0% found this document useful (0 votes)
2 views7 pages

Lecture Note 2

The document discusses eigenvalues and eigenvectors of triangular matrices, emphasizing that the eigenvalues are the entries on the main diagonal. It explains concepts such as algebraic and geometric multiplicities, invertibility of matrices, and diagonalizability, providing examples and theorems related to these topics. Additionally, it highlights the relationship between eigenvalues of powers of matrices and the conditions for diagonalizability based on the linear independence of eigenvectors.
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
0% found this document useful (0 votes)
2 views7 pages

Lecture Note 2

The document discusses eigenvalues and eigenvectors of triangular matrices, emphasizing that the eigenvalues are the entries on the main diagonal. It explains concepts such as algebraic and geometric multiplicities, invertibility of matrices, and diagonalizability, providing examples and theorems related to these topics. Additionally, it highlights the relationship between eigenvalues of powers of matrices and the conditions for diagonalizability based on the linear independence of eigenvectors.
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

Theorem

If 𝐴 is an 𝑛 × 𝑛 triangular matrix (upper triangular, lower triangular or diagonal), then the


eigenvalues of 𝐴 are the entries on the main diagonal of 𝐴.
Bases for Eigen spaces
Example 5:

A 3 × 3 matrix with one Eigenvalue and two Linearly Independent Eigenvectors.

−1 −3 −9
𝐴=( 0 5 18 )
0 −2 −7
−1 − 𝜆 −3 −9
det(𝐴 − 𝜆𝐼) = | 0 5−𝜆 18 |
0 −2 −7 − 𝜆
= −(𝜆 + 1)3 = 0

⇒ 𝜆 = −1 is an eigenvalue of algebraic multiplicity 3.

To find the eigenspace 𝐸−1 let us solve

0 −3 −9 𝑥1 0
𝑥
(0 6 18 ) ( 2 ) = (0)
0 −2 −6 𝑥3 0 −3
𝑅1 → 𝑅1 + ( ) 𝑅3
0 −3 −9 0 2 0 0 0 0
(0 6 18 0) 𝑅2 → 𝑅2 + 3𝑅3 (0 0 0 0)
0 −2 −6 0 0 −2 −6 0
Therefore −2𝑥2 − 6𝑥3 = 0

⇒ 𝑥2 = −3𝑥3 and 𝑥1 is arbitrary.

Take 𝑥1 = 𝑡 and 𝑥3 = 𝑠 parameters;


𝑥1 𝑡 1 0
We get (𝑥2 ) = (−3𝑠) = 𝑡 (0) + 𝑠 (−3)
𝑥3 𝑠 0 1
1 0
Thus 𝐸−1 = 𝑠𝑝𝑎𝑛 {(0) , (−3)}
0 1
1 0
Since (0) and (−3) are linearly independent (check), these vectors form a basis for the
0 1
eigenspace 𝐸−1 .

Definition: Geometric Multiplicity.

Let 𝜆 be an eigenvalue of the matrix 𝐴. Then the geometric multiplicity of 𝜆 is the dimension of
the eigenspace corresponds to 𝜆.

Geometric multiplicity of λ =
𝑑𝑖𝑚 𝐸𝜆

In Example 5,

𝜆 = −1 has algebraic multiplicity 3.

1 0
𝐸−1 = 𝑠𝑝𝑎𝑛 {(0) , (−3)} ⇒ 𝑑𝑖𝑚 𝐸−1 = 2.
0 1

∴ Geometric multiplicity of 𝜆 = −1 is 2.

Theorem:

Let 𝜆 be an eigenvalue of 𝐴. Then,

Geometric multiplicity of λ ≤ algebraic multiplicity


of λ

Remark:

The geometric multiplicity of an eigenvalue is never zero. Why?


Eigenvalues and Invertibility

Theorem
A square matrix 𝐴 is invertible if and only if 𝜆 = 0 is not an eigenvalue of 𝐴.

Proof:
Assume that 𝐴 is an 𝑛 × 𝑛 matrix, and 𝜆𝑛 + 𝑐1 𝜆𝑛−1 + ⋯ + 𝑐𝑛−1 𝜆 + 𝑐𝑛 = 0 → ①
is the characteristic equation.

That is

det(𝐴 − 𝜆𝐼) = 𝜆𝑛 + 𝑐1 𝜆𝑛−1 + ⋯ + 𝑐𝑛−1 𝜆 + 𝑐𝑛 → ②

If 𝜆 = 0 is an eigenvalue, then 𝜆 = 0 is a solution of ①.

Setting 𝜆 = 0 in ② we get det(𝐴) = 𝑐𝑛 .

Thus

𝐴 is invertible

iff det 𝐴 ≠ 0

iff 𝑐𝑛 ≠ 0

-----------------------------------------------------------------------

Finding eigenvalues of 𝑨𝒌 for a positive integer 𝒌:

If 𝜆 is an eigenvalue of 𝐴 and 𝑣 is a corresponding eigenvector, then

𝐴2 𝑣 = 𝐴(𝐴𝑣) = 𝐴(𝜆𝑣) = 𝜆𝐴𝑣 = 𝜆2 𝑣

Thus 𝜆2 is an eigenvalue of 𝐴2 .

Hence, we can show that


Theorem:

If 𝑘 is a positive integer, 𝜆 is an eigenvalue of a matrix 𝐴, and 𝑣 is a corresponding eigenvector,


then 𝜆𝑘 is an eigenvalue of 𝐴𝑘 and 𝑣 is a corresponding eigenvector.

Example:

Eigenvalues of powers of a matrix.

1 −1 4
In Example 2, we showed that the eigenvalues of 𝐴 = (3 2 −1) are 𝜆 = 1, 𝜆 = −2 and
2 1 −1
𝜆 = 3.

Thus 𝜆1 = 14 = 1, 𝜆2 = (−2)4 = 16 and 𝜆3 = 34 = 81are eigenvalues of 𝐴4 .

Eigenvectors calculated are corresponding eigenvectors for 𝜆1 = 1, 𝜆2 = 16 and 𝜆3 = 81

Definition: Similar Matrices

𝐴 and 𝐵 are square matrices (of the same order).

We say that 𝐵 is similar to 𝐴 if there is an invertible matrix 𝑃 such that 𝐵 = 𝑃−1 𝐴𝑃.

Remarks:

1. The relation similarity is an equivalence relation.


2. 𝐴 and 𝑃 −1 𝐴𝑃 have the same determinant.
3. 𝐴 is invertible iff 𝑃 −1 𝐴𝑃 is invertible.
4. 𝐴 and 𝑃 −1 𝐴𝑃 have the same characteristic polynomial.
5. 𝐴 and 𝑃 −1 𝐴𝑃 have the same eigenvalues.

Try to prove the above remarks.

Definition: Diagonalizable matrix

A square matrix 𝐴 is said to be diagonalizable if it similar to some diagonal matrix.

(That is 𝑃−1 𝐴𝑃 is a diagonal matrix for some invertible matrix 𝑃)


Finding a matrix 𝑃 that Diagonalize a matrix 𝐴

Example 1:

Diagonalizing a 2 × 2 matrix.

4 2
Let 𝐴 = ( )
3 3
In Example 1, we found two distinct eigenvalues 𝜆1 = 1 and 𝜆2 = 6.

−2 1
Thus eigenvectors we found 𝑣1 = ( ) and 𝑣2 = ( ) must be linearly independent. (Why?)
3 1

𝑣1 𝑣2
↓ ↓
−2 1
Then, setting 𝑃 = ( ) we see that
3 1
−1 1 −1 4 2 −2 1
𝑃−1 𝐴𝑃 = ( 5 ) ( )( )( )
−3 −2 3 3 3 1
−1 1 −1 −2 1
= ( 5 )( )( )
−18 −12 3 1
−1 −5 0
= ( 5 )( )
0 −30
1 0
=( ) which is the matrix whose diagonal components are the eigenvalues of 𝐴.
0 6

Example 2:

Diagonalizing a 3 × 3 matrix with Three Distinct Eigenvalues.

HW Exercise: (Hint: Use matrix 𝐴 given in Example 2. Take the eigenvectors 𝑣1 , 𝑣2 & 𝑣3 calculated
and let 𝑃 = [𝑣1 𝑣2 𝑣3 ] ←columns of 𝑃 are eigenvectors. Then you will see that 𝑃−1 𝐴𝑃 =
1 0 0
[0 −2 0] )
0 0 3
Remark: Since there are finite number of ways to choose an eigenvector, there are infinite
number of ways to choose the diagonalizing matrix 𝑃. The only advice is to choose the
eigenvectors and matrix 𝑃 that are, arithmetically, the easiest to work with. Hence insert as many
0’s and 1’s as possible.

Example 3:

Diagonalizing a 3 × 3 matrix with Two Distinct Eigenvalues and Three Linearly Independent
Eigenvectors.

(Source: El. Lin. Alg. By Grossman)

3 2 4
Let 𝐴 = (2 0 2).
4 2 3
One can show that 𝐴 has two distinct eigenvalues and three linearly independent eigenvectors
𝑣1 , 𝑣2 & 𝑣3 .

𝜆1 0 0
Setting 𝑃 = [𝑣1 𝑣2 𝑣3 ] we can get 𝑃−1 𝐴𝑃 = [ 0 𝜆2 0 ] ← diagonal matrix
0 0 𝜆3

Here 𝜆1 , 𝜆2 and 𝜆3 are eigenvalues of 𝐴.

Note: This example illustrates that 𝐴 is a digonalizable even though its eigenvalues are not
distinct.

Example 4:

A 2 × 2 Matrix with Only One Linearly Independent Eigenvector Cannot be Diagonalized.

(Source: El. Linear Algebra by Grossman)

4 1
Solution: Let 𝐴 = ( )
0 4
𝐴 is an upper triangular matrix.

⇒ 𝜆=4 is an eigenvalue with algebraic multiplicity 2.

Suppose that 𝐴 were diagonalizable (we must get a contradiction here;

since 𝐴 is diagonalizable iff eigenvectors of 𝐴 are linearly independent)

4 0
If 𝐴 is diagonalizable, then the diagonal form must be 𝐷 = ( )
0 4
⇒There is an invertible matrix 𝑃 such that
𝑃 −1 𝐴𝑃 = 𝐷

⇒ 𝐴𝑃 = 𝑃𝐷

⇒ 𝐴 = 𝑃𝐷𝑃−1

4 0 −1
= 𝑃( )𝑃
0 4
= 𝑃(4𝐼)𝑃−1

= 4𝑃𝑃−1

= 4𝐼

=𝐷

a contradiction, since 𝐴 ≠ 𝐷.

Hence no such P exists.


0 0 1 𝑥1 0
(Note that solving (𝐴 − 4𝐼)𝑣 = ( ) we get ( )( ) = ( )
0 0 0 𝑥2 0
𝑥2 0
That is ( ) = ( )
0 0
⇒ 𝑥2 =0 and 𝑥1 is a free variable.
𝑥1 𝑡 1
⇒ (𝑥 ) = ( ) = 𝑡 ( )
2 0 0
1 1
⇒ ( ) is an eigenvector and 𝐸4 =span{( )})
0 0

Read Textbook for more examples.

Theorem :

Let 𝐴 be a 𝑛 × 𝑛 matrix. 𝐴 is diagonalizable iff 𝐴 has 𝑛 linearly independent eigenvectors.

Theorem:

If 𝑣1 , 𝑣2 , … , 𝑣𝑘 are eigenvectors of a matrix 𝐴 corresponding to distinct eigenvalues, then


{𝑣1 , … , 𝑣𝑘 } is a linearly independent set.

You might also like