PARTIAL SOLUTION OF EIGENPROBLEMS BY A DECOMPOSITION METHOD .1. THEORY

被引:0
作者
NG, KYK
SANCHO, NGF
WONG, ML
机构
[1] UNIV OTTAWA,FAC ADM,OTTAWA,ON K1N 6N5,CANADA
[2] MCGILL UNIV,DEPT MATH & STAT,MONTREAL,PQ,CANADA
[3] CITY POLYTECH HONG KONG,DEPT APPL STAT & OPERAT RES,KOWLOON,HONG KONG
关键词
DYNAMIC PROGRAMMING; MATRIX EIGENVALUES; PARTIAL SOLUTION;
D O I
10.1080/03052159508941258
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
This paper demonstrates the feasibility of using a recently developed technique, the 'r-variable algorithm' (see references under Ng), to solve the partial eigenproblem. The method involves decomposing the Rayleigh quotient minimization problem into smaller subproblems, where each subproblem is comprised of the solution of a system of linear equations using backward and forward substitutions. The striking features of the method lie in its simplicity and the minimal storage requirements. Examples are included to illustrate the accuracy of the approach in solving symmetric eigenproblems.
引用
收藏
页码:115 / 129
页数:15
相关论文
共 28 条