A = \begin{bmatrix}{-C} & {-M} \\ {I} & {0}\end{bmatrix}, \quad B = \begin{bmatrix}{K} & {0} \\ {0} & {I}\end{bmatrix}, \quad z = \begin{bmatrix}{x} \\ {\lambda x}\end{bmatrix}. \tag{9.14}
注意,从因式分解
B - \lambda A =\begin{bmatrix}{I} & {\lambda M} \\ {0} & {I}\end{bmatrix}\begin{bmatrix}{\lambda^2 M + \lambda C + K} & {0} \\ {0} & {I}\end{bmatrix}\begin{bmatrix}{I} & {0} \\ {-\lambda I} & {I}\end{bmatrix}
我们知道矩阵束B - \lambda A等价于
\begin{bmatrix}{\lambda^2 M + \lambda C + K} & {0} \\ {0} & {I}\end{bmatrix} .
由于\det(B - \lambda A) = \det( \lambda^2 M + \lambda C + K ),我们得出结论:矩阵束B - \lambda A是正则的当且仅当二次矩阵多项式\lambda^2 M + \lambda C + K是正则的,并且原始QEP(9.1)的特征值与矩阵束B - \lambda A的特征值一致。