摘要
设A和B是两个任意的n阶方阵,其特征值分别为{λ_1,…,λ_n}和{μ_1,…,μ_n}.本文对此两组特征值的如下“距离”的界给出了若干估计: B对于A的谱改变量 A与B的特征值的改变量这里的结果包含了Bauer-Fike定理,并且优于Kahan-Parlett/Jiang定理及Chu,施和肖所得出的结果.
For arbitrary matrices A and B with spectra {λ_1,…,λ_n}and{μ_1,…,μ_n}the following are estimated, i.e. the bound of the following 'distances' between spectra the spectral variation of B with respect to A and the eigenvalue variation of A and B. The results include the Bauer-Fike Theorem and are better than the Kahan-Parlett-Jiang Theorem and the theorem of Chu, Shi and Xiao.
出处
《应用数学》
CSCD
北大核心
1992年第4期19-25,共7页
Mathematica Applicata
关键词
矩阵
特征值
扰动
估计
Matrix
Eigenvalue
Perturbation
