線性代數

Linear Algebra

<Part 4：特徵值與特徵向量>

長榮大學 科工系 106 級

姓名：

學號：

Textbook: 高等工程數學(Advanced Engineering Mathematics)，第 7 版，Peter V. O’Neil

著，黃孟槺譯，華泰文化總經銷，2012 年 5 月初版，ISBN 978-1-285-10502-4。
Chapter 8 Eigenvalues, Diagonalization, and Special Matrices
(特徵值、對角線化及特殊矩陣)

Suppose A is an × matrix of real numbers, and E is an n-vector and can be written

as a column

= ⋮ .

Then AE is an × 1 matrix, which we may also think of as an n-vector. (可以將 A 視為

在 n 度空間中移動向量的一種運算子)

Depending on A, the direction of AE will generally be different from that of E. It may

happen, however, that for some vector E, AE and E are parallel. In this event, there is a
number such that = . Then is called an eigenvalue (特徵值) of A, with E an
associate eigenvector (特徵向量). (對任意方陣 A 而言，當存在一向量 E，使得
= ，則稱 為 A 的一個特徵值，且稱 E 為對應於該特徵值之特徵向量)

8.1 Eigenvalues (特徵值) and Eigenvectors (特徵向量)

Let A be an × matrix of real or complex numbers.

‧ 定義 8.1 Eigenvalues and Eigenvectors

A real or number is an eigenvalue of A if there is a nonzero × 1 matrix (or vector)

E such that
= .
Any nonzero vector E satisfying this relationship is called an eigenvector (特徵向量)
associated with the eigenvalue (特徵值).

Eigenvalues are also known as characteristic values (特徵向量) of a matrix, and eigen-
vectors can be called characteristic vectors (特徵值).

If is a nonzero scalar and = , then

(α ) = ( ) = (λ ) = λ( ).

- 100 -
 This means that nonzero scalar multiples of eigenvectors are again eigenvectors (特徵
向量的非零常數倍，還是特徵向量).

Example 8.1 特徵值可為零，但特徵向量不能為零向量

Example 8.2 特徵值與特徵向量

- 101 -
‧ Meanings of Eigenvalue and Eigenvector (特徵值與特徵向量之意義)

- 102 -
‧ Consider = , then we have
− =
⟹ ( − ) =
This makes E a non-trivial solution of the × system of linear equations
( − ) = .
But this system can have a non-trivial solution if and only if the coefficient matrix has
determinant zero (若齊性線性方程組有非零解，則其係數矩陣的行列式值為零), that
is
| − | = 0.

‧ 定理 8.1 特徵值與特徵向量方程式

Let A be an × matrix of real or complex numbers. Then

1. is an eigenvalue of A if and only if | − | = 0. (特徵值之求解)

2. If is an eigenvalue of A, then any nontrivial solution of ( − ) = is an

associated eigenvector. (特徵向量之求解)

‧ 定義 8.2 Characteristic Polynomial (特徵多項式)

The polynomial | − | is the characteristic polynomial of A, and is denoted ( ).

Finding the eigenvalues of a matrix is equivalent to finding the roots of an nth degree
polynomial, and if ≥ 3 this may be difficult.

Example 8.3 求矩陣之特徵值與特徵向量（實數）

- 103 -
Example 8.4 求矩陣之特徵值與特徵向量（複數）

- 104 -
‧ 定理 8.2 Gerschgorin’s Theorem (喬斯哥林定理)

Let A be an × matrix of real or complex numbers. For = 1,2, … , , let

= | |

Let be the circle of radius centered at ( , ), where + = . Then

each eigenvalue of A, when plotted as appoint in the complex plane, lies on or within one
of the circles , , …, .

The circles are called Gerschgorin circles (喬斯哥林圓).

- 105 -
‧ 喬斯哥林圓之繪製程序―

The procedures to plot the circle are

1) read across row k (讀取第 k 列元素)

2) evaluate the radius of by adding the magnitudes of the row elements while
omitting the diagonal element (計算半徑 )

3) plot the circle centered at in the complex plane. (以 為圓心畫出圓)

Example 8.5 喬斯哥林圓之繪製

- 106 -
‧ 喬斯哥林圓的半徑可能很大，因此僅供參考，不一定對λ的求解有幫助。

Homework for sec. 8.1 ⟹ #1, 7

- 107 -
8.2 Diagonalization of Matrices (矩陣之對角線化)

We have referred to the elements of a square matrix as its main diagonal elements
(主對角線元素). All other elements are called off-diagonal elements (非主對角線元素).

‧ 定義 8.3 Diagonal Matrix (對角線矩陣)

A square matrix having all off-diagonal elements equal to zero is called a diagonal matrix.
(非對角線元素全為零之方陣，稱之為對角線矩陣)

‧ 定理 8.3 對角線矩陣之運算特性

- 108 -
 (證明留給同學自行驗證)

‧ 定理 8.4 Diagonalizability of Matrix (矩陣之可對角線化)

(若方陣 A 有 n 個線性獨立的特徵向量，則該方陣為可對角線化)
(若方陣 P 的行元素由 A 的特徵向量所組成，則 為一對角線矩陣，且主對角
線元素為 A 的特徵值)

Example 8.6 2 × 2矩陣

- 109 -
Example 8.9 3 × 3矩陣 (特徵值可能有重複，但特徵向量不會重複)

- 110 -
‧ 定理 8.5 Diagonalizable Matrices (可對角線化矩陣)

Proof:

- 111 -
(並非所有矩陣均可對角線化)

Example 8.10

- 112 -
‧ 定理 8.6 特徵值不同，對應的特徵向量必線性獨立

Proof: by induction (使用歸納法)

‧ 輔助定理 8.1 n 個特徵值都不同的 × 矩陣，必可對角線化

- 113 -
Example 8.11

Homework for sec. 8.2 ⟹ #1, 3, 15

試判定矩陣是否可對角線化，若為是，則請求
出對應之 P 矩陣。

- 114 -
8.3 Orthogonal (正交) and Symmetric (對稱) Matrices

‧ 定義 8.5 Orthogonal Matrix (實數的正交矩陣)

A real square matrix A is orthogonal if and only if = = .

(轉置矩陣等於反矩陣之實數矩陣，稱之為正交矩陣)

An orthogonal matrix is therefore non-singular, and we find its inverse simply by taking
its transpose. (正交矩陣必為非奇異)

Because the transpose of the transpose of a matrix is the original matrix, a matrix is or-
thogonal exactly when its transpose is orthogonal. (一正交矩陣之轉置矩陣仍是正交矩
陣)

Example 8.12 判定是否為正交矩陣

‧ 定理 8.7 正交矩陣之轉置矩陣，仍為正交矩陣

A is an orthogonal matrix if and only if is an orthogonal matrix.

‧ 定理 8.8 正交矩陣之行列式值，必為 1 或−1

If A is an orthogonal matrix, then | | = ±1.

- 115 -
 Proof: Since = ,

then | | = | |= 1 = | |∙| | =| |∙| |= | |∙| |= | |

∴ | | = ±1

‧ 一組向量之正交與正規正交
A set of vectors in is said to be orthogonal (正交) if any two distinct vectors in the
set are orthogonal (that is, their dot product is zero). The set is orthonormal (正規正交或
標準正交) if, in addition, each vector has length 1.

定理 8.9 「矩陣之正交」相當於 「列向量及行向量之正規正交」

Let A be a real × matrix. Then,

1) A is orthogonal if and only if the row vectors form an orthonormal set of vectors
in .

2) A is orthogonal if and only if the column vectors form an orthonormal set of vec-
tors in .

正交矩陣之例 1

- 116 -
 正交矩陣之例 2

‧ 定義 8.6 Symmetric Matrix (對稱矩陣)

A square matrix is symmetric if = .

(矩陣轉置後不改變，稱之為對稱矩陣)

For example, the matrix

is symmetric.

定理 8.10 一實數對稱矩陣之特徵值，均為實數

The eigenvalues of a real, symmetric matrix are real numbers.

定理 8.11 一實數對稱矩陣之特徵向量，彼此相互正交

Let A be a real, symmetric matrix. Then eigenvectors associated with distinct eigenvalues
are orthogonal.

Example 8.13 對稱矩陣之特徵向量具正交性

- 117 -
特徵向量之單位化與矩陣之對角線化 ―

定理 8.12 一實數對稱矩陣，必可被對角線化

Let A be a real, symmetric matrix. Then there is a real, orthogonal matrix that
diagonalizes A.

Example 8.14 求對稱矩陣 A 之對角線化正交矩陣 Q

- 118 -
Homework for sec. 8.3 ⟹ #1, 3, 5
求矩陣之特徵值及對應之特徵向量。檢查
特徵向量間是否正交，且檢查特徵向量組
成之正交矩陣是否能將原矩陣對角線化。

- 119 -
8.4 Quadratic Forms (二次式)

‧ 定義 8.7 二次式的 表示式

A (complex) quadratic form is an expression

in which each , and is a complex number.

For = 2, this quadratic form is

+ + + .
The terms involving with ≠ are the mixed product terms.

If each , and is real. We have

+( + ) + .

The terms involving and are the squared terms, and is the mixed product
term.

二次式的矩陣表示式―
It is often convenient to write a quadratic form in matrix form as

- 120 -
Example 8.15 二次式及對稱矩陣

‧ 定理 8.13 Principal Axis Theorem (主軸定理)

Let A be a real symmetric matrix with eigenvalues , , …, . Let Q be an orthogo-

nal matrix that diagonalizes A. Then the change of variables =QY transforms
∑ ∑ to ∑ . (若 Q 是可將 A 對角線化之正交矩陣，則 =QY 之
變數變換，可將二次式的 對稱式∑ ∑ 轉換成 標準式∑ )

- 121 -
Example 8.16

驗證:

[ ] 1 −1 (對稱式；非主軸)
−1 1
＝ −2 +
1 1 1 1 1 1 1 1
=( + ) −2 + ( − )+( − )
√2 √2 √2 √2 √2 √2 √2 √2
1 1 1 1 1 1
= + + −2 − + − +
2 2 2 2 2 2

=2 = [ ] 0 0 (標準式；主軸)
0 2

Example 8.17 將一橢圓從非主軸的 , 平面(對稱式)轉換至主軸的 , 平面(標準式)

- 122 -
Homework for sec. 8.4 ⟹ #1, 3, 7, 9

- 123 -
8.5 Unitary (么正矩陣), Hermitian (赫米特矩陣), and Skew-Hermitian
Matrices (反赫米特矩陣)

么正矩陣―
‧ 定義 8.8 Unitary Matrix (複數的么正矩陣)

An × complex matrix U is unitary if and only if = . (一矩陣之共軛矩

陣之反矩陣等於其轉置矩陣，稱之為么正矩陣)

The condition = is equivalent to = .

Example 8.18

‧ Unitary matrices are the complex analogues of orthogonal matrices. (么正矩陣是正交

矩陣之複數類似物)

‧ 定理 8.15 么正矩陣之特徵值之大小為 1

證明―

- 124 -
赫米特矩陣與反赫米特矩陣―
‧ 定義 8.10
1. Hermitian matrix (一矩陣之共軛矩陣等於其轉置矩陣，稱之為赫米特矩陣)
An × complex matrix H is hermitian if and only if = .

2. Skew-Hermitian matrix (反赫米特矩陣)

An × complex matrix S is skew-hermitian if and only if =− .

Example 8.20 判定為赫米特或反赫米特矩陣

- 125 -
‧ 定理 8.16 赫米特矩陣與反赫米特矩陣之二次式

Then

證明―

‧ 定理 8.17 赫米特矩陣與反赫米特矩陣之特徵值

證明―

- 126 -
‧ 各矩陣的特徵值在複數平面的位置

- 127 -
Homework for sec. 8.5 ⟹ #1, 3, 5

- 128 -