等式约束优化非单调信赖域算法(英文)  

A Nonmonotone Trust Region Method for Nonlinear Equality Constrained Optimization Problems

作　　者：张新华[1] 张浩[1] 

机构地区：[1]南京农业大学工学院,江苏南京210031

出　　处：《应用数学》2011年第4期684-690,共7页Mathematica Applicata

基　　金：the Natural Science Foundation of China(71001053)

摘　　要：设计了一个新的求解等式约束优化问题的非单调信赖域算法.该算法不需要罚函数也无需滤子.在每次迭代过程中只需求解满足下降条件的拟法向步及切向步.新算法产生的迭代步比滤子方法更易接受,计算量比单调算法小.在一般条件下,算法具有全局收敛性.We propose and analyze a class of nonmonotone trust region algorithms for solving nonlinear equality constrained optimization problems.This method does not need penalty function and filter.Each trial step is composed of a quasi-normal step and a tangential step.Both steps are required to satisfy a decrease condition for their respective trust region subproblems.This new method has more flexibility for the acceptance of the trial step compared to the filter methods,and requires less computational costs compared with the monotone methods.Under reasonable conditions,the global convergence is established.

关 键 词：非线性规划 非单调 滤子 信赖域算法 

分 类 号：O221.2[理学—运筹学与控制论]