约束Minimax问题的SQP-Filter算法
首发时间:2010-08-26
摘要:SQP方法是求解约束非线性规划问题最有效的方法之一,而其罚因子的适当选取往往是比较困难的,Filter技巧则可避免SQP方法中罚因子的选取。本文提出了一个求解带等式和不等式约束的Minimax问题的SQP-filter算法,每步通过求解两个二次规划子问题来得到搜索方向,并沿该方向做线搜索。该算法避免了较难的罚因子的选取,克服了Maratos效应。并在适当的假设下,得到了算法的全局收敛性和局部超线性收敛速度。
关键词: 运筹学 Minimax问题 SQP-Filter算法 全局收敛性 超线性收敛性
For information in English, please click here
SQP-Filter Method for Constrained Minimax Problem
Abstract:SQP method is one of the most effective algorithm to solve constrained nonlinear programming problem, however, which is difficult to choose the penalty factor, whereas filter techique can avoid it. In this paper,a SQP-filter method is proposed for solving minimax problems with equality and inequality constraints. In each step, a quadratic programming subproblem is solved to get the search direction which is used as line search. This method avoids the difficult of selecting the penalty factor and overcomes Maratos effects successfully. Its global convegence and local superlinear convergence rate are obtained under some suitable conditions.
Keywords: Operations research minimax problem SQP-filter method global convergence superlinear convergence
论文图表:
引用
No.4382277194250128****
同行评议
共计0人参与
勘误表
约束Minimax问题的SQP-Filter算法
评论
全部评论0/1000