You are on page 1of 17

Determinants and Diagonalization

(Continue)

Department of Mathematics, FPT University

Quy Nhon, 2022


Table of Contents

1 Diagonalization And Eigenvalues

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 2 / 15


Diagonalization And Eigenvalues

Definition (Eigenvalues and Eigenvectors of a Matrix)


If A is an n × n matrix, a number λ is called an eigenvalue of A if AX = λX
for some column X ̸= 0 in Rn .
Such a vector X is called an eigenvector corresponding to the eigenvalue λ, or
a λ-eigenvector.

Example (Eigenvalues and Eigenvectors of a Matrix)


   
3 5 5
Let A = and X = . Then, AX = 4X so λ = 4 is an eigenvalue of A
1 −1 1
with corresponding to eigenvector X.

Theorem (Find Eigenvalues and Eigenvectors)


Let A be an n × n matrix.
The characteristic polynomial of A is cA (x) = det(xIn − A).
Eigenvalues of A are the solutions of the equation cA (x) = 0.
To find λ-eigenvectors, solve the system (λIn − A)X = 0.
(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 3 / 15
Diagonalization And Eigenvalues

Example (Find Eigenvalues and Eigenvectors)


 
3 5
Find eigenvalues and eigenvectors of the matrix A = .
1 −1
Solution.
First, find the characteristic polynomial. Since
     
x 0 3 5 x−3 −5
xI 2 − A = − = ,
0 x 1 −1 −1 x+1

we obtain
 
x−3 −5
cA (x) = det = (x − 3)(x + 1) − (−5)(−1) = (x + 2)(x − 4).
−1 x+1

Second, find the eigenvalues. The roots of cA (x) = 0 are −2 and 4, thus, the
eigenvalues of A are λ1 = −2, λ2 = 4.

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 4 / 15


Diagonalization And Eigenvalues

Example (Find Eigenvalues and Eigenvectors (Continue))


Finally, find the eigenvectors corresponding to each eigenvalue. For λ1 = −2,
       
−2 0 3 5 x1 −5(x1 + x2 )
(λ1 I2 − A)Xλ1 = − =
0 −2 1 −1 x2 −(x1 + x2 )

which implies (λ1 I2 − A)Xλ1 = 0 means that x1 + x2 = 0 or x1 = −x2 . So,


     
x1 −x2 −1
Xλ1 = = = x2 where x2 ∈ R \ {0}.
x2 x2 1

Therefore, the eigenvector Xλ1 corresponding to λ1 are


 
−1
Xλ1 =t , t ∈ R \ {0}.
1

Similarly, for λ2 = 4, λ2 -eigenvectors are


 
5
Xλ2 = s , s ∈ R \ {0}.
1

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 5 / 15


Diagonalization And Eigenvalues

Note
If A is a square matrix, then A and AT have the same characteristic polynomial, and
hence they have the same eigenvalues.

Quiz question
Find eigenvalues and eigenvectors of the matrix
 
2 0 0
A = 1 2 −1 .
1 3 −2

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 6 / 15


Diagonalization And Eigenvalues

Definition (Diagonal Matrix)


An n × n matrix D is called diagonal matrix if all its entries off the main diagonal
are zero,
 
λ1 0 . . . 0
 0 λ2 . . . 0 
D=. ..  = diag(λ1 , λ2 , . . . , λn )
 
.. ..
 .. . . . 
0 0 . . . λn
where λ1 , λ2 , . . . , λn are numbers.

Properties (Diagonal Matrix)


Let D = diag(λ1 , λ2 , . . . , λn ) and E = diag(µ1 , µ2 , . . . , µn ). Then
DE = diag(λ1 µ1 , λ2 µ2 , . . . , λn µn ).
D + E = diag(λ1 + µ1 , λ2 + µ2 , . . . , λn + µn ).
Dk = diag(λk1 , λk2 , . . . , λkn ), k ∈ N∗ .

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 7 / 15


Diagonalization And Eigenvalues

Example (Diagonal Matrix)


   
1 0 0 4 0 0
diag(1, 2, 3) = 0 2 0 , diag(4, −5, −6) = 0 −5 0 .
0 0 3 0 0 −6

diag(1, 2, 3) + diag(4, −5, −6) = diag(5, −3, −3).


diag(1, 2, 3) · diag(4, −5, −6) = diag(4, −10, −18).
[diag(1, 2, 3)]k = diag(1, 2k , 3k ).
[diag(4, −5, −6)]k = diag(4k , (−5)k , (−6)k ).

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 8 / 15


Diagonalization And Eigenvalues

Definition (Diagonalizable Matrix)


An n × n matrix A is diagonalizable if P −1 AP is diagonal for some invertible n × n
matrix P . Here, the invertible matrix P is called a diagonalizing matrix for A.

Example (Diagonalizable Matrix)


 
2 −4
Let A = . Then, a diagonalizing matrix for A and its inverse matrix
−1 −1
   
4 1 1 1 −1
P = , P −1 = .
−1 1 5 1 4

Thus  
−1 3 0
P AP = .
0 −2

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 9 / 15


Diagonalization And Eigenvalues

Remark (Diagonalizable Matrix)


Let D denote the diagonal matrix, we have P −1 AP = D implies A = P DP −1 .

Theorem
If A = P DP −1 then Ak = P Dk P −1 for each k = 1, 2, . . .

Example
 
2009 2 −4
Find A where A = .
−1 −1
     
4 1 1 1 −1 3 0
Solution. We have P = , P −1 = and D = . Then,
−1 1 5 1 4 0 −2

A2009 = P D2009 P −1
 
32009 0
=P P −1 .
0 (−2)2009

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 10 / 15


Diagonalization And Eigenvalues

Theorem
An eigenvalue λ of a square matrix A is said to have multiplicity m if it occurs m times
as a root of the characteristic polynomial cA (x).

Theorem
A square matrix A is diagonalizable if and only if every eigenvalue λ of multiplicity m
yields exactly m basic eigenvectors

Algorithm (Diagonalization Algorithm)


Let A be an n × n matrix.
Find all eigenvalues λ1 , λ2 , . . . , λn of A.
For each eigenvalue compute basic eigenvectors.
If there are exactly a total of n basic eigenvectors, then A is diagonalizable.
In this case, D is the diagonal matrix with λ1 , λ2 , . . . , λn in the diagonal, and P is is
the matrix whose columns are the basic eigenvectors.

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 11 / 15


Diagonalization And Eigenvalues

Example (Diagonalize A Matrix)


 
0 1 1
Diagonalize the matrix A = 1 0 1.
1 1 0
Eigenvalues. Since cA (x) = (x − 2)(x + 1)2 then the eigenvalues are 2, −1, −1.
Eigenvectors.
" #
1
+ λ = 2. X = t 1 , t ̸= 0.
1
" # " #
−1 −1
+ λ = −1. X = t 1 + s 0 , t2 + s2 ̸= 0.
0 1
Diagonal and Diagonalizing Matrices. P −1 AP = D where
     
2 0 0 1 −1 −1 1/3 1/3 1/3
D = 0 −1 0 , P = 1 1 0 , P −1 = −1/3 2/3 −1/3 .
0 0 −1 1 0 1 −1/3 −1/3 2/3

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 12 / 15


Diagonalization And Eigenvalues

Quiz question
Diagonalize the matrix
 
2 0 −1
A = −1/3 5/3 −1/3 .
−2/3 −2/3 7/3

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 13 / 15


Diagonalization And Eigenvalues

Theorem
An n × n matrix with n distinct eigenvalues is diagonalizable.

Example
 
1 1
Show that A = is not diagonalizable.
0 1
Solution. Since cA (x) = (x − 1)2 then λ = 1 is the only eigenvalue. For λ = 1, there
 
1
is only one basic eigenvector X = t , thus, A is not diagonalizable.
0

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 14 / 15


THANK YOU!

(VyNHT – FU Quy Nhon) Determinants and Diagonalization Quy Nhon, 2022 15 / 15


Digonalize A Matrix
Diagonalize the matrix
1 1 0
 

A= 1
0 1 .


1 1 0


Solution.

• Eigenvalues. Since cA (x) = det(xI3 − A) = (x − 2)(x + 1)2 then the eigenvalues


of A are the roots of cA (x) = 0. Thus,

λ1 = 2 and λ2 = −1 (multiplicity 2).

• Eigenvectors. Consider two cases as the following.

+ λ1 = 2 We now find Xλ1 ∈ R3 such that (λ1 I3 − A)Xλ1 = 0. We have

2 0 0 0 1 1 x1   2 −1 −1 x1 


        

(λ1 I3 − A)Xλ1 = 0


 x2  = −1
2 0 − 1 0 1 2 −1 x2  = 0.
        
  

0 0 2 1 1 0 −1 −1 2 x3
        
x3

To solve the above system, we transform the coefficient matrix into (re-
duced) row-echelon form as follows

 2  1 −1/2 −1/2 r2 +r1 →r2 1 −1/2 −1/2


     
−1 −1 1
 2 r1 →r1   r3 +r1 →r3 
−1 2 −1 −−−−→ −1 2 −1  −− −−−→ 0 3/2 −3/2
 

−1 −1 2 2 0 −3/2 3/2
     
−1 −1

1 1 0 −1
   
r3 +r2 →r3
2
−1/2 −1/2
3 r2 →r2  r1 + 21 r2 →r1 
−−−−−→ 0 1 −1  −−−−−−→ 0 1 −1
 

0 0 0 0 0 0
   

 
x1 − x3 = 0 x 1 = x3

 

which implies leading-variables are x1 , x2 and . So 
− x3 = 0 x x = x .

 
2 2 3
Let x3 = t. Then, we obtain the λ1 -eigenvectors are

1
     
x1  t
Xλ1 = = = t 1 , t ̸= 0.
     
x  t
 2  
1
     
x3 t
2

+ λ2 = −1 We now find Xλ2 ∈ R3 such that (λ2 I3 − A)Xλ2 = 0. We have


0 0  0 1 1 x1  −1 −1 −1 x1 
        
−1
(λ2 I3 − A)Xλ2 =
 0  x2  = −1 −1 −1 x2  = 0.
−1 0  − 1 0 1
        
     

0 0 −1 1 1 0
        
x3 −1 −1 −1 x3
To solve the above system, we transform the coefficient matrix into (re-
duced) row-echelon matrix as follows

−1 −1 −1 r2 −r1 →r2 1 1 1


   
−1r1 →r1
 r3 −r1 →r3 
 −−−−→ 0 0 0
−1 −1 −1 −
 
 

0 0 0
   
−1 −1 −1
which implies leading-variable is x1 and x1 + x2 + x3 = 0. Thus,
x1 = −x2 − x3 .
Therefore, we get
           
x1  −x2 − x3  −x2  −x3  −1 −1
Xλ 2 = x2  =  x2  =  x  +  0  = x  1  + x  0 .
           
  2    2  3 
0 0 1
           
x3 x3 x3
Let x2 = t, x3 = s. Then, the λ2 -eigenvectors are
   
−1 −1
Xλ 2 = t  1  + s  0  , t2 + s2 ̸= 0.
   

0 1
   

1
     
−1 −1
• There are three basic eigenvectors, they are 1 ,  1 ,  0  which are
     
   
1 0 1
     

the columns of the diagonalizing matrix for A.

• Diagonal and Diagonalizing Matrices. P −1 AP = D where

2 0 0 1 −1 −1  1/3 1/3 1/3 


     

D = 0 −1 0  , P = 1 1 0  , P −1 = −1/3 2/3 −1/3 .


     

0 0 −1 1 0 1 −1/3 −1/3 2/3


     

or
2 0 0 1 −1 −1  1/3 1/3 1/3 
     

D = 0 −1 0  , P = 1 0 1  , P −1 = −1/3 −1/3 2/3  .


     

0 0 −1 1 1 0 −1/3 2/3 −1/3


     

You might also like