求解优化问题的混合PSO-Solver算法  

Hybrid PSO-Solver algorithm for solving optimization problems

作　　者：高艳卉[1] 诸克军[1] 

机构地区：[1]中国地质大学(武汉)经济管理学院,武汉430074

出　　处：《计算机应用》2011年第6期1648-1651,共4页journal of Computer Applications

基　　金：中央高校基本科研业务费专项资金资助项目(CUG090113);中国博士后基金资助项目(20090461293)

摘　　要：融合了粒子群算法(PSO)和Solver加载宏,形成混合PSO-Solver算法进行优化问题的求解。PSO作为全局搜索算法首先给出问题的全局可行解,Solver则是基于梯度信息的局部搜索工具,对粒子群算法得出的解再进行改进,二者互相结合,既加快了全局搜索的速度,又有效地避免了陷入局部最优。算法用VBA语言进行编程,简单且易于实现。通过对无约束优化问题和约束优化问题的求解,以及和标准PSO、其他一些混合算法的比较表明,PSO-Solver算法能够有效地提高求解过程的收敛速度和解的精确性。Combined Particle Swarm Optimization（PSO） and Solver add-in,this paper proposed a hybrid PSO-Solver algorithm to solve the optimization problems.As a global search algorithm,PSO looks for the global feasible solution,and Solver is a local search tool based on gradient information,which refines the solution obtained by PSO.The hybrid algorithm could speed up the global search,as well as avoid getting into local minima.VBA was used to code,which is simple and easily conducted.Results of solving some unconstrained and constrained examples,compared to the standard PSO and other heuristic algorithms,show that this hybrid PSO-Solver algorithm can improve the speed of convergence and the accuracy of solutions significantly.

关 键 词：粒子群算法 SOLVER VISUAL Basic应用程序 优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]