摘要
文章基于前人的工作 ,在哈密尔顿矩阵约化过程中 ,采用了辛相似变换 ,使得哈密尔顿矩阵在辛相似变换下仍保持Hamilton结构 ,这样从根本上确保了特征值的正确性和稳定性 ,也能保证特征值成对出现且在每个半平面上都只求得 n个特征值 。
The symplectic algorithm method for Hamiltonian matrix eigenvalue and eigenvector problem is presented based upon predecessors' papers. This method works with symplectic similarity transformation which reflects the structure of the spectrum of Hamiltonian matrices, and has a higher numerical accuracy than ordinary algorithm.It is therefore guaranteed that the computed eigenvalues occur in plus minus pairs and ensure numerical stabilities.This algorithm can find exactly n eigenvalues in each half plane. Even small perturbations may not cause the computed eigenvalues to cross the imaginary axis.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
2000年第3期336-340,共5页
Journal of Hefei University of Technology：Natural Science
基金
国家自然科学基金资助项目! (1 9392 30 5)
安徽省教委基金资助项目 !(99j1 o2 2 6zc)
关键词
辛算法
哈密尔顿矩阵
辛相似变换
特征问题
symplectic algorithm
Hamiltonian matrix
symplectic similarity transformation
eigen problem
