一个求解非线性不等式约束优化问题的带有共轭梯度参数的广义梯度投影算法(英文)  

A Generalized Gradient Projection Algorithmwith Conjugate Gradient Parameter for NonlinearInequality Constrained Optimization Problems

作　　者：陈翠玲[1] 李明[2] 李略[1] 

机构地区：[1]广西师范大学数学科学学院,广西桂林541004 [2]桂林理工大学理学院,广西桂林541004

出　　处：《应用数学》2011年第2期399-406,共8页Mathematica Applicata

基　　金：Supported by the National Natural Science Foundation (10961006);Guangxi Education Department Science Foundation (200911LX53);Guangxi Normal University Research Foundation for the Youth Backbone Teacher of China

摘　　要：本文提出一个求解非线性不等式约束优化问题的带有共轭梯度参数的广义梯度投影算法.算法中的共轭梯度参数是很容易得到的,且算法的初始点可以任意选取.而且,由于算法仅使用前一步搜索方向的信息,因而减少了计算量.在较弱条件下得到了算法的全局收敛性.数值结果表明算法是有效的.In this paper,we propose a new generalized gradient projection algorithm with conjugate gradient parameter for nonlinear inequality constrained optimizatiori problems. In this algorithm,it is very simple to obtain the conjugate gradient parameter, and it is arbitrary to choose the initial point. Moreover it only uses the last search direction information, which reduces the computational cost. The global convergence of this algorithm is obtained under weak conditions. The numerical results show the algorithm is effect.

关 键 词：不等式约束优化 广义梯度投影算法 共轭梯度参数 全局收敛性 

分 类 号：O221[理学—运筹学与控制论]