特征值和特征向量
在数学上,特别是线性代数中,对于一个给定的线性变换
A
A
A, 它的特征向量
v
v
v 经过这个线性变换的作用之后,得到的新向量仍然与原来的
v
v
v 保持在同一条直线上, 但其长度或方向也许会改变。 即
A
v
=
λ
v
Av= \lambda v
Av=λv
其中
λ
\lambda
λ 为标量,即特征向量的长度在该线性变换下的缩放比例,程为特征值。
举例:
A
=
[
1
1
4
−
2
]
A = \begin{bmatrix} 1 & 1 \\ 4 & -2 \\ \end{bmatrix}
A=[141−2]
v 1 = [ 1 2 ] v_1 = \begin{bmatrix} 1 \\ 2 \\ \end{bmatrix} v1=[12] \quad A v 1 = [ 3 0 ] Av_1 = \begin{bmatrix} 3 \\ 0 \\ \end{bmatrix} Av1=[30]
v 2 = [ 1 1 ] v_2 = \begin{bmatrix} 1 \\ 1 \\ \end{bmatrix} v2=[11] \quad A v 2 = [ 2 2 ] = 2 [ 1 1 ] Av_2 = \begin{bmatrix} 2 \\ 2 \\ \end{bmatrix} = 2\begin{bmatrix} 1 \\ 1 \\ \end{bmatrix} Av2=[22]=2[11]
A
v
2
Av_2
Av2 与
v
2
v_2
v2 在一条直线上
A
v
2
=
2
v
2
Av_2 = 2v_2
Av2=2v2
\quad
v
2
v_2
v2是特征向量,2是特征值
求解特征值,特征向量
A
v
=
λ
v
Av= \lambda v
Av=λv
A
v
−
λ
v
=
0
Av - \lambda v = 0
Av−λv=0
(
A
−
λ
I
)
v
=
0
(A - \lambda I)v = 0
(A−λI)v=0 有非零解
∣
A
−
λ
I
∣
=
0
\mid A - \lambda I \mid = 0
∣A−λI∣=0
A
=
[
1
1
4
−
2
]
A = \begin{bmatrix} 1 & 1 \\ 4 & -2 \\ \end{bmatrix}
A=[141−2]
A
−
λ
I
=
[
1
−
λ
1
4
−
2
−
λ
]
=
0
A - \lambda I = \begin{bmatrix} 1 - \lambda & 1 \\ 4 & -2-\lambda \\ \end{bmatrix} =0
A−λI=[1−λ41−2−λ]=0
(
1
−
λ
)
(
−
2
−
λ
)
−
1
∗
4
=
0
(1-\lambda)(-2-\lambda)-1*4=0
(1−λ)(−2−λ)−1∗4=0
(
λ
−
2
)
(
λ
+
3
)
=
0
(\lambda-2)(\lambda+3)=0
(λ−2)(λ+3)=0
λ
1
=
2
λ
2
=
−
3
\lambda_1=2 \quad \lambda_2=-3
λ1=2λ2=−3
带入原方程
当
λ
1
=
2
\lambda_1 = 2
λ1=2
[ 1 − 2 1 4 − 2 − 2 ] v 1 = 0 \begin{bmatrix} 1-2 & 1 \\ 4 & -2-2 \\ \end{bmatrix} v_1 = 0 [1−241−2−2]v1=0
[ − 1 1 4 − 4 ] [ v 11 v 12 ] = 0 \begin{bmatrix} -1 & 1 \\ 4 & -4 \\ \end{bmatrix} \begin{bmatrix} v_{11} \\ v_{12} \\ \end{bmatrix}= 0 [−141−4][v11v12]=0
− v 11 + v 12 = 0 4 v 11 − 4 v 12 = 0 → v 11 = v 12 -v_{11}+v_{12}=0 \quad 4v_{11}-4v_{12}=0 \rightarrow v_{11} = v_{12} −v11+v12=04v11−4v12=0→v11=v12
任取一组解
当 λ 1 = 2 \lambda_1 = 2 λ1=2 \quad v 1 = [ 1 1 ] v_1= \begin{bmatrix} 1 \\ 1 \\ \end{bmatrix} v1=[11]
同理,当 λ 2 = − 3 \lambda_2 = -3 λ2=−3
[ 1 − ( − 3 ) 1 4 − 2 − ( − 3 ) ] [ v 21 v 22 ] = 0 \begin{bmatrix} 1-(-3) & 1 \\ 4 & -2-(-3) \\ \end{bmatrix} \begin{bmatrix} v_{21} \\ v_{22} \\ \end{bmatrix}= 0 [1−(−3)41−2−(−3)][v21v22]=0
4 v 21 + v 22 = 0 4v_{21}+v_{22}=0 4v21+v22=0
任取一组解
同理,当 λ 2 = − 3 \lambda_2 = -3 λ2=−3 \quad v 2 = [ 1 − 4 ] v_2= \begin{bmatrix} 1 \\ -4 \\ \end{bmatrix} v2=[1−4]
应用
代对角矩阵,解耦 Decouple
P = [ v 1 v 2 ] P= \begin{bmatrix} v_1 & v_2 \\ \end{bmatrix} P=[v1v2] P : c o o r d i n a t e t r a n s f o r m a t i o n m a t r i x \quad \quad P:coordinate \;\; transformation \;\; matrix P:coordinatetransformationmatrix
A P = A [ v 1 v 2 ] = A [ v 11 v 21 v 12 v 22 ] = [ A [ v 11 v 12 ] A [ v 21 v 22 ] ] = [ λ 1 [ v 11 v 12 ] λ 2 [ v 21 v 22 ] ] = [ λ 1 v 11 λ 2 v 21 λ 1 v 12 λ 2 v 22 ] = [ v 11 v 21 v 12 v 22 ] [ λ 1 0 0 λ 2 ] \begin{aligned} AP &=A \begin{bmatrix} v_1 & v_2 \\ \end{bmatrix}=A \begin{bmatrix} v_{11} & v_{21} \\ v_{12} & v_{22} \\ \end{bmatrix} = \begin{bmatrix} A\begin{bmatrix} v_{11} \\ v_{12} \end{bmatrix} & A \begin{bmatrix} v_{21} \\ v_{22} \end{bmatrix} \end{bmatrix} = \begin{bmatrix} \lambda_1 \begin{bmatrix} v_{11} \\ v_{12} \end{bmatrix} & \lambda_2 \begin{bmatrix} v_{21} \\ v_{22} \end{bmatrix} \end{bmatrix} \\ &= \begin{bmatrix} \lambda_1 v_{11} & \lambda_2 v_{21} \\ \lambda_1 v_{12} & \lambda_2 v_{22} \\ \end{bmatrix} = \begin{bmatrix} v_{11} & v_{21} \\ v_{12} & v_{22} \\ \end{bmatrix} \begin{bmatrix} \lambda_1 & 0 \\ 0 & \lambda_2 \\ \end{bmatrix} \end{aligned} AP=A[v1v2]=A[v11v12v21v22]=[A[v11v12]A[v21v22]]=[λ1[v11v12]λ2[v21v22]]=[λ1v11λ1v12λ2v21λ2v22]=[v11v12v21v22][λ100λ2]
A
P
=
P
Λ
AP=P \Lambda
AP=PΛ
P
−
1
A
P
=
P
−
1
P
Λ
=
Λ
P^{-1}AP=P^{-1}P \Lambda= \Lambda
P−1AP=P−1PΛ=Λ
原视频:
https://www.bilibili.com/video/av15201777?from=search&seid=15999928462480899873