Scribd
Upload
37 views14 pages

Lec 10

The document discusses similar matrices and diagonalization of matrices. It defines similar matrices as matrices A and B where B = P^-1AP for some nonsingular matrix P. A matrix is diagonalizable if it is similar to a diagonal matrix. The key points are: 1. Similar matrices have the same eigenvalues. 2. A matrix is diagonalizable if all its eigenvalues are distinct real numbers. 3. Any real symmetric matrix is diagonalizable as its eigenvalues are real and eigenvectors are orthogonal. 4. To diagonalize a matrix, the steps are to find eigenvalues and corresponding eigenvectors and use them to construct the matrix P such that P^-1AP is a diagonal matrix of eigenvalues.

Uploaded by

Mohamed Mostafa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
37 views14 pages

Lec 10

The document discusses similar matrices and diagonalization of matrices. It defines similar matrices as matrices A and B where B = P^-1AP for some nonsingular matrix P. A matrix is diagonalizable if it is similar to a diagonal matrix. The key points are: 1. Similar matrices have the same eigenvalues. 2. A matrix is diagonalizable if all its eigenvalues are distinct real numbers. 3. Any real symmetric matrix is diagonalizable as its eigenvalues are real and eigenvectors are orthogonal. 4. To diagonalize a matrix, the steps are to find eigenvalues and corresponding eigenvectors and use them to construct the matrix P such that P^-1AP is a diagonal matrix of eigenvalues.

Uploaded by

Mohamed Mostafa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

Computational

Linear Algebra
Similar Matrices
Diagonalization of a Matrix

Lecture # 10
Dr. Ahmad Moursy
Eigenvalues and Eigenvectors
Main Objectives
• Define Similar matrices.
• Introduce diagonalization of matrices.

Recommended Reading
• D. Lay, S. Lay, J. McDonald, Linear Algebra and Its Applications.
• To increase your knowledge, you can consult “Elementary Linear
algebra” by H. Anton & C. Rorres: Chapter 5
Dr. Ahmad Moursy 2
Similar Matrices and Diagonalization
Definition 1: A matrix B is said to be similar to another matrix A if there is a nonsingular
matrix P such that B = P–1A P.
Definition 2: We shall say that the matrix A is diagonalizable if it is similar to a diagonal
matrix. In this case we also say that A can be diagonalized.

Theorem: (Similar matrices have the same eigenvalues)


If A and B are similar 𝑛 × 𝑛 matrices, then they have the same characteristic polynomial.
Thus, A and B have the same eigenvalues.

 1 1
Example: Let A=  and P = 1 1
 then  2 − 1
P −1 =  
and
 − 2 4 1 2 − 1 1 

2 −1 1 1 1 1 2 0
𝐵 = 𝑃−1 𝐴𝑃 = =
−1 1 −2 4 1 2 0 3
It follows that the matrix A is diagonlalizable.
Dr. Ahmad Moursy 3
Theorem: A matrix A is diagonalizable if all its eigenvalues are distinct real numbers.
Moreover, 𝐵 = 𝑃−1 𝐴𝑃 = diag 𝜆1 , 𝜆2 , ⋯ , 𝜆𝑛 Sufficient Condition
for Diagonalization

Theorem: 1- The eigenvalues of any real symmetric matrix are all real numbers.
2- The eigenvectors of any real symmetric matrix are orthogonal.
Then they are
Consequences: Any real symmetric matrix is diagonalizable. linearly independent

How we diagonalize a matrix:


In order to perform this diagonalization we proceed as follows:
(1) Find the real distinct eigenvalues of A.
(2) Find the corresponding (linearly independent) eigenvectors 𝑋1 𝑋2 ⋯ 𝑋𝑛 .
(3) Construct the non singular matrix P as the matrix whose columns are those eigenvectors:
𝑃 = [𝑋1 𝑋2 ⋯ 𝑋𝑛 ]
(4) 𝑃−1 𝐴 𝑃 = 𝐴𝑑 = 𝑑𝑖𝑎𝑔(𝜆1 𝜆2 ⋯ 𝜆𝑛 ) Since 𝐴 𝑋𝑖 = 𝜆𝑖 𝑋𝑖
Dr. Ahmad Moursy 4
Example 1: (Determining whether a matrix is diagonalizable)
1 −2 1
𝐴= 0 0 1
0 0 −3

Solution
Since A is a triangular matrix, its eigenvalues are the main diagonal entries.

𝜆1 = 1, 𝜆2 = 0, 𝜆3 = −3
which are distinct, so A is diagonalizable.

Remark:
The order of the eigenvalues used to form P will determine the order
in which the eigenvalues appear on the main diagonal of D.

Dr. Ahmad Moursy 5


Example 2 : (Diagonalizing a matrix)
 1 − 1 − 1
A= 1 3 1
 
− 3 1 − 1
Find a matrix P such that P −1 AP is diagonal.

Solution: Characteristic equation:


 −1 1 1
I − A = − 1  − 3 − 1 = ( − 2)( + 2)( − 3) = 0
3 −1  +1
Distinct
Eigenvalue s : 1 = 2, 2 = −2, 3 = 3 Eigenvalues

Dr. Ahmad Moursy 6


1 1 1 1 0 1
𝜆1 = 2 ⇒ 𝜆1 I − 𝐴 = −1 −1 −1 ~ 0 1 0
3 −1 3 0 0 0
𝑥1 −𝑡 −1 −1
⇒ 𝑥2 = 0 = 𝑡 0 ⇒ Eigenvector: 𝑝1 = 0
𝑥3 𝑡 1 1
1
−3 1 1 1 0 −
4
𝜆2 = −2 ⇒ 𝜆2 I − 𝐴 = −1 −5 −1 ~ 1
3 −1 −1 0 1
4
0 0 0
1
𝑥1 𝑡 1 1
4 1
⇒ 𝑥2 = 1 = 𝑡 −1 ⇒ Eigenvector: 𝑝2 = −1
𝑥3 − 𝑡 4
4 4 4
𝑡
Dr. Ahmad Moursy 7
 2 1 1  1 0 1 
𝜆3 = 3  3I − A = − 1 0 − 1 ~ 0 1 − 1
   
 3 − 1 4  0 0 0 
 x1  − t  − 1 − 1
  x2  =  t  = t  1   Eigenvecto r : p3 =  1 
 x3   t   1   1 
− 1 1 − 1
Let P = [ p1 p2 p3 ] =  0 − 1 1 
 1 4 1 
 2 0 0
 P −1 AP = 0 − 2 0
0 0 3

Dr. Ahmad Moursy 8


Example 3: Find a nonsingular matrix P that diagonalizes the symmetric matrix A and
then diagonalize A.  0 0 − 2
A= 0 −2 0 
 
− 2 0 3 

Solution: The eigenvalues of A are 1 = −2, 2 = 4, and 3 = −1 , which are distinct


0  − 1 2
X 1 = 1 X2 =  0  X 3 = 0 
The corresponding linear independent eigenvectors are      
0  2  1 

0 − 1 2  Eigenvectors
Hence, the required matrix P is P = 1 0 0 are orthogonal,
  then they are
0 2 1 linearly
independent

The corresponding diagonal matrix is 𝐷 = 𝑑𝑖𝑎𝑔(−2 4 − 1)

Dr. Ahmad Moursy 9


Remarks:
(1) Not Every square matrix is diagonalizable.
(2) If the eigenvalues of A are not all distinct, we should find all eigenvectors
- If all eigenvectors are linearly Independent then 𝑃, 𝑃−1 exist and A is diagonalizable.
- If the eigenvectors are linearly dependent then 𝑃, 𝑃−1 DNE and A is not diagonalizable.
(3) If 𝐷 = 𝑑𝑖𝑎𝑔(𝜆1 𝜆2 ⋯ 𝜆𝑛 ), then 𝐷 𝑘 = 𝑑𝑖𝑎𝑔(𝜆1𝑘 𝜆2𝑘 ⋯ 𝜆𝑛𝑘 ).
(4) 𝐴𝑘 = 𝑃 𝐷𝑘 𝑃−1
If 𝐷 = 𝑃−1 𝐴 𝑃 ⇒ 𝐴 = 𝑃𝐷 𝑃−1 ⇒ 𝐴𝑘 = 𝑃 𝐷 𝑃−1 𝑘
⇒ 𝐴𝑘 = 𝑃 𝐷 𝑃−1 𝑃𝐷 𝑃−1 ⋯ 𝑃 𝐷 𝑃−1 𝑃 𝐷 𝑃−1
⇒ 𝐴𝑘 = 𝑃 𝐷 𝑃−1 𝑃 𝐷 𝑃−1 𝑃 𝐷 𝑃−1 ⋯ 𝑃 𝐷 (𝑃−1 𝑃)𝐴𝑑 𝑃−1

⇒ 𝐴𝑘 = 𝑃 𝐴𝑑 𝑘 𝑃−1

Dr. Ahmad Moursy 10


Example 5:
1 0 0
Evaluate 𝑨𝟓 for the matrix, 𝑨 = −1 1 1.
−1 −2 4

Solution
Eigenvalues of A are 1, 2, 3 (why?). Calculating the eigenvectors for the
distinct eigenvalues, we obtain

1 0 0
𝑃 = [𝑋1 𝑋2 𝑋3 ] = 1 1 1
1 1 2
and
1 0 0
D = 𝑃−1 𝐴𝑃 = 0 2 0
0 0 3

Dr. Ahmad Moursy 11


1 0 0 1 0 0
𝐷5 = 0 32 0 and 𝑃−1 = −1 2 −1 (why?)
0 0 729 0 −1 1

Now, we have 𝐴5 = 𝑃𝐷5 𝑃−1

1 0 0 1 0 0 1 0 0
𝐴5 = 1 1 1 0 32 0 −1 2 −1
1 1 2 0 0 729 0 −1 1

1 0 0
= −31 665 697
−31 −1394 1426

Dr. Ahmad Moursy 12


Example: The matrix 𝐴 = 1 −1 is not diagonalizable since:
0 1

𝑑𝑒𝑡 𝜆 𝐼 − 𝐴 = 1 − 𝜆 2 ⇒ 𝜆1 = 𝜆2 = 1 (not distinct), so we need to check eigenvectors


0 −1 𝑥 0
𝜆𝐼−𝐴 𝑋 = 𝑦 = ⇒ 𝑦 = 0, 𝑥 = 𝑡
0 0 0
1
So, we have only one eigenvector , spans the set of solutions:
0

Example: In example 3 of slide 10 in lecture 9, the eigenvalues are not distinct with 𝜆1
having multiplicity 2, so we need to check eigenvectors. In that case the eigenvectors were
linearly independent. Hence the matrix A is diagonalizable.

Dr. Ahmad Moursy 13


Dr. Ahmad Moursy 14

You might also like