求解二层规划的混合微粒群算法

对于二层规划问题有许多经典的求解方法,如极点搜索法、分支定界法和罚函数法等。文中给出了基于微粒群算法的二层规划的一种新的求解方法。提出了分别先用单纯形法和内部映射牛顿法的子空间置信域法求解下层规划,然后用微粒群算法求解上层规划的求解方法,这两种混合微粒群算法分别用于求解线性二层规划和非线性二层规划。并结合实例的对比分析,说明了这两种混合微粒群算法求解二层规划的可行性和有效性。

Hybrid Particle Swarm Optimization for Bilevel Programming

There are many classical solution methods for bilevel programming such as culmination searching algorithm,branch and bound algorithm and penalty function algorithm.Put forward a new solution algorithm for bilevel programming based on particle swarm optimization.It resolves the lower-programming by simplex method and subspace trust region method based on the interior-reflective Newton method,and resolve the upper-programming by particle swarm optimization.The two hybrid particle swarm optimizations resolve the linear bilevel programming and non-linear bilevel programming independently.With many examples,it is feasible that two hybrid particle swarm optimizations resolve bilevel programming.