Eigenvalues and

Eigenvectors

Chapter 4
Linear Algebra
Dr. Akhlaq Husain Slide 1
 Similarity and
Diagonalization

Chapter 4, Section 4.4

Linear algebra
Dr. Akhlaq Husain Slide 29
 Similarity
Definition: Let 𝐴 and 𝐵 be 𝑛 × 𝑛 matrices. We say that 𝐴 is
similar to 𝐵 if there is an invertible 𝑛 × 𝑛 matrix 𝑃 such
that 𝑃−1 𝐴𝑃 = 𝐵.
If 𝐴 is similar to 𝐵, we write 𝐴 ∼ 𝐵.

Remark: If 𝐴 ∼ 𝐵 then 𝐴 = 𝑃𝐵𝑃−1 ⟹ 𝐴𝑃 = 𝑃𝐵.

1 2 1 0
Example: Let 𝐴 = , 𝐵=
0 −1 −2 −1
Then 𝐴 ∼ 𝐵 since
1 2 1 −1 1 −1 1 0
=
0 −1 1 1 1 1 −2 −1
Linear algebra
Dr. Akhlaq Husain Slide 30
Therefore,
1 −1
𝐴𝑃 = 𝑃𝐵 with 𝑃 =
1 1

Theorem: Let 𝐴 and 𝐵 be similar i.e., 𝐴 ∼ 𝐵. Then

(𝑎) det 𝐴 = det(𝐵)
(𝑏) 𝐴 is invertible if and only if 𝐵 is invertible.
(𝑐) rank 𝐴 = rank(𝐵)
(𝑑) 𝐴 and 𝐵 have same characteristic polynomial.
(𝑑) 𝐴 and 𝐵 have same eigenvalues.
(𝑒) 𝐴𝑚 ∼ 𝐵𝑚 for all integers 𝑚 ≥ 0.

Linear algebra
Dr. Akhlaq Husain Slide 31
Example:
1 2 2 1
𝐴= , 𝐵=
2 1 1 2
Then 𝐴 and 𝐵 are not similar since det 𝐴 = −3 and
det 𝐵 = 3

Example:
1 3 1 1
𝐴= , 𝐵=
2 2 3 −1
Then 𝐴 and 𝐵 are not similar since
𝑝𝜆 𝐴 = 𝜆2 − 3𝜆 − 4
and
𝑝𝜆 𝐵 = 𝜆2 − 4

Linear algebra
Dr. Akhlaq Husain Slide 32
 Diagonalization
Definition: Let 𝐴 be 𝑛 × 𝑛 matrix. is said to be
diagonalizable if 𝐴 is similar to a diagonal matrix i.e.,
there is an invertible 𝑛 × 𝑛 matrix 𝑃 such that 𝑃−1 𝐴𝑃 = 𝐷.

1 3
Example: Let 𝐴 = , then 𝐴 is diagonalizable
2 2
since if
1 3 4 0
𝑃= , 𝐷= .
1 −2 0 −1
Then
𝑃−1 𝐴𝑃 = 𝐷 (check!)

Linear algebra
Dr. Akhlaq Husain Slide 33
Theorem: Let 𝐴 be 𝑛 × 𝑛 matrix. Then 𝐴 is diagonalizable
if and only if 𝐴 has 𝑛 linearly independent eigenvectors.
More precisely,
there exist an invertible matrix 𝑃 and a diagonal matrix 𝐷
such that 𝑃−1 𝐴𝑃 = 𝐷 if and only if the columns of 𝑃 are 𝑛
linearly independent eigenvectors of 𝐴 and the diagonal
entries of 𝐷 are the eigenvalues of 𝐴 corresponding to the
eigenvectors in 𝑃 in the same order.

Linear algebra
Dr. Akhlaq Husain Slide 34
Example: If possible diagonalize the matrix
−1 0 1
𝐴 = 3 0 −3
1 0 −1
Solution: The eigenvalues of 𝐴 are
𝜆1 = 0, 𝜆2 = 0, 𝜆3 = −2 (check!)
Eigenvectors corresponding to 𝜆1 = 0 are
0 1
𝐸0 = 𝛼 1 + 𝛽 0 , 𝛼, 𝛽 ∈ ℝ
0 1
Two independent eigenvectors are
0 1
𝑣1 = 1 , 𝑣2 = 0
ด
0 ด1
𝑣1 𝑣2
Linear algebra
Dr. Akhlaq Husain Slide 35
Eigenvectors corresponding to 𝜆3 = −2 are
−1
𝐸−2 = 𝛼 3 , 𝛼∈ℝ
ถ1
𝑣3
Therefore,
𝑃 = 𝑣1 , 𝑣2 , 𝑣3
0 1 −1
⟹𝑃= 1 0 3
0 1 1
Note that, 𝑃 is invertible and
0 0 0
𝑃−1 𝐴𝑃 = 𝐷 = 0 0 0 .
0 0 −2
Linear algebra
Dr. Akhlaq Husain Slide 36
Theorem: If 𝐴 be 𝑛 × 𝑛 matrix with 𝑛 distinct eigenvalues
then 𝐴 is diagonalizable.

2 −3 7
Example: Let 𝐴 = 0 5 1 , then 𝐴 is
0 0 −1
diagonalizable since the eigenvalues of A are
𝜆1 = 2, 𝜆1 = 5, 𝜆1 = −1. (why? )

Linear algebra
Dr. Akhlaq Husain Slide 37
 Theorem: If 𝐴 is an 𝑛 × 𝑛 matrix, then the geometric
multiplicity of each eigenvalue is less than or equal to its
algebraic multiplicity.

Example: Determine 𝐴10 if

0 1
𝐴= .
2 1
The eigenvalues of 𝐴 are 𝜆1 = −1 and 𝜆2 = 2 with
1
corresponding eigenvectors 𝑣1 = and 𝑣2 =
−1
1
. (Check!)
2
Linear algebra
Dr. Akhlaq Husain Slide 38
𝐴 is diagonalizable sine it has 2 linearly independent
eigenvectors. Therefore,
1 1
𝑃 = 𝑣1 , 𝑣2 =
−1 2
and
−1 0
𝐷=
0 2
such that
𝑃−1 𝐴𝑃 = 𝐷 ⟹ 𝐴 = 𝑃𝐷𝑃−1
From this, we get
𝐴10 = 𝑃𝐷10 𝑃−1
1 1 −1 10 0 1 1 −1
=
−1 2 0 210 −1 2

Linear algebra
Dr. Akhlaq Husain Slide 39
 10 2/3 −1/3
1 1 −1 0
⟹ 𝐴10 =
−1 2 0 210 1/3 1/3
2 −1 10 + 210 −1 10 + 211
= 3 3
2 −1 11 + 211 −1 12 + 211
3 3
342 341
= .
682 683

Linear algebra
Dr. Akhlaq Husain Slide 40
Exercise: Show that the following matrices are not similar.
2 1 4 1 0 0
𝐴= 0 2 3 , 𝐵 = −1 4 0 .
0 0 4 2 3 4
Exercise: Determine if the following matrices are
diagonalizable. If yes, then find an invertible matrix 𝑃 and a
diagonal matrix 𝐷 such that 𝑃−1 𝐴𝑃 = 𝐷.
1 0 0
5 2
𝑖 𝐴= , 𝑖𝑖 𝐴 = 2 2 1 .
2 5
3 0 1
Exercise: Compute the given powers of matrices
2015
10 1 1 1
−1 6
𝑖 , 𝑖𝑖 𝐴 = 0 −1 0 .
1 0
0 0 −1
Linear algebra
Dr. Akhlaq Husain Slide 41