摘要
本文提出了一个解广义特征值问题的方法,即用半分法确定近似特征值,用改进的Rayleigh商迭代校正近似特征值,同时求出相应的特征向量。理论证明它是一个收敛的方法。
An algorithm of solving the generalized symmetric eigenvalue problem Ax=λBx (B is nearly singular or positive semidefinite) is presented. Firstly, the approximate eigenvalue is determined by the semidefinite method. Then, the approximate eigenvalue is corrected by using the modified Rayleigh quotient iteration and the corresponding eigenvector is obtained. The convergence of this algorithm is proved.
出处
《南京邮电学院学报》
北大核心
1991年第4期84-92,共9页
Journal of Nanjing University of Posts and Telecommunications(Natural Science)
关键词
广义待征值
半分法
瑞利商迭代
Generalized eigenvalue, Semidefinite, Rayleigh quotient iteration