由FR共轭梯度法控制的两类优化算法的全局收敛性  

GLOBAL CONVERGENCE PROPERTIES OF TWO CLASSES OF OPTIMAL ALGORITHMS CONSTRAINED BY THE FR CONJUGATE GRADIENT METHOD

作　　者：杜学武[1] 徐成贤[2] 

机构地区：[1]焦作工学院基础科学部,焦作454159 [2]西安交通大学科学计算与应用软件系,西安710049

出　　处：《高等学校计算数学学报》2000年第4期311-318,共8页Numerical Mathematics A Journal of Chinese Universities

基　　金：河南省教委自然科学基金资助课题!2000110004.

摘　　要：In this paper, we prove that two classes of general methods for unconstrained optimization which is constrained by the Fletcher-Reeves conjugate gradient method are globally convergent under a kind of inexact line search conditions. Results of numerical experiments for several methods are presented.In this paper, we prove that two classes of general methods for unconstrained optimization which is constrained by the Fletcher-Reeves conjugate gradient method are globally convergent under a kind of inexact line search conditions. Results of numerical experiments for several methods are presented.

关 键 词：FR共轭梯度法 优化算法 全局收敛性 

分 类 号：O242.23[理学—计算数学]