒐𝒖𝒔 𝒔𝒚𝒕𝒆𝒎)

(Recall: Homogeneous system 𝐴𝑋 = 0 have trivial solution or non-trivial (infinite many

solution).
We know homogenous system has trivial solution if |𝐴| = 𝑑𝑒𝑡(𝐴) ≠ 0 We are not interested in
trivial solution (useless). Moreover, homogenous system has non-trivial solution if |𝐴| =
𝑑𝑒𝑡(𝐴) = 0.)

𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝑠𝑦𝑡𝑒𝑚 (2) have non-trivial solution only if |𝐴 − 𝜆𝐼| = 𝑑𝑒𝑡(𝐴 − 𝜆𝐼) = 0 − (3)
Equation (3) is called “characteristic equation” and gives “characteristic polynomial”.
Roots of this “characteristic polynomial” are called eigenvalues, and corresponding eigenvectors
can be found by putting these eigenvalues in (2).
3 −5
A Tiny Example: Find eigenvalues and eigenvectors of matrix 𝐴 = [ ]
1 −3
3−𝜆 −5
Solution:- Consider |𝐴 − 𝜆𝐼| = 0 → | |=0
1 −3 − 𝜆
[ (3 − 𝜆)(−3 − 𝜆) + 5] = 0 → 𝜆2 − 4 = 0 → 𝜆 = −2, 2
𝑥1
Case1: when 𝜆 = −2, Solve 𝐴𝑋 = 𝜆𝑋 𝑖𝑒 (𝐴 − 𝜆𝐼)𝑋 = 𝑂 where 𝑋 = [𝑥 ]
2

3 −5 1 0 5 −5
𝐴 − 𝜆𝐼 = 𝐴 + 2𝐼 = [ ] + 2[ ]=[ ]
1 −3 0 1 1 −1
 | |
[𝐴 − 𝜆𝐼|𝑂] = [5 −5 0] 𝑅𝑅𝐸𝐹 [1 −1 0]
1 −1 | 0 0 0 | 0
𝑟 1
𝑥1 − 𝑥2 = 0 ; 𝑥1 = 𝑥2 = 𝑟 𝑎𝑛𝑑 𝑋 = [ ] = 𝑟 [ ]
𝑟 1
3 −5 1 −2 1
Verify: 𝐴𝑋 = [ ] [ ] = [ ] = −2 [ ] = 𝜆𝑋
1 −3 1 −2 1
Case2: when 𝜆 = 2, Solve 𝐴𝑋 = 𝜆𝑋 𝑖𝑒 (𝐴 − 𝜆𝐼)𝑋 = 𝑂
3 −5 1 0 1 −5
𝐴 − 𝜆𝐼 = 𝐴 − 2𝐼 = [ ] − 2[ ]=[ ]
1 −3 0 1 1 −5
| 0 1 −5 | 0
[𝐴 − 𝜆𝐼|𝑂] = [1 −5 ] 𝑅𝑅𝐸𝐹 [ ]
1 −5 | 0 0 0 | 0
5𝑠 5
𝑥1 − 5𝑥2 = 0 ; 𝑥1 = 5𝑥2 ; 𝑥2 = 𝑠 𝑡ℎ𝑒𝑛 𝑥1 = 5𝑠 𝑎𝑛𝑑 𝑋 = [ ] = 𝑠 [ ]
𝑠 1
3 −5 5 10 5
Verify: 𝐴𝑋 = [ ] [ ] = [ ] = 2 [ ] = 𝜆𝑋
1 −3 1 2 1
2 0 −1 −2
𝐴𝑋 = [ ][ ] = [ ]
0 0.5 −1 −0.5
 𝟏 𝟏 𝟏
Example 1: (Repeated roots) 𝑨 = [𝟐 𝟏 −𝟏 ]
𝟎 −𝟏 𝟏
−𝜆3 + 3𝜆2 + 0𝜆 − 4 = 0 (𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝑃𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙)
𝜆 = −1, 2, 2
𝑥1
𝑥
Case 1: For 𝜆 = −1; (𝐴 − 𝜆𝐼 )𝑋 = 𝟎 where 𝑋 = [ 2 ]
𝑥3
2 1 1 | 0 −3/2𝑟 −3/2
[2 2 −1 | 0] (Do yourself, here I am using LAT) 𝑋 = [ 2𝑟 ] = 𝑟 [ 2 ].
0 −1 2 | 0 𝑟 1
𝑥1
Case 1: For 𝜆 = 2 (𝐴 − 𝜆𝐼)𝑋 = 0 where 𝑋 = [𝑥2 ]
𝑥3
 −1 1 1 | 0 0 0
[2 −1 −1 | 0] (Do yourself, here I am using LAT) 𝑋 = [−𝑠] = 𝑠 [−1].
0 −1 −1 | 0 𝑠 1
Example 2: (Repeated roots)
0 0 −2
𝐴 = [1 2 1 ]
1 0 3
−𝜆 0 −2
Solution: consider |𝐴 − 𝜆𝐼| = 0 → | 1 2 − 𝜆 1 |=0
1 0 3−𝜆
2−𝜆 1 1 2−𝜆
𝐸𝑥𝑝𝑎𝑛𝑑 𝑤𝑖𝑡ℎ 𝑅𝑜𝑤 1 → −𝜆 | |− 2| |=0
0 3−𝜆 1 0
−𝜆[(2 − 𝜆)(3 − 𝜆) − 0] − 2[0 − (2 − 𝜆)] = 0 → (2 − 𝜆)[−𝜆(3 − 𝜆) + 2] = 0
Either 2 − 𝜆 = 0 or [−𝜆(3 − 𝜆) + 2] = 0.
Gives 𝜆 = 2 and 𝜆2 − 3𝜆 + 2 = 0 → 𝜆 = 1, 2
Case 1: For 𝜆 = 1, Solve homogeneous system 𝐴𝑋 = 𝜆𝑋 Or (𝐴 − 𝜆𝐼)𝑋 = 𝑂
−1 0 −2 𝑥1 0
𝑥
[ 1 1 1 ] [ 2 ] = [0]
1 0 2 𝑥3 0
−1 0 −2 | 0 1 0 2 | 0
[𝐴|𝑂] = [ 1 1 1 | 0] 𝑅𝑅𝐸𝐹 [0 1 −1 | 0]
1 0 2 | 0 0 0 0 | 0
𝑥3 = 𝑠 ; 𝑥2 = 𝑥3 = 𝑠
𝑥1 = −2𝑥3 = −2𝑠
𝑥1 −2𝑠 −2
𝑋 = [𝑥2 ] = [ 𝑠 ] = 𝑠 [ 1 ]
𝑥3 𝑠 1
Case 2: For 𝜆 = 2, Solve homogeneous system 𝐴𝑋 = 𝜆𝑋 Or (𝐴 − 𝜆𝐼)𝑋 = 𝑂
(𝐴 − 2𝐼)𝑋 = 𝑂
−2 0 −2 𝑥1 0
𝑥
[ 1 0 1 ] [ 2 ] = [0]
1 0 1 𝑥3 0
−2 0 −2 | 0 1 0 1| 0
[𝐴|𝑂] = [ 1 0 1 | 0] 𝑅𝑅𝐸𝐹 [0 0 0 | 0]
1 0 1 | 0 0 0 0| 0
 𝑥3 = 𝑟 ; 𝑥2 = 𝑡 ; 𝒙𝟏 + 𝒙𝟑 = 𝟎 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 𝑥1 = −𝑥3 = −𝑟
𝑥1 −𝑟 −1 0
𝑋 = [𝑥2 ] = [ 𝑡 ] = 𝑟 [ 0 ] + 𝑡 [1]
𝑥3 𝑟 1 0
Observation: If for 3X3 matrix you have repeated eigenvalues then either you have less than 3
eigenvectors or 3 eigenvectors.

1 2 3
Solution:- Hint (a) Let 𝐴 = [2 −1 5]; consider |𝐴 − 𝜆𝐼| = 0 (𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛)
3 2 1
1−𝜆 2 3
| 2 −1 − 𝜆 5 |=0
3 2 1−𝜆
𝜆3 − 𝜆2 − 24𝜆 − 36 = 0 (𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝑃𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙)
No need to find eigenvalues and eigenvectors. Only show, 𝐴3 − 𝐴2 − 24𝐴 − 36𝐼 = 𝑂
1 2 3 1 2 3 1 2 3
𝐴3 = [2 −1 5] [2 −1 5] [2 −1 5]
3 2 1 3 2 1 3 2 1
1 2 3 1 2 3
𝐴2 = [2 −1 5 ] [2 −1 5]
3 2 1 3 2 1
1 2 3
24𝐴 = 24 [2 −1 5]
3 2 1
36 0 0
36𝐼 = [ 0 36 0]
0 0 36
Similar Matrices:

Note: Similar matrices have same determinant; have same eigenvalues, have same properties.
Diagonalizable Matrix:
A square matrix 𝐴 is said to be “diagonalizable” if 𝐴 is similar to a “diagonal matrix” that is

𝑫 = 𝑷−𝟏 𝑨𝑷
Note: Columns of 𝑷 are eigenvectors of 𝑨 and diagonal entries of 𝑫 are eigenvalues of 𝑨.
Example:- Diagonalize the following matrix, if possible.
0 0 −2
𝐴 = [1 2 1 ]
1 0 3
−𝜆 0 −2
Solution: consider |𝐴 − 𝜆𝐼| = 0 → | 1 2 − 𝜆 1 |=0
1 0 3−𝜆
2−𝜆 1 1 2−𝜆
𝐸𝑥𝑝𝑎𝑛𝑑 𝑤𝑖𝑡ℎ 𝑅𝑜𝑤 1 → −𝜆 | |− 2| |=0
0 3−𝜆 1 0
−𝜆[(2 − 𝜆)(3 − 𝜆) − 0] − 2[0 − (2 − 𝜆)] = 0 → (2 − 𝜆)[−𝜆(3 − 𝜆) + 2] = 0
Either 2 − 𝜆 = 0 or [−𝜆(3 − 𝜆) + 2] = 0.
Gives 𝜆 = 2 and 𝜆2 − 3𝜆 + 2 = 0 → 𝜆 = 1, 2
Case 1: For 𝜆 = 1, Solve homogeneous system 𝐴𝑋 = 𝜆𝑋 Or (𝐴 − 𝜆𝐼)𝑋 = 𝑂
−1 0 −2 𝑥1 0
𝑥
[ 1 1 1 ] [ 2 ] = [0]
1 0 2 𝑥3 0
−1 0 −2 | 0 1 0 2 | 0
[𝐴|𝑂] = [ 1 1 1 | 0] 𝑅𝑅𝐸𝐹 [0 1 −1 | 0]
1 0 2 | 0 0 0 0 | 0
𝑥3 = 𝑠 ; 𝑥2 = 𝑥3 = 𝑠
𝑥1 = −2𝑥3 = −2𝑠
𝑥1 −2𝑠 −2
𝑥
𝑋 = [ 2] = [ 𝑠 ] = 𝑠 [ 1 ]
𝑥3 𝑠 1
Case 2: For 𝜆 = 2, Solve homogeneous system 𝐴𝑋 = 𝜆𝑋 Or (𝐴 − 𝜆𝐼)𝑋 = 𝑂
(𝐴 − 2𝐼)𝑋 = 𝑂
−2 0 −2 𝑥1 0
[ 1 0 1 ] [𝑥2 ] = [0]
1 0 1 𝑥3 0
−2 0 −2 | 0 1 0 1| 0
[𝐴|𝑂] = [ 1 0 1 | 0] 𝑅𝑅𝐸𝐹 [0 0 0 | 0]
1 0 1 | 0 0 0 0| 0
𝑥3 = 𝑟 ; 𝑥2 = 𝑡 ; 𝒙𝟏 + 𝒙𝟑 = 𝟎 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 𝑥1 = −𝑥3 = −𝑟
𝑥1 −𝑟 −1 0
𝑥 𝑡
𝑋 = [ 2 ] = [ ] = 𝑟 [ 0 ] + 𝑡 [1 ]
𝑥3 𝑟 1 0
------------------------------------------------------------------------------------------------------------------
−2 −1 0
𝑃=[ 1 0 1]
1 1 0
Verify (Do yourself) using row operations or Adjoint method
−1 0 −1
−1
𝑃 =[ 1 0 2]
1 1 1
−1 0 −1 0 0 −2 −2 −1 0
𝑃−1 𝐴𝑃 = [ 1 0 2 ] [1 2 1][1 0 1]
1 1 1 1 0 3 1 1 0
1 0 0
⟹ = [0 2 0]
0 0 2
Example:-

Solution:-

Question No 8 (Exercise 7.2) ( Question 9 Home work )

Target 𝐴 =?
−1 1
Given 𝜆 = 2, 𝑋 = [ ] and when 𝜆 = −3, 𝑋 = [ ]
2 1
Solution: We know 𝑃−1 𝐴𝑃 = 𝐷
𝑃𝑃 −1 𝐴𝑃 = 𝑃𝐷
𝐴𝑃𝑃−1 = 𝑃𝐷𝑃 −1
This implies 𝐴 = 𝑃𝐷𝑃−1
−1 1 2 0
Where 𝑃 = [ ] And 𝐷 = [ ]
2 1 0 −3
−1 1 −1
𝑃 −1 = [ ]
3 −2 −1
−1 −1 1 2 0 1 −1
𝐴 = 𝑃𝐷𝑃−1 = [ ] [ ] [ ]
3 2 1 0 −3 −2 −1
−1 −1 1 2 −2
= [ ] [ ]
3 2 1 6 3
−1 4 5 −4/3 −5/3
= [ ]=[ ]
3 10 −1 −10/3 1/3
Hint for Question 9:-
1 0 1
Given 𝑃 = [0 1 1] Find 𝑃−1 using Row operations
1 1 1
−2 0 0
𝐷=[ 0 −2 0]
0 0 3
Compute 𝐴 = 𝑃𝐷𝑃−1 and you are done
--------------------------------------------------------------------------------------------------------------
Idea for taking higher powers of matrix A
(Application of diagonalization: We know 𝑃−1 𝐴𝑃 = 𝐷
𝑃𝑃−1 𝐴𝑃 = 𝑃𝐷 Multiplying both sides on left with 𝑃.
𝐴𝑃𝑃−1 = 𝑃𝐷𝑃−1 Multiplying both sides on right with 𝑃−1 .
𝐴 = 𝑃𝐷𝑃−1
𝐴9 = (𝑃𝐷𝑃−1 )9 Taking power 9 of both sides.
𝐴9 = 𝑃𝐷𝑃 −1 𝑃𝐷𝑃−1 𝑃𝐷𝑃−1 𝑃𝐷𝑃−1 𝑃𝐷𝑃 −1 𝑃𝐷𝑃−1 𝑃𝐷𝑃−1 𝑃𝐷𝑃 −1 𝑃𝐷𝑃−1
𝐴9 = 𝑃𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝐼𝐷𝑃−1
𝐴9 = 𝑃𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝑃−1 = 𝑃𝐷9 𝑃−1
Similarly, in general 𝐴𝑛 = 𝑃𝐷𝑛 𝑃−1 )
3 −5
Question19 (Ex 7.2) 𝐴=[ ] 𝑡ℎ𝑒𝑛 𝐴9 =?
1 −3
We use formula 𝐴9 = 𝑃𝐷9 𝑃−1

|𝐴 − 𝜆𝐼| = 0 → |3 − 𝜆 −5
|=0
1 −3 − 𝜆
[ (3 − 𝜆)(−3 − 𝜆) + 5] = 0 → 𝜆2 − 4 = 0 𝑎𝑛𝑑 𝜆 = −2, 2
Case1: when 𝜆 = −2, Solve 𝐴𝑋 = 𝜆𝑋 𝑖𝑒 (𝐴 − 𝜆𝐼)𝑋 = 𝑂
| |
[𝐴|𝑂] = [5 −5 0] 𝑅𝑅𝐸𝐹 [1 −1 0]
1 −1 | 0 0 0 | 0
𝑟 1
𝑥1 − 𝑥2 = 0 ; 𝑥1 = 𝑥2 = 𝑟 𝑎𝑛𝑑 𝑋 = [ ] = 𝑟 [ ]
𝑟 1
Case2: when 𝜆 = 2, Solve 𝐴𝑋 = 𝜆𝑋 𝑖𝑒 (𝐴 − 𝜆𝐼)𝑋 = 𝑂
 | |
[𝐴|𝑂] = [1 −5 0] 𝑅𝑅𝐸𝐹 [1 −5 0]
1 −5 | 0 0 0 | 0
5𝑠 5
𝑥1 − 5𝑥2 = 0 ; 𝑥1 = 5𝑥2 ; 𝑥2 = 𝑠 𝑡ℎ𝑒𝑛 𝑥1 = 5𝑠 𝑎𝑛𝑑 𝑋 = [ ] = 𝑠 [ ]
𝑠 1
−2 0 1 5
We know 𝐷 = [ ] and 𝑃 = [ ]
0 2 1 1
1 1 −5
𝑃−1 = [ ]
−4 −1 1
−1 1 5 (−2)9 0 1 −5 −1 1 5 −512 0 1 −5
𝐴9 = 𝑃𝐷9 𝑃 −1 = [ ][ 9] [ ]= [ ][ ][ ]
4 1 1 0 2 −1 1 4 1 1 0 512 −1 1
−512 1 5 −1 0 1 −5 1 5 −1 5 −𝟔 𝟏𝟎
𝐴9 = [ ][ ][ ] = −128 [ ][ ] = −𝟏𝟐𝟖 [ ]
4 1 1 0 1 −1 1 1 1 −1 1 −𝟐 𝟔
−𝟕𝟔𝟖 𝟏𝟐𝟖𝟎
=[ ]
−𝟐𝟓𝟔 𝟕𝟔𝟖
--------------------------------------------------------------------------------------------------------------------
(1) Eigenvalues of a matrix A (3x3) are 1, 2, -3 then Is A diagonalizable? (T/F)
(2) Eigenvalues of a matrix A are 2, 2, -3 then Is A diagonalizable? (T/F)
(3) Eigenvalues of a matrix A are 1, 2, -3 then eigenvalues of 𝐴2 are 1, 4 and 9. (T/F)
(4) If A is 3x3 matrix then it has three distinct eigenvalues. (T/F)
(5) If A is 3x3 real symmetric matrix then it has three distinct eigenvalues. (T/F)