The global convergence of the Polak-Ribiere-Polyak conjugate gradient algorithm under inexact line search for nonconvex functions

作     者：Yuan, Gonglin Wei, Zengxin Yang, Yuning 

作者机构：Guangxi Univ Coll Math & Informat Sci Nanning Guangxi Peoples R China 

出 版 物：《JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS》 (计算与应用数学杂志)

年 卷 期：2019年第362卷第0期

页      面：262-275页

核心收录：

学科分类：07[理学] 070104[理学-应用数学] 0701[理学-数学] 

基　　金：National Natural Science Foundation of China Guangxi Natural Science Fund for Distinguished Young Scholars [2015GXNSFGA139001] Guangxi Natural Science Key Fund [2017GXNSFDA198046] 

主　　题：PRP method Global convergence Inexact line search Nonconvex functions 

摘      要：Powell (1984) and Dai (2003) constructed respectively a counterexample to show that the Polak-Ribiere-Polyak (PRP) conjugate gradient algorithm fails to globally converge for nonconvex functions even when the exact line search technique is used, which implies similar failure of the weak Wolfe-Powell (WWP) inexact line search technique. Does another inexact line search technique exist that can ensure global convergence for nonconvex functions? This paper gives a positive answer to this question and proposes a modified WWP (MWWP) inexact line search technique. An algorithm is presented that uses the MWWP inexact line search technique;the next point chi(k+1) generated by the PRP formula is accepted if a positive condition holds, and otherwise, chi(k+1) is defined by a technique for projection onto a parabolic surface. The proposed PRP conjugate gradient algorithm is shown to possess global convergence under inexact line search for nonconvex functions. Numerical performance of the proposed algorithm is shown to be competitive with those of similar algorithms. (C) 2018 Elsevier B.V. All rights reserved.