Linear Algebra Part 4

Dr. Edwin Sybingco

Outline
• Eigenvalue and Eigenvector
• Matrix Diagonalization
• Matrix Diagonalization of a Symmetric Matrix
Eigenvalues and Eigenvectors
Ԧ we want to find a vector 𝑥Ԧ such that 𝐴Ԧ𝑥Ԧ = 𝜆𝑥,
• Given a matrix 𝐴, Ԧ
where 𝜆 is real or complex numbers.

• One possible solution is 𝑥Ԧ = 0, but we are interested in finding 𝑥Ԧ ≠ 0.

𝐴Ԧ𝑥Ԧ = 𝜆𝑥Ԧ = 𝜆𝐼Ԧ𝑥Ԧ

𝐴Ԧ𝑥Ԧ − 𝜆𝐼Ԧ𝑥Ԧ = 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ = 0
Trivial solution nontrivial solution

𝑥Ԧ = 0 𝐴Ԧ − 𝜆𝐼Ԧ = 0
Eigenvalues and Eigenvectors
• Characteristics Equation

𝐴Ԧ − 𝜆𝐼Ԧ = 0

• Where the values 𝜆 satisfy the characteristic equation called

eigenvalues.
• If 𝐴Ԧ is an NxN matrix, the characteristic equation is a polynomial
equation of order N
• The matrix 𝐴Ԧ can have up to N distinct eigenvalues
Eigenvalues and Eigenvectors
• To find the non-trivial solution 𝑥Ԧ𝑖 for 𝑖 = 1,2, … , 𝑁, for the specific
eigenvalue 𝜆𝑖 , we use the method in the homogeneous system of
equations

𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ = 0

Ԧ
• 𝑥Ԧ𝑖 are the eigenvectors of the matrix 𝐴.
Theorems
• Given an NxN matrix 𝐴Ԧ with the corresponding eigenvalue 𝜆, then
• 𝜆𝑚 is the eigenvalue of 𝐴Ԧ𝑚 , for 𝑚 = 1,2,3, …
• If 𝐴Ԧ is invertible, then 1Τ𝜆 is the eigenvalue of 𝐴Ԧ−1.
• 𝐴Ԧ is not invertible if an only if one of the eigenvalue of 𝜆 of 𝐴Ԧ is zero, 𝜆 = 0.
• If 𝛼 is any value, then 𝛼 + 𝜆 is the eigenvalue of 𝐴Ԧ + α𝐼. Ԧ
• If 𝐴Ԧ and 𝐵 are similar, then they have the same characteristic polynomial
which implies they also have the same eigenvalues
Matrix Diagonalization
Ԧ find an invertible NxN matrix 𝑃 and an NxN
• Given an NxN matrix 𝐴,
diagonal matrix 𝐷 such that
𝐴Ԧ = 𝑃𝐷𝑃−1 Ԧ =𝐷
𝑃 −1 𝐴𝑃

• Matrix 𝐴Ԧ can be diagonalized if 𝐴Ԧ has N linearly independent vectors

• Let 𝑥Ԧ1 , 𝑥Ԧ2 , … , 𝑥Ԧ𝑁 represents the linearly independent eigenvectors,

then
𝑃 = 𝑥Ԧ1 𝑥Ԧ2 𝑥Ԧ3 ⋯ 𝑥Ԧ𝑁
Matrix Diagonalization
0 0 2
• Example 1: Diagonalize the matrix 𝐴Ԧ = −2 2 2
0 0 2
0 0 2 1 0 0 −𝜆 0 2 −𝜆 0 2
𝐴Ԧ − 𝜆𝐼Ԧ = −2 2 2 −𝜆 0 1 0 = −2 2−𝜆 2 𝐴Ԧ − 𝜆𝐼Ԧ = −2 2−𝜆 2 =0
0 0 2 0 0 1 0 0 2−𝜆 0 0 2−𝜆

−𝜆 0 −𝜆 2 − 𝜆 2 − 𝜆 = 0
2−𝜆 =0
−2 2−𝜆 𝜆 = 0,2,2

𝐹𝑜𝑟 𝜆1 = 0
0 0 2 𝑥1 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ1 = −2 2 2 𝑥2 = 0
0 0 2 𝑥3 0
Matrix Diagonalization
𝑎=1
0 0 2 0 𝑅1 ↔ 𝑅2 −2 2 2 0 𝑥1 = 𝑥2 = 𝑎 𝑎 1
−2 2 2 ቮ0 0 0 2 ቮ0 𝑥2 = 𝑎 𝑥Ԧ1 = 𝑎 𝑥Ԧ1 = 1
0 0 2 0 0 0 2 0 𝑥3 = 0 0 0

𝐹𝑜𝑟 𝜆2 = 2

−2 0 2 𝑥1 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ2 = −2 0 2 𝑥2 = 0
0 0 0 𝑥3 0

𝑎 = 1, 𝑏 = 0 𝑎 = 0, 𝑏 = 1
−2 0 2 0 𝑥1 = 𝑥3 = 𝑎 𝑎 1 0
−2 0 2 ቮ0 𝑥2 = 𝑏 𝑥Ԧ2 = 𝑏 𝑥Ԧ2 = 0 𝑥Ԧ3 = 1
0 0 0 0 𝑥3 = 𝑎 𝑎 1 0
Matrix Diagonalization
Calculate 𝑃−1
0 0 0 1 1 0 1 1 0 1 0 0
𝐷= 0 2 0 𝑃= 1 0 1 1 0 1 ቮ0 1 0 𝑅2 ← 𝑅2 − 𝑅1
0 0 2 0 1 0 0 1 0 0 0 1

1 1 0 1 0 0 1 1 0 1 0 0
0 −1 1 ቮ−1 1 0 0 −1 1 ቮ−1 1 0 𝑅2 ← 𝑅2 − 𝑅3
0 1 0 0 0 1 𝑅3 ← 𝑅3 + 𝑅2 0 0 1 −1 1 1

1 1 0 1 0 0 𝑅1 ← 𝑅1 + 𝑅2 1 0 0 1 0 −1 1 0 0 1 0 −1
0 −1 0 ቮ 0 0 −1 0 −1 0 ቮ 0 0 −1 𝑅2 ← −𝑅2 0 1 0ቮ 0 0 1
0 0 1 −1 1 1 0 0 1 −1 1 1 0 0 1 −1 1 1

1 0 −1
𝑃−1 = 0 0 1
−1 1 1
Matrix Diagonalization
• Example 2: Determine if the given matrix can be diagonalized

1 0 0
𝐴Ԧ = 2 1 0
3 4 2
1 0 0 1 0 0 1−𝜆 0 0 1−𝜆 0 0
𝐴Ԧ − 𝜆𝐼Ԧ = 2 1 0 −𝜆 0 1 0 = 2 1−𝜆 0 𝐴Ԧ − 𝜆𝐼Ԧ = 2 1−𝜆 0 =0
3 4 2 0 0 1 3 4 2−𝜆 3 4 2−𝜆

1−𝜆 0 1−𝜆 1−𝜆 2−𝜆 = 0

1−𝜆 =0
4 2−𝜆 𝜆 = 1,1,2
Matrix Diagonalization
𝐹𝑜𝑟 𝜆1 = 1

1−𝜆 0 0 𝑥1 0 0 0 𝑥1 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ1 = 2 1−𝜆 0 𝑥2 = 2 0 0 𝑥2 = 0 2𝑥1 = 0 𝑥1 = 0 𝑥2 = 𝑎
3 4 2 − 𝜆 𝑥3 3 4 1 𝑥3 0 4𝑥2 + 𝑥3 = 0 𝑥3 = −4𝑎

0 0 With the expression of 𝑥Ԧ1, it is not possible to get two

𝑥Ԧ1 = 𝑎 𝑥Ԧ1 = 1 linearly independent vectors
−4𝑎 −4
Matrix Diagonalization
• A given NxN matrix is diagonalized by selecting one eigenvector from
the corresponding eigenvalue to form N linearly independent
eigenvectors

• Eigenvectors are linearly independent if the eigenvalues of a given

square matrix are distinct

• A given NxN matrix with less than N distinct eigenvalues may or may
not be diagonalized
Matrix Diagonalization of a Symmetric Matrix
• Symmetric matrix is a matrix in which 𝑎𝑖𝑗 = 𝑎𝑗𝑖 for all values of i and j.
Thus, and is always square. 𝐴Ԧ = 𝐴Ԧ𝑇

• Example
1 3 5 1 3 5
𝐴Ԧ = 3 0 6 Ԧ𝑇
𝐴 = 3 0 6
5 6 9 5 6 9
Matrix Diagonalization of a Symmetric Matrix
• Theorem. Given an NxN symmetric matrix whose elements are real
numbers then its eigenvalues are all real values.

• If the symmetric matrix 𝐴Ԧ has N distinct eigenvalues 𝜆1 , 𝜆2 , … , 𝜆𝑁 with

unit norm eigenvectors 𝑥Ԧ1 , 𝑥Ԧ2 , … , 𝑥Ԧ𝑁 then 𝑃 is an orthogonal matrix,
𝑃𝑇 = 𝑃−1

𝐴Ԧ = 𝑃 𝐴𝑃
Ԧ 𝑇
Matrix Diagonalization of a Symmetric Matrix
1 0 1
• Diagonalize the given matrix 𝐴Ԧ = 0 1 1
1 1 0
1 0 1 1 0 0 1−𝜆 0 1 1−𝜆 0 1
𝐴Ԧ − 𝜆𝐼Ԧ = 0 1 1 −𝜆 0 1 0 = 0 1−𝜆 1 𝐴Ԧ − 𝜆𝐼Ԧ = 0 1−𝜆 1 =0
1 1 0 0 0 1 1 1 −𝜆 1 1 −𝜆

1 − 𝜆 −𝜆 1 − 𝜆 − 1 − 1 − 𝜆 = 0
1−𝜆 1 0 1−𝜆 𝜆3 − 2𝜆2 − 𝜆 + 2 = 0
1−𝜆 + =0
1 −𝜆 1 1
𝜆 = −1,1,2
𝐹𝑜𝑟 𝜆1 = −1
2 0 1 𝑥1 2 0 1 𝑥1 0 2 0 1 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ1 = 0 2 1 𝑥2 = 0 2 1 𝑥2 = 0 0 2 1 ቮ0
1 1 1 𝑥3 1 1 1 𝑥3 0 1 1 1 0
Matrix Diagonalization of a Symmetric Matrix
1 1
2 0 1 0 2 0 1 0 2 0 1 0 2𝑥1 + 𝑥3 = 0 𝑥1 = −2𝑥3= −2𝑎
0 2 1 ቮ0 0 2 1 ቮ0 0 2 1 ቮ0 2𝑥2 + 𝑥3 = 0 𝑥2 = −12𝑥3= −12𝑎
1 1 1 0 𝑅3 ← 2𝑅3 − 𝑅1 0 2 1 0 𝑅3 ← 𝑅3 − 𝑅2 0 0 0 0 𝑥3 = 𝑎

−12𝑎 𝐹𝑜𝑟 𝜆2 = 1 0 0 1 𝑥1 0 𝑥3 = 0
1 1 −1
𝑥Ԧ1 = −2𝑎 =
1 −1 𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ2 = 0 0 1 𝑥2 = 0
1 2 1 2 6 2 1 1 −1 𝑥3 0 𝑥1 + 𝑥2 = 0
2𝑎 + 2𝑎 + 𝑎2 𝑎
𝑥2 = 𝑎 𝑥1 = −𝑎

−𝑎 𝐹𝑜𝑟 𝜆3 = 2 −1 0 1 𝑥1 0
1 1 −1
𝑥Ԧ2 = 𝑎 = 1 𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ3 = 0 −1 1 𝑥2 = 0
−𝑎 2 + 𝑎 2 + 02 0 2 0 1 1 −2 𝑥3 0
Matrix Diagonalization of a Symmetric Matrix
𝐹𝑜𝑟 𝜆3 = 2
−1 0 1 𝑥1 0 −1 0 1 0
𝐴Ԧ − 𝜆𝐼Ԧ 𝑥Ԧ3 = 0 −1 1 𝑥2 = 0 0 −1 1 ቮ0
1 1 −2 𝑥3 0 1 1 −2 0 𝑅3 ← 𝑅3 + 𝑅1

−1 0 1 0 −1 0 1 0 𝑥1 = 𝑥3 = 𝑎
0 −1 1 ቮ0 0 −1 1 ቮ0 𝑥2 = 𝑥3 = 𝑎
0 1 −1 0 𝑅3 ← 𝑅3 + 𝑅2 0 0 0 0 𝑥3 = 𝑎

1 1 1
𝑎 − −
1 1 1 6 2 3
𝑥Ԧ3 = 𝑎 = 1 𝑃= −
1 1 1
𝑎2 + 𝑎2 + 𝑎2 𝑎 3 1 6 2 3
2 1
0
6 3
Matrix Diagonalization of a Symmetric Matrix

1 1 1 1 1 2
− − − −
1 0 1 6 2 3 −1 0 0 6 6 6
1 1 1 1 1
𝐴Ԧ = 0 1 1 = −
6 2 3 0 1 0 −
2 2
0
1 1 0 2
0 1 0 0 2 1 2 1
6 3 3 3 3
Assignment 1D-1
• Find the eigenvalues and eigenvectors of the following matrix

5 −10 −5
𝐴Ԧ = 2 14 2
−4 −8 6

• Determine the diagonalization of 𝐴Ԧ

Assignment 1D-2
• Find the eigenvalues and eigenvectors of 𝐴Ԧ
3 0 0 0
0 3 0 0
𝐴Ԧ =
1 1 3 0
1 1 1 3