首页 | 官方网站   微博 | 高级检索  
     

解约束优化问题的QP-free可行域方法
引用本文:周岩,濮定国.解约束优化问题的QP-free可行域方法[J].运筹学学报,2007,11(3):31-43.
作者姓名:周岩  濮定国
作者单位:1. 青岛大学管理科学与工程系,青岛,266071
2. 同济大学应用效学系,上海,200092
摘    要:本文利用一个新的分片线性NCP函数提出一个新的可行的QP-free方法解非线性不等式约束优化问题.不同于其他的QP-free方法,这个方法只考虑在工作集中的约束函数,工作集是积极集的一个估计,因此子问题的维数不是满秩的.这个方法可行的并且不需假定严格互补条件、聚点的孤立性得到算法的全局收敛性,并且积极约束函数的梯度不要求线性独立的,其中由拟牛顿法得到的子矩阵不需要求一致正定性.

关 键 词:运筹学  不等式约束优化  积极集  收敛性
修稿时间:2006-02-17

A New QP-free Feasible Method for Inequality Constrained Optimization
Zhou Yan,Pu Dingguo.A New QP-free Feasible Method for Inequality Constrained Optimization[J].OR Transactions,2007,11(3):31-43.
Authors:Zhou Yan  Pu Dingguo
Affiliation:1.Depaxtment of Management Science and Engineering, Qingda, University, Qingda, 266071, China ;2.Department of Applied Mathematics, Tongji University, Shanghai 200092, China
Abstract:In this paper,a new QP-free feasible method is proposed for solving inequality constrained optimization problems,by a new piecewise linear NCP functions. Unlike the existing QP-free algorithms,the proposed method is concerned with only the constraints in the working set,which is an estimate of the active set.Consequently,the dimension of the subproblems is not full dimensional.This method is implementable and globally convergent without assuming that the strict complomentarity condition,the isolatedness of the accumulation points.Furthermore,the gradients of active constraints are not requested to be linearly independent.The submatrix,which is obtained by quasi Newton methods,isn't requested to be uniformly positive definite.
Keywords:Operations research  inequality constrained optimization  active set strategy  convergence
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号