一种新的无罚函数无滤子的SQP算法  

An SQP Method without a Penalty Function or a Filter for Inequality Constrained Optimization

作　　者：王祥玲 WANG Xiangling(Department of Primary Education,Yichun Early Childhood Teachers College,Yichun Jiangxi 336000,China)

机构地区：[1]宜春幼儿师范高等专科学校初等教育学院,江西宜春336000

出　　处：《保山学院学报》2023年第2期33-37,共5页JOURNAL OF BAOSHAN UNIVERSITY

基　　金：江西省教育厅科学技术研究项目“约束优化问题的无罚函数无滤子的SQP法研究”(项目编号:GJJ2209202)。

摘　　要：通过修正QP子问题和构造适当的线搜索方向,提出了一种新的无罚函数无滤子的方法来求解不等式约束优化问题。该方法有效地避免了罚函数选择的困难,避开了传统滤子法所需要的可行恢复阶段。最后在适当的假设条件下,给出了算法的有效性分析和全局收敛性分析。In this paper,a new method without penalty function or a filter is proposed to solve inequali⁃ty constrained optimization problems by modifying the QP sub-problem and constructing the appropri⁃ate line search direction.The method avoids the difficulty of selecting penalty function and the feasible recovery stage required by traditional filter method.Finally,the global convergence and effectiveness of the algorithm are given under appropriate assumptions.

关 键 词：不等式约束优化 SQP 线搜索 全局收敛性 

分 类 号：O1[理学—数学]