均衡约束数学规划问题序列二次规划算法的研究.pdfVIP

摘要 摘 要 含均衡约束的数学规划问题（mathematical programs with equilibrium constraints,简写MPEC）是近年来运筹学领域中的一个热点问题该问题起源于经. 济问题，与著名的对策论有着紧密的联系.由于这类问题在经济、工程技术、运 输网络的设计和对策论等领域，有着极为广泛的应用，所以受到人们的关注.然 而由于互补约束的存在，该问题又是难解的，现有的求解非线性约束优化问题的 算法不能直接用来求解此类问题. 本文给出了两个求解互补约束优化问题的有效算法，具体研究成果有以下两 个方面： 第一部分：对线性互补约束优化问题，利用一个连续可微的光滑互补函数， 光滑系数趋于零时，原问题转化为光滑非线性规划问题，再利用SQP算法来求 解该光滑非线性规划问题在适当的条件下该算法具有全局收敛性. , . 第二部分：针对一般互补约束优化问题，提出一个新的SQP 算法，该方法 通过引入Lagrange函数，通过乘子修正为此问题提供一个新的算法.在较弱的条 件下，证明了该算法收敛到分片稳定点，进一步在MPEC-LICQ条件下，证明了 算法收敛于S-稳定点. 最后，通过数值试验的结果证明了算法的可行性和有效性. 关键词：均衡问题; 罚函数; SQP法; 全局收敛. I 万方数据 Abstract Abstract In recent years，mathematical programs with equilibrium constraints (MPEC for short) have become a focus in operational research. It stems from the economical problem, and has close relation with game theory .Because this kind of problem are widely used in economy, engineering design, the design of transportation networks and game theory etc, so more and more people pay attention to it in recent years .However, due to existence of equilibrium constraints, the discussed problem become difficult. So many existing algorithms for nonlinear problems can not be appliedheredirectly. In this paper, two efficient algorithm for MPEC are presented. The achievements canbesummarizedintothefollowingtwoaspects. Firstly, equilibrium problem with linear equilibrium constrains is studied. By using a smoothing complimentarily function which is differentiable everywhere, when smoot