我的龟孙啊,这么个简单的玩意卡住了好几天
-
有矩阵$P$:
\begin{equation}
P=
\left(
\begin{matrix}
3&1\\
1&2
\end{matrix}
\right)
\end{equation}
$P$的单位化特征向量:
\begin{equation}
v_1=\frac{1}{0.25(1+\sqrt{5})^2+0.25}\left(
\begin{matrix}
0.5(1+\sqrt{5})\\
1
\end{matrix}
\right)=
\left(
\begin{matrix}
0.4472135954999579\\
0.276393202250021
\end{matrix}
\right),
v_1=\frac{1}{0.25(1-\sqrt{5})^2+0.25}\left(
\begin{matrix}
0.5(1-\sqrt{5})\\
1
\end{matrix}
\right)
\end{equation}
$P$的特征值:
\begin{equation}
\lambda_1=0.5(5+\sqrt{5}),\lambda_2=0.5(5-\sqrt{5})
\end{equation}$P$的谱分解(spectral decomposition):
\begin{equation}\label{4}
P=Q\cdot\Lambda\cdot Q^T
\end{equation}
\begin{equation}
Q=(v_1,v_2)
\end{equation}
将$Q$,$\Lambda$代入到方程\eqref{4},有:
\begin{equation}\label{5}
Q\cdot\Lambda\cdot Q^T=\left(
\begin{matrix}
3&-1\\
-1&2
\end{matrix}
\right)
\end{equation}
竟然不是原来的$P$!!咋回事呢