非线性极大极小问题一个修正FR共轭梯度法  

A Modified FR Conjugate Gradient Algorithm for Nonlinear Minimax Problems

作　　者：靳文慧 刘美杏 JIN Wen-hui;LIU Mei-xing(Office Statistics,of Academic Affairs,Yulin Normal University Guangxi Colleges and Universities Key Lab of Yulin Normal University,Yulin,Guangxi 537000;School of Mathematics and Complex System Optimization and LargeData Processing,Yulin,Guangxi 537000)

机构地区：[1]玉林师范学院教务处,广西玉林537000 [2]玉林师范学院数学与统计学院/广西高校复杂系统优化与大数据处理重点实验室,广西玉林537000

出　　处：《玉林师范学院学报》2018年第2期14-18,共5页Journal of Yulin Normal University

基　　金：广西自然科学基金项目(2015GXNSFBA139001);广西高校科研项目(KY2015YB242)

摘　　要：基于指数光滑化函数,提出了一个求解非线性极大极小问题修正FR共轭梯度算法.在每一次迭代中,通过Armjio非精确线搜产生步长,利用几何减小策略调整指数光滑化函数中的罚参数.在适当的假设条件下,新算法具有全局收敛性.Based on the exponential smoothing function,we proposed a modified FR conjugate gradient algorithm for solving nonlinear minimax problems.At each iteration,Armjio inexact line search was used to generate step size and the penalty parameter in exponential smoothing function was adjusted by geometric reductionstrategy.Under some mild assumptions,the proposed algorithm was globally convergent.

关 键 词：非线性极大极小问题 指数光滑化函数 修正FR共轭梯度法 全局收敛性 

分 类 号：O221.2[理学—运筹学与控制论]