矩阵乘法的定义
$$ \left\{\begin{align*} y_1 &= a_{11} x_1 + a_{12} x_2 + \cdots + a_{1N} x_N \\ y_2 &= a_{21} x_1 + a_{22} x_2 + \cdots + a_{2N} x_N \\ & \ \ \vdots \\ y_N &= a_{N1} x_1 + a_{N2} x_2 + \cdots + a_{NN} x_N \\ \end{align*}\right.
\left\{\begin{align*} z_1 &= b_{11} y_1 + b_{12} y_2 + \cdots + b_{1N} y_N \\ z_2 &= b_{21} y_1 + b_{22} y_2 + \cdots + b_{2N} y_N \\ & \ \ \vdots \\ z_N &= b_{N1} y_1 + b_{N2} y_2 + \cdots + b_{NN} y_N \\ \end{align*}\right. \\
\boldsymbol y = \boldsymbol A \boldsymbol x, \ \ \boldsymbol z = \boldsymbol B \boldsymbol y
$$
可以根据矩阵乘法原理结合,把第一式中的 $\boldsymbol y_i$ 代入到第二式中,即等同于:
$$ \left\{\begin{align*}
z_1 &= b_{11} (a_{11} x_1 + a_{12} x_2 + \cdots + a_{1N} x_N ) + b_{12} (a_{21} x_1 + a_{22} x_2 + \cdots + a_{2N} x_N) + \cdots + b_{1N} (a_{N1} x_1 + a_{N2} x_2 + \cdots + a_{NN} x_N) \\
z_2 &= b_{21} (a_{11} x_1 + a_{12} x_2 + \cdots + a_{1N} x_N ) + b_{22} (a_{21} x_1 + a_{22} x_2 + \cdots + a_{2N} x_N) + \cdots + b_{2N} (a_{N1} x_1 + a_{N2} x_2 + \cdots + a_{NN} x_N) \\
& \ \ \vdots \\
z_N &= b_{N1} (a_{11} x_1 + a_{12} x_2 + \cdots + a_{1N} x_N ) + b_{N2} (a_{21} x_1 + a_{22} x_2 + \cdots + a_{2N} x_N) + \cdots + b_{NN} (a_{N1} x_1 + a_{N2} x_2 + \cdots + a_{NN} x_N) \\
\end{align*}\right. $$
整理得:
$$ \left\{\begin{align*}
z_1 &= (b_{11} a_{11} + b_{12}a_{21} + \cdots b_{1N}a_{N1})x_1 + (b_{11} a_{12} + b_{12}a_{22} + \cdots b_{1N}a_{N2})x_2 + \cdots + (b_{11} a_{1N} + b_{12}a_{2N} + \cdots b_{1N}a_{NN})x_N \\
z_2 &= (b_{21} a_{11} + b_{22}a_{21} + \cdots b_{2N}a_{N1})x_1 + (b_{21} a_{12} + b_{22}a_{22} + \cdots b_{2N}a_{N2})x_2 + \cdots + (b_{21} a_{1N} + b_{22}a_{2N} + \cdots b_{2N}a_{NN})x_N \\
& \ \ \vdots \\
z_N &= (b_{N1} a_{11} + b_{N2}a_{21} + \cdots b_{NN}a_{N1})x_1 + (b_{N1} a_{12} + b_{N2}a_{22} + \cdots b_{NN}a_{N2})x_2 + \cdots + (b_{N1} a_{1N} + b_{N2}a_{2N} + \cdots b_{NN}a_{NN})x_N \\
\end{align*}\right. $$
记为:
$$ \boldsymbol z = \boldsymbol B \boldsymbol A \boldsymbol x, \ \ (\boldsymbol B \boldsymbol A){u, v} = \sum{s=1}^{N} b_{us} a_{sv} $$
推广到非方阵:
$$ \boldsymbol B_{N\times S} \boldsymbol A_{S\times M} = (\boldsymbol B \boldsymbol A){N\times M}, \ \ \ (\boldsymbol B \boldsymbol A){u, v} = \sum_{s=1}^{S} b_{us} a_{sv} $$
矩阵乘法的性质:
矩阵乘法不具备的性质:
【不满足】交换律(commutative law),不保证 $\boldsymbol A \boldsymbol B = \boldsymbol B \boldsymbol A$。
【不满足】消去律(cancellation law):
此二者的经典例子:
$$ \boldsymbol A \boldsymbol B = \begin{bmatrix} 1 & 0 \\ 1 & 0 \end{bmatrix}\begin{bmatrix} 0 & 0 \\ 1 & 1 \end{bmatrix} = \begin{bmatrix} 0 & 0 \\ 0 & 0 \end{bmatrix} \\ \boldsymbol B \boldsymbol A = \begin{bmatrix} 0 & 0 \\ 1 & 1 \end{bmatrix}\begin{bmatrix} 1 & 0 \\ 1 & 0 \end{bmatrix} = \begin{bmatrix} 0 & 0 \\ 2 & 0 \end{bmatrix} $$
方幂的定义:
$$ \boldsymbol A^k = \underbrace{\boldsymbol A\boldsymbol A \cdots \boldsymbol A}_{k 个} $$
方幂的性质:
转置的运算律:
对称矩阵(symmetric matrix):$\boldsymbol A^{\rm T} = \boldsymbol A$
反对称矩阵(antisymmetric matrix):$\boldsymbol A^{\rm T} = -\boldsymbol A$