摘要
本文提出了一个关于大型带状稀疏矩阵特征值问题的于空间反迭代法的几种算法和二分法相结合的优化算法.它特别注意那些带幅较宽且有多个密集特征值群的情形,也考虑了所期望的特征值的个数较多且其精度也要高的情形CPU和USE的负担等问题.
In this paper we discuss composite optimzing computational algorithms that combine several computational algorithms of subspace inverse iteration method in the eigenvalues problems of a large scale sparse band symmetric matrix with bisection method. Especially we consider following two cases. The first has wider range bandwidth and some groups of dense eigenvalues, the second has more number of expected eigenvalues and has high precision required, i. e. the problems of CPU and USE burden.
出处
《延边大学学报(自然科学版)》
1995年第1期1-7,共7页
Journal of Yanbian University(Natural Science Edition)
关键词
迭代法
特征值
矩阵
子空间迭代法
Subspace inverse iteration method
block inverse power method
biscction method