Scribd
Upload
22 views

Additional Solved Problems

1) The matrix A is not diagonalizable as its geometric and algebraic multiplicities are not equal for one of the eigenvalues. Specifically, the geometric multiplicity of λ1 = -2 is 1 but its algebraic multiplicity is 2. 2) The eigenvalues of the matrix A are λ1 = -2 with algebraic multiplicity 2 and λ2 = 5 with algebraic multiplicity 1. 3) For the matrix in problem 1, the expression for A4 in terms of the eigenvalues is A4 = PD4P-1, where D is the diagonal matrix of eigenvalues.

Uploaded by

chem sou
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
22 views

Additional Solved Problems

1) The matrix A is not diagonalizable as its geometric and algebraic multiplicities are not equal for one of the eigenvalues. Specifically, the geometric multiplicity of λ1 = -2 is 1 but its algebraic multiplicity is 2. 2) The eigenvalues of the matrix A are λ1 = -2 with algebraic multiplicity 2 and λ2 = 5 with algebraic multiplicity 1. 3) For the matrix in problem 1, the expression for A4 in terms of the eigenvalues is A4 = PD4P-1, where D is the diagonal matrix of eigenvalues.

Uploaded by

chem sou
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

Diagonalization (Solved Problems)


1 3 3
Problem 1: Let 𝐴 = (−3 −5 −3)
3 3 1
1- Find the eigenvalues of the matrix A and their respective
eigenspaces.
2- Determine whether A is diagonalizable or not.
3- Find an expression for 𝐴4 in terms of the eigenvalues.

Solution

1- Eigenvalues and Eigenspaces:

Characteristic polynomial:
𝜆 − 1 −3 −3
𝑃(𝜆) = det(𝜆𝐼 − 𝐴) = 𝑑𝑒𝑡 ( 3 𝜆+5 3 )
−3 −3 𝜆 − 1
𝜆+5 3 3 3 3 𝜆+5
= (𝜆 − 1) | | − (−3) | |− 3| |
−3 𝜆 − 1 −3 𝜆 − 1 −3 −3
= (𝜆 − 1)[(𝜆 + 5)(𝜆 − 1) − 3(−3)] + 3[3𝜆 − 3 − 3(−3)]
− 3[−9 + 3𝜆 + 15]
= (𝜆 − 1)2 (𝜆 + 5) + 9(𝜆 − 1) + 3(3𝜆 + 6) − 3(3𝜆 + 6)
= (𝜆 − 1)[(𝜆 − 1)(𝜆 + 5) + 9] = (𝜆 − 1)(𝜆2 + 4𝜆 − 5 + 9)
= (𝜆 − 1)(𝜆 + 2)²

Therefore , the eigenvalues of the matrix 𝐴 are :


𝜆 = −2 𝑤𝑖𝑡ℎ 𝐴𝑚(𝜆1 ) = 2
{ 1
𝜆2 = 1 𝑤𝑖𝑡ℎ 𝐴𝑚(𝜆2 ) = 1
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

−3 −3 −3
 For 𝝀𝟏 = −𝟐 : 𝜆1 𝐼 − 𝐴 = ( 3 3 3)
−3 −3 −3
𝑥1
Let 𝑋 = (𝑥2 ) be an eigenvector associated with 𝜆1 ⇒ (𝜆1𝐼 − 𝐴)𝑋 = 𝟎
𝑥3
−3 −3 −3 𝑥1 0
⇒( 3 3 3 ) (𝑥2 ) =(0)
−3 −3 −3 𝑥3 0
⇒ 3𝑥1 + 3𝑥2 + 3𝑥3 = 0 ⇒ 𝑥1 = −𝑥2 − 𝑥3

Therefore, the eigenvectors associated with 𝜆1 have the general form :


𝑥1 −𝑥2 − 𝑥3 −𝑥2 −𝑥3
𝑋 = (𝑥2 ) = ( 𝑥2 ) = ( 𝑥2 ) + ( 0 )
𝑥3 𝑥3 0 𝑥3
−1 −1
= 𝑥2 ( 1 ) + 𝑥3 ( 0 ) 𝑤𝑖𝑡ℎ (𝑥2 , 𝑥3 ) ≠ (0,0)
0 1
The associated eigenspace is then:
−1 −1 −1 −1
𝐸(𝜆1 ) = {𝑥2 ( 1 ) + 𝑥3 ( 0 ), 𝑥2 , 𝑥3 ∈ 𝑅} = Span{( 1 ) , ( 0 )}
0 1 0 1
0 −3 −3
 For 𝝀𝟐 = 𝟏 ∶ 𝜆2 𝐼 − 𝐴 = (−3 −6 −3)
3 3 0
𝑥1
Let 𝑋 = (𝑥2 ) be an eigenvector associated with 𝜆1 ⇒ (𝜆1𝐼 − 𝐴)𝑋 = 𝟎
𝑥3
0 −3 −3 𝑥1 0
⇒ (−3 −6 −3) ( 2 ) =(0)
𝑥
3 3 0 𝑥3 0
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

−3𝑥2 − 3𝑥3 = 0 ⇒ 𝑥3 = −𝑥2


⇒ {−3𝑥1 − 6𝑥2 − 3𝑥3 = 0 ⇒ −𝑥1 − 2𝑥2 − (−𝑥2 ) = 0 ⇒ 𝑥1 = −𝑥2
3𝑥1 + 3𝑥2 ⇒ 𝑥1 = −𝑥2

Therefore, the eigenvectors associated with 𝜆2 have the general form:


𝑥1 −𝑥2 −1
𝑋 = (𝑥2 ) = ( 𝑥2 ) = 𝑥2 ( 1 ) 𝑤𝑖𝑡ℎ 𝑥2 ≠ 0
𝑥3 −𝑥2 −1
The associated eigenspace is then
−1 −1
𝐸(𝜆1 ) = {𝑥2 ( 1 ) , 𝑥2 ∈ 𝑅} = Span {( 1 )}
−1 −1

2- Is the matrix diagonalizable?

We need to find the geometric multiplicities of the eigenvalues:

 Method 1: Using the rank-nullity theorem

𝐺𝑚(𝜆) = 𝐷𝑖𝑚(𝐸 (𝜆)) = 𝑛𝑢𝑙𝑙 (𝜆𝐼 − 𝐴) = 𝑛 − 𝑟𝑎𝑛𝑘(𝜆𝐼 − 𝐴)

𝑟𝑎𝑛𝑘(𝜆𝐼 − 𝐴): 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑛𝑜𝑛𝑧𝑒𝑟𝑜 𝑟𝑜𝑤𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑜𝑤 𝑒𝑐ℎ𝑒𝑙𝑜𝑛 𝑓𝑜𝑟𝑚


−3 −3 −3
For 𝝀𝟏 = 𝟑 : 𝜆1 𝐼 − 𝐴 = ( 3 3 3)
−3 −3 −3
Using row-operations, we compute the row echelon form for the
matrix (𝜆1 𝐼 − 𝐴)

−3 −3 −3 𝑅𝑅2 →𝑅
→𝑅
1 +𝑅2 −3 −3 −3 𝑅1 → −1𝑅1 1 1 1
3 2 −𝑅3 3
(3 3 3 )→ (0 0 0 )→ (0 0 0)
−3 −3 −3 0 0 0 0 0 0
⇒ 𝑟𝑎𝑛𝑘(𝜆1 𝐼 − 𝐴) = 1 ⇒ 𝐺𝑚(𝜆1 ) = 3 − 1 = 2
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

For 𝝀𝟐 = 𝟎 ∶ 𝐴𝑚(𝜆2 ) = 1, and since 1 ≤ 𝐺𝑚 ≤ 𝐴𝑚, it follows that


𝐺𝑚(𝜆1 ) = 1

Since : 𝐺𝑚(𝜆1 ) = 𝐴𝑚(𝜆1 ) = 2 and 𝐺𝑚(𝜆2 ) = 𝐴𝑚(𝜆2 ) = 1, we conclude


that 𝐴 is diagonalizable.

 Method 2: Using the eigenspaces


−1 −1
For 𝝀𝟏 = 𝟑 ∶ 𝐸 (𝜆1) = Span {( 1 ) , ( 0 )}
0 1
−1 −1
The vectors ( 1 ) , ( 0 ) that span the eigenspace 𝐸 (𝜆1 ) are linearly
0 1
independent ⇒ They form a basis for 𝐸(𝜆1 )

Therefore : 𝐷𝑖𝑚(𝐸(𝜆1 )) = 2 ⇒ 𝐺𝑚(𝜆1 ) = 2

−1 −1
𝐸(𝜆2 ) = Span{( 1 )} ⇒ The vector ( 1 ) forms a basis for 𝐸 (𝜆2 )
−1 −1
Therefore : 𝐷𝑖𝑚(𝐸(𝜆2 )) = 1 ⇒ 𝐺𝑚(𝜆2 ) = 1

For 𝝀𝟐 = 𝟎 ∶ 𝐴𝑚(𝜆2 ) = 1, and since 1 ≤ 𝐺𝑚 ≤ 𝐴𝑚, it follows that


𝐺𝑚(𝜆1 ) = 1

Since : 𝐺𝑚(𝜆1 ) = 𝐴𝑚(𝜆1 ) = 2 and 𝐺𝑚(𝜆2 ) = 𝐴𝑚(𝜆2 ) = 1 we conclude


that the matrix is diagonalizable.
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

 Diagonal form:

Since 𝐴 is diagonalizable, there exists a diagonal matrix 𝐷 such that:

𝐷 = 𝑃 −1 𝐴𝑃

Where 𝑃 is the matrix of basis eigenvectors


−1 −1 −1
𝑃 = [𝑋11 𝑋12 𝑋2 ] = ( 1 0 1)
0 1 1
𝑋11 𝑎𝑛𝑑 𝑋12 are the basis eigenvectors associated with 𝜆1

𝑋2 is the basis eigenvector associated with 𝜆2

𝜆1 0 0 −2 0 0
and 𝐷 = ( 0 𝜆1 0 ) = ( 0 −2 0)
0 0 𝜆2 0 0 1

3- Find an expression for 𝐴4 in terms of the eigenvalues:

𝐷 = 𝑃 −1 𝐴𝑃 ⇒ 𝐴 = 𝑃𝐷𝑃 −1

𝐴4 = (𝑃𝐷𝑃 −1 )(𝑃𝐷𝑃 −1 )(𝑃𝐷𝑃 −1 )(𝑃𝐷𝑃 −1 ) = 𝑃𝐷𝑃 −1 𝑃𝐷𝑃 −1 𝑃𝐷𝑃 −1 𝑃𝐷𝑃 −1

= 𝑃𝐷 4 𝑃 −1 Since 𝑃 −1 𝑃 = 𝐼

(−2)4 0 0 16 0 0
Where 𝐷 = ( 0
4
(−2) 4
0 ) = ( 0 16 0) (This property
0 0 14 0 0 1
holds only for diagonal matrices)
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

6 3 −8
Problem 2: Let 𝐴 = (0 −2 0 ) determine whether 𝐴 is
1 0 −3
diagonalizable or not.

Solution

Characteristic polynomial:
𝜆 − 6 −3 −8
𝑃(𝜆) = det(𝜆𝐼 − 𝐴) = 𝑑𝑒𝑡 ( 0 𝜆+2 0 ) = (𝜆 + 2)²(𝜆 − 5)
−1 0 𝜆+3
Therefore , the eigenvalues of the matrix 𝐴 are :
𝜆 = −2 𝑤𝑖𝑡ℎ 𝐴𝑚(𝜆1 ) = 2
{ 1
𝜆2 = 5 𝑤𝑖𝑡ℎ 𝐴𝑚(𝜆2 ) = 1

To determine whether 𝐴 is diagonalizable or not, we need to compute


the geometric multiplicities of the eigenvalues.
−8 −3 8
For 𝝀𝟏 = −𝟐 : 𝜆1 𝐼 − 𝐴 = ( 0 0 0)
−1 0 −1
𝑥1
Let 𝑋 = (𝑥2 ) be an eigenvector associated with 𝜆1 ⇒ (𝜆1𝐼 − 𝐴)𝑋 = 𝟎
𝑥3
−8 −3 8 𝑥1 0
⇒( 0 0 0) ( 2 ) =(0)
𝑥
−1 0 1 𝑥3 0
−8𝑥1 − 3𝑥2 + 8𝑥3 = 0 −8𝑥3 − 3𝑥2 + 8𝑥3 = 0 ⇒ 𝑥2 = 0
⇒{ ⇒
𝑥1 − 𝑥3 = 0 𝑥1 = 𝑥3
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

Therefore, the eigenvectors associated with 𝜆1 have the general form:


𝑥1 −𝑥3 −1
𝑥
𝑋 = ( 2 ) = ( 0 ) = 𝑥3 ( 0 ) 𝑤𝑖𝑡ℎ 𝑥3 ≠ 0
𝑥3 𝑥3 1
The associated eigenspace is then:
−1 −1
𝐸(𝜆1 ) = {𝑥3 ( 0 ), 𝑥3 ∈ 𝑅} = Span {( 0 )}
1 1
⇒ 𝐷𝑖𝑚(𝐸 (𝜆1 )) = 1 ⇒ 𝐺𝑚(𝜆1 ) = 1

Since: 𝐺𝑚(𝜆1 ) < 𝐴𝑚(𝜆1 ), it follows that 𝐴 is not diagonalizable.


Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

1−𝑎 𝑎
Problem 3: Let 𝐴 = ( ), determine the values of 𝑎 that
−𝑎 1+𝑎
make 𝐴 diagonalizable.

Solution

Characteristic polynomial:
𝜆−1+𝑎 −𝑎
𝑃(𝜆) = det(𝜆𝐼 − 𝐴) = det ( )
𝑎 𝜆−1−𝑎
= (𝜆 − 1 + 𝑎 )(𝜆 − 1 − 𝑎 ) + 𝑎 2 = (𝜆 − 1)2 − 𝑎 2 + 𝑎 2
= (𝜆 − 1)²

Therefore 𝐴 has only one eigenvalue 𝜆1 = 1 with 𝐴𝑚(𝜆1 ) = 2


𝑎 𝑎
𝜆1 𝐼 − 𝐴 = ( )
−𝑎 −𝑎
𝑥1
Let 𝑋 = (𝑥 ) be an eigenvector associated with 𝜆1 ⇒ (𝜆1 𝐼 − 𝐴)𝑋 = 0
2

𝑎 𝑎 𝑥1 0
⇒( ) (𝑥 ) =( )
−𝑎 −𝑎 2 0
We have two cases:

 If 𝒂 ≠ 𝟎

Then 𝑎𝑥1 + 𝑎𝑥2 = 0 ⇒ 𝑥1 = −𝑥2

Therefore, the eigenvectors associated with 𝜆1 have the general form


𝑥1 𝑥1 1
𝑋 = (𝑥 ) = (−𝑥 ) = 𝑥1 ( ) 𝑤𝑖𝑡ℎ 𝑥1 ≠ 0
2 1 −1
1
𝐸(𝜆1 ) = Span{( ) } ⇒ 𝐷𝑖𝑚(𝐸(𝜆1 )) = 1 ⇒ 𝐺𝑚(𝜆) = 1
−1
Since: 𝐺𝑚(𝜆1 ) < 𝐴𝑚(𝜆1 ), It follows that 𝐴 is not diagonalizable when
𝑎≠0
Linear Algebra/Differential Equations - L2 Institute of Electrical and Electronics Engineering IGEE

 If 𝒂 = 𝟎
0 0
Then 𝜆1 𝐼 − 𝐴 = ( ) / We cannot conclude from this result, so we
0 0
must go back to the matrix 𝐴 with 𝑎 = 0 :
1−𝑎 𝑎 1 0
𝐴=( )= ( ) which is already in diagonal form
−𝑎 1+𝑎 0 1
Therefore , 𝐴 is diagonalizable if and only if 𝑎 = 0.

You might also like