正定矩阵的分解方法
设三阶
正定矩阵
A
A
A
,若矩阵
A
A
A
的特征值为
λ
1
,
λ
2
,
λ
3
\lambda_1,\lambda_2,\lambda_3
λ
1
,
λ
2
,
λ
3
,对应的
单位化
特征向量分别为
α
1
,
α
2
,
α
3
\alpha_1,\alpha_2,\alpha_3
α
1
,
α
2
,
α
3
且
两两正交
,则存在正交矩阵
Q
=
(
α
1
,
α
2
,
α
3
)
Q = (\alpha_1,\alpha_2,\alpha_3)
Q
=
(
α
1
,
α
2
,
α
3
)
,使得
Q
T
A
Q
=
Λ
=
[
λ
1
λ
2
λ
3
]
Q^{\mathrm{T}}AQ = \Lambda = \begin{bmatrix} \lambda_1 & & \\ & \lambda_2 & \\ & & \lambda_3 \end{bmatrix}
Q
T
A
Q
=
Λ
=
λ
1
λ
2
λ
3
现得到
A
=
Q
Λ
Q
T
A = Q \Lambda Q^{\mathrm{T}}
A
=
Q
Λ
Q
T
,令
Λ
1
=
[
λ
1
λ
2
λ
3
]
\Lambda_1 = \begin{bmatrix} \sqrt{\lambda_1} & & \\ & \sqrt{\lambda_2} & \\ & & \sqrt{\lambda_3} \end{bmatrix}
Λ
1
=
λ
1
λ
2
λ
3
,则正定矩阵
A
A
A
有以下分解方法:
【方法一】
A
=
Q
Λ
Q
T
=
Q
(
Λ
1
Λ
1
)
Q
T
=
(
Λ
1
Q
T
)
T
(
Λ
1
Q
T
)
=
C
T
C
(令
C
=
Λ
1
Q
T
)
A = Q \Lambda Q^{\mathrm{T}} = Q (\Lambda_1 \Lambda_1) Q^{\mathrm{T}} = (\Lambda_1 Q^{\mathrm{T}})^{\mathrm{T}} (\Lambda_1 Q^{\mathrm{T}}) = C^{\mathrm{T}} C(令 C = \Lambda_1 Q^{\mathrm{T}})
A
=
Q
Λ
Q
T
=
Q
(
Λ
1
Λ
1
)
Q
T
=
(
Λ
1
Q
T
)
T
(
Λ
1
Q
T
)
=
C
T
C
(令
C
=
Λ
1
Q
T
)
【方法二】
A
=
Q
Λ
Q
T
=
Q
Λ
1
Λ
1
Q
T
=
Q
Λ
1
(
Q
T
Q
)
Λ
1
Q
T
=
(
Q
Λ
1
Q
T
)
(
Q
Λ
1
Q
T
)
=
C
2
(令
C
=
Q
Λ
1
Q
T
)
A = Q \Lambda Q^{\mathrm{T}} = Q \Lambda_1 \Lambda_1 Q^{\mathrm{T}} = Q \Lambda_1 (Q^{\mathrm{T}} Q) \Lambda_1 Q^{\mathrm{T}} = (Q \Lambda_1 Q^{\mathrm{T}})(Q \Lambda_1 Q^{\mathrm{T}}) = C^2 (令 C = Q \Lambda_1 Q^{\mathrm{T}})
A
=
Q
Λ
Q
T
=
Q
Λ
1
Λ
1
Q
T
=
Q
Λ
1
(
Q
T
Q
)
Λ
1
Q
T
=
(
Q
Λ
1
Q
T
)
(
Q
Λ
1
Q
T
)
=
C
2
(令
C
=
Q
Λ
1
Q
T
)
此时由
谱分解定理
得:
-
A=
λ
1
α
1
α
1
T
+
λ
2
α
2
α
2
T
+
λ
3
α
3
α
3
T
A = \lambda_1 \alpha_1 \alpha_1^{\mathrm{T}} + \lambda_2 \alpha_2 \alpha_2^{\mathrm{T}} + \lambda_3 \alpha_3 \alpha_3^{\mathrm{T}}
A
=
λ
1
α
1
α
1
T
+
λ
2
α
2
α
2
T
+
λ
3
α
3
α
3
T
-
B=
λ
1
α
1
α
1
T
+
λ
2
α
2
α
2
T
+
λ
3
α
3
α
3
T
B = \sqrt{\lambda_1} \alpha_1 \alpha_1^{\mathrm{T}} + \sqrt{\lambda_2} \alpha_2 \alpha_2^{\mathrm{T}} + \sqrt{\lambda_3} \alpha_3 \alpha_3^{\mathrm{T}}
B
=
λ
1
α
1
α
1
T
+
λ
2
α
2
α
2
T
+
λ
3
α
3
α
3
T
【拓展】
若要将正定矩阵
A
A
A
分解为
C
3
C^3
C
3
,则可令
Λ
2
=
[
λ
1
3
λ
2
3
λ
3
3
]
\Lambda_2 = \begin{bmatrix} \sqrt[3]{\lambda_1} & & \\ & \sqrt[3]{\lambda_2} & \\ & & \sqrt[3]{\lambda_3} \end{bmatrix}
Λ
2
=
3
λ
1
3
λ
2
3
λ
3
,所以有:
A
=
(
Q
Λ
2
Q
T
)
(
Q
Λ
2
Q
T
)
(
Q
Λ
2
Q
T
)
=
C
3
(令
C
=
Q
Λ
2
Q
T
)
A = (Q \Lambda_2 Q^{\mathrm{T}})(Q \Lambda_2 Q^{\mathrm{T}})(Q \Lambda_2 Q^{\mathrm{T}}) = C^3 (令 C = Q \Lambda_2 Q^{\mathrm{T}})
A
=
(
Q
Λ
2
Q
T
)
(
Q
Λ
2
Q
T
)
(
Q
Λ
2
Q
T
)
=
C
3
(令
C
=
Q
Λ
2
Q
T
)
相关例题
【例 1】已知矩阵
A
=
[
0
0
1
0
1
0
1
0
0
]
A = \begin{bmatrix} 0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0 \end{bmatrix}
A
=
0
0
1
0
1
0
1
0
0
,求一个可逆矩阵
C
C
C
,使得
C
3
=
A
C^3 = A
C
3
=
A
。
【解】
C
3
=
A
C^3 = A
C
3
=
A
的特征值为
1
,
1
,
−
1
1,1,-1
1
,
1
,
−
1
,则
C
C
C
的特征值为
1
3
,
1
3
,
−
1
3
\sqrt[3]{1},\sqrt[3]{1},\sqrt[3]{-1}
3
1
,
3
1
,
3
−
1
(即
1
,
1
,
−
1
1,1,-1
1
,
1
,
−
1
),所以
C
−
E
C-E
C
−
E
的特征值为
0
,
0
,
−
2
0,0,-2
0
,
0
,
−
2
。
A
A
A
的特征值
−
1
-1
−
1
所对应的特征向量为
α
3
=
1
2
(
1
,
0
,
−
1
)
T
\alpha_3 = \frac{1}{\sqrt{2}} (1,0,-1)^{\mathrm{T}}
α
3
=
2
1
(
1
,
0
,
−
1
)
T
,则
C
−
E
C-E
C
−
E
的特征值
λ
3
=
−
2
\lambda_3 = -2
λ
3
=
−
2
所对应的特征向量也为
α
3
=
1
2
(
1
,
0
,
−
1
)
T
\alpha_3 = \frac{1}{\sqrt{2}} (1,0,-1)^{\mathrm{T}}
α
3
=
2
1
(
1
,
0
,
−
1
)
T
。
由谱分解定理得
C
−
E
=
λ
3
α
3
α
3
T
=
−
2
⋅
1
2
[
1
0
−
1
]
⋅
1
2
(
1
,
0
,
−
1
)
=
−
[
1
0
−
1
0
0
0
−
1
0
1
]
=
[
−
1
0
1
0
0
0
1
0
−
1
]
∴
C
=
(
C
−
E
)
+
E
=
[
0
0
1
0
1
0
1
0
0
]
\begin{aligned} C-E &= \lambda_3 \alpha_3 \alpha_3^{\mathrm{T}} \\ &= -2 \cdot \frac{1}{\sqrt{2}} \begin{bmatrix} 1 \\ 0 \\ -1 \end{bmatrix} \cdot \frac{1}{\sqrt{2}} (1,0,-1) \\ &= -\begin{bmatrix} 1 & 0 & -1 \\ 0 & 0 & 0 \\ -1 & 0 & 1 \end{bmatrix} = \begin{bmatrix} -1 & 0 & 1 \\ 0 & 0 & 0 \\ 1 & 0 & -1 \end{bmatrix} \\ \therefore C &= (C-E) + E = \begin{bmatrix} 0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0 \end{bmatrix} \end{aligned}
C
−
E
∴
C
=
λ
3
α
3
α
3
T
=
−
2
⋅
2
1
1
0
−
1
⋅
2
1
(
1
,
0
,
−
1
)
=
−
1
0
−
1
0
0
0
−
1
0
1
=
−
1
0
1
0
0
0
1
0
−
1
=
(
C
−
E
)
+
E
=
0
0
1
0
1
0
1
0
0
【例 2】已知矩阵
A
=
[
0
0
1
0
1
0
1
0
0
]
A = \begin{bmatrix} 0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0 \end{bmatrix}
A
=
0
0
1
0
1
0
1
0
0
,求一个正定矩阵
C
C
C
,使得
C
2
=
A
+
2
E
C^2 = A+2E
C
2
=
A
+
2
E
。
【解】
A
A
A
的特征值为
1
,
1
,
−
1
1,1,-1
1
,
1
,
−
1
,则
C
2
=
A
+
2
E
C^2 = A+2E
C
2
=
A
+
2
E
的特征值为
3
,
3
,
1
3,3,1
3
,
3
,
1
,则
C
C
C
的特征值为
3
,
3
,
1
\sqrt{3},\sqrt{3},1
3
,
3
,
1
,所以
C
−
3
E
C-\sqrt{3}E
C
−
3
E
的特征值为
0
,
0
,
1
−
3
0,0,1-\sqrt{3}
0
,
0
,
1
−
3
。
A
A
A
的特征值
−
1
-1
−
1
所对应的特征向量为
α
3
=
1
2
(
1
,
0
,
−
1
)
T
\alpha_3 = \frac{1}{\sqrt{2}} (1,0,-1)^{\mathrm{T}}
α
3
=
2
1
(
1
,
0
,
−
1
)
T
,则
C
−
3
E
C-\sqrt{3}E
C
−
3
E
的特征值
λ
3
=
1
−
3
\lambda_3 = 1-\sqrt{3}
λ
3
=
1
−
3
所对应的特征向量也为
α
3
=
1
2
(
1
,
0
,
−
1
)
T
\alpha_3 = \frac{1}{\sqrt{2}} (1,0,-1)^{\mathrm{T}}
α
3
=
2
1
(
1
,
0
,
−
1
)
T
。
由谱分解定理得
C
−
3
E
=
λ
3
α
3
α
3
T
=
(
1
−
3
)
⋅
1
2
[
1
0
−
1
]
⋅
1
2
(
1
,
0
,
−
1
)
=
1
−
3
2
[
1
0
−
1
0
0
0
−
1
0
1
]
=
[
1
−
3
2
0
3
−
1
2
0
0
0
3
−
1
2
0
1
−
3
2
]
∴
C
=
(
C
−
3
E
)
+
3
E
=
[
3
+
1
2
0
3
−
1
2
0
3
0
3
−
1
2
0
3
+
1
2
]
\begin{aligned} C-\sqrt{3}E &= \lambda_3 \alpha_3 \alpha_3^{\mathrm{T}} \\ &= (1-\sqrt{3}) \cdot \frac{1}{\sqrt{2}} \begin{bmatrix} 1 \\ 0 \\ -1 \end{bmatrix} \cdot \frac{1}{\sqrt{2}} (1,0,-1) \\ &= \frac{1-\sqrt{3}}{2} \begin{bmatrix} 1 & 0 & -1 \\ 0 & 0 & 0 \\ -1 & 0 & 1 \end{bmatrix} \\ &= \begin{bmatrix} \frac{1-\sqrt{3}}{2} & 0 & \frac{\sqrt{3}-1}{2} \\ 0 & 0 & 0 \\ \frac{\sqrt{3}-1}{2} & 0 & \frac{1-\sqrt{3}}{2} \end{bmatrix} \\ \therefore C &= (C-\sqrt{3}E) + \sqrt{3}E = \begin{bmatrix} \frac{\sqrt{3}+1}{2} & 0 & \frac{\sqrt{3}-1}{2} \\ 0 & \sqrt{3} & 0 \\ \frac{\sqrt{3}-1}{2} & 0 & \frac{\sqrt{3}+1}{2} \end{bmatrix} \end{aligned}
C
−
3
E
∴
C
=
λ
3
α
3
α
3
T
=
(
1
−
3
)
⋅
2
1
1
0
−
1
⋅
2
1
(
1
,
0
,
−
1
)
=
2
1
−
3
1
0
−
1
0
0
0
−
1
0
1
=
2
1
−
3
0
2
3
−
1
0
0
0
2
3
−
1
0
2
1
−
3
=
(
C
−
3
E
)
+
3
E
=
2
3
+
1
0
2
3
−
1
0
3
0
2
3
−
1
0
2
3
+
1
【例 3】已知矩阵
A
=
[
0
0
1
0
1
0
1
0
0
]
A = \begin{bmatrix} 0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0 \end{bmatrix}
A
=
0
0
1
0
1
0
1
0
0
,求一个正定矩阵
C
C
C
,使得
C
n
=
A
+
2
E
C^n = A+2E
C
n
=
A
+
2
E
。
【解】该题与上题的解法是相似的,现直接给出答案:
C
=
[
3
n
+
1
2
0
3
n
−
1
2
0
3
n
0
3
n
−
1
2
0
3
n
+
1
2
]
C = \begin{bmatrix} \frac{\sqrt[n]{3}+1}{2} & 0 & \frac{\sqrt[n]{3}-1}{2} \\ 0 & \sqrt[n]{3} & 0 \\ \frac{\sqrt[n]{3}-1}{2} & 0 & \frac{\sqrt[n]{3}+1}{2} \end{bmatrix}
C
=
2
n
3
+
1
0
2
n
3
−
1
0
n
3
0
2
n
3
−
1
0
2
n
3
+
1