结构J-对称矩阵特征值问题的算法

Algorithms for Structured J-Symmetric Eigenvalue Problems

本文给出了计算结构J-对称矩阵特征值的算法。这些算法以Van Loan的平方约化方法和Cullum与Willoughby提出的计算复对称三对角矩阵特征值的QL过程为基础,利用了J-对称矩阵的特殊结构,可以节省计算量和存储量。

Algorithms for computing eigenvalues of structured J-symmetric matrices are presented. Based on an analog of Van Loan＇s square reduced method for Hamiltonian matrices and a QL procedure proposed by Cullum and Willoughby for computing eigenvalues of complex symmetric tridiagonal matrices, these algorithms are structure-preserving, and hence the cost of computation and storage is reduced.