跟小D每日学口语

微分方程解耦

  对于实系数线性微分方程组

\[X' = AX: = \left\{ \begin{array}{l}{{x'}_1} = a{x_1} + b{x_2}\\{{x'}_2} = c{x_1} + d{x_2}\end{array} \right.\]

  很显然,每一个微分方程都是两个自变量的函数,如果矩阵\(A\)有完备的特征向量(且不能有复特征值),即存在两个线性无关的特征向量\(\overrightarrow {{\alpha _1}} ,\overrightarrow {{\alpha _2}} \),

\[\begin{array}{l}AE = A\left[ {\overrightarrow {{\alpha _1}} \;\;\overrightarrow {{\alpha _2}} } \right] = \left[ {A\overrightarrow {{\alpha _1}} \;\;A\overrightarrow {{\alpha _2}} } \right]\\\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; = \left[ {{\lambda _1}\overrightarrow {{\alpha _1}} \;\;{\lambda _2}A\overrightarrow {{\alpha _2}} } \right]\\\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; = \left[ {\overrightarrow {{\alpha _1}} \;\;\overrightarrow {{\alpha _2}} } \right]\left[ \begin{array}{l}{\lambda _1}\;\;\;0\\0\;\;\;{\lambda _2}\end{array} \right]\\\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; = E\left[ \begin{array}{l}{\lambda _1}\;\;\;0\\0\;\;\;{\lambda _2}\end{array} \right]\end{array}\]

从而有

\[{E^{ - 1}}AE = \left[ \begin{array}{l}{\lambda _1}\;\;\;0\\0\;\;\;{\lambda _2}\end{array} \right]\]

以上就是对角化的过程,即用线性无关的特征向量组成矩阵\(E\)。尽管\(\overrightarrow {{\alpha _1}} ,\overrightarrow {{\alpha _2}} \)线性无关,但未必正交,如果能找到一个变换,将\(\overrightarrow {{\alpha _1}} ,\overrightarrow {{\alpha _2}} \)变换为两个正交的向量,通过变量代换,则可以将原方程组中的两个耦合的方程组化为不耦合的两个方程组,即完成解耦。

  可以基于这一想法来寻找变换,即如果将原坐标系下的\(\left( \begin{array}{l}{x_1}\\{x_2}\end{array} \right) = \overrightarrow {{\alpha _1}} \to \left( \begin{array}{l}0\\1\end{array} \right)\),\(\left( \begin{array}{l}{x_1}\\{x_2}\end{array} \right) = \overrightarrow {{\alpha _2}} \to \left( \begin{array}{l}1\\0\end{array} \right)\),即完成了正交化,意思是原坐标系下坐标分别为\(\overrightarrow {{\alpha _1}} ,\overrightarrow {{\alpha _2}} \)的向量,在新坐标系下变成了\(\left( \begin{array}{l}0\\1\end{array} \right),\left( \begin{array}{l}1\\0\end{array} \right)\),如果新坐标系用\(\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right)\)来表示,那么,\(\left( \begin{array}{l}{x_1}\\{x_2}\end{array} \right)\)和\(\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right)\)之间是什么关系呢?这种关系即是我们需要寻找的变换,其实上述思想已经暗含了变换,即

\[\left( \begin{array}{l}{x_1}\\{x_2}\end{array} \right) = \left[ {\overrightarrow {{\alpha _1}} \;\;\overrightarrow {{\alpha _2}} } \right]\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right) = E\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right)\]

取\(\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right) = \left( \begin{array}{l}0\\1\end{array} \right)\)和\(\left( \begin{array}{l}{u_1}\\{u_2}\end{array} \right) = \left( \begin{array}{l}1\\0\end{array} \right)\)代入上式即可验证。从变换中,可以解出\({x_1} = p{u_1} + q{u_2},{x_2} = r{u_1} + s{u_2}\),代入原微分方程组,得到以\({u_1},{u_2}\)为自变量的微分方程组,可以证明,方程组中的两个方程不耦合。由以上看出,由特征向量组成的矩阵在对角化和解耦合中的意义。

posted on 2013-10-31 12:12  湘厦人  阅读(2857)  评论(0编辑  收藏  举报

导航