面向多目标的自适应动态概率粒子群优化算法  

Self-adaptive Dynamic Probabilistic Particle Swarm Optimization Algorithm for Multiple Objectives

作　　者：贾兆红[1] 陈华平[1] 唐俊[2] 卢冰原[3] 

机构地区：[1]中国科学技术大学信息管理与决策科学系,合肥230026 [2]安徽大学智能计算与信号处理国家教育部重点实验室,合肥230039 [3]南京工程学院经管学院,南京211167

出　　处：《系统仿真学报》2008年第18期4959-4963,共5页Journal of System Simulation

基　　金：国家自然科学基金(70671096);安徽省自然科学基金(050460404);中国科学技术大学研究生创新基金(KD2006059);安徽省高校自然科学基金重点项目(2006kj013A);安徽省教育厅自然科学基金项目(kj2008B142;kj2008B024)

摘　　要：将基于动态概率搜索的粒子群优化(Particle swarm optimization,PSO)算法应用于多目标作业车间调度问题(Flexible job shop scheduling problem,FJSP),提出一种新算法。算法在搜索初期利用粒子近邻的平均最优代替传统的单个最优引导搜索,后期用Gaussian动态概率搜索来提高算法的局部开挖能力。然后,引入Pareto优的概念,采用精英集来存放非劣解,提出一种新的适应度值分配方法。此外,在算法中还引入了一种自适应的变异算子来增强解的多样性。最后,用新算法对多组FJSP实例进行测试,并与其他几种方法进行比较,结果表明提出的算法具有较好的搜索性能,是求解多目标FJSP的一种可行方法。A novel algorithm was proposed to solve multi-objective flexible job shop scheduling problem （FJSP）, based on dynamic probabilistic particle swarm optimization （PSO）. At the earlier stage, the average of the neighboring best individuals instead of the general single one was employed to guide the search. In the latter stage, Gaussian dynamic probabilistic search was used to improve the local exploiting ability of our method. Moreover, borrowing ideas from Pareto optimization, the non-dominated solutions were stored by using an elitism repository, and a new fitness allocation approach was proposed. Meanwhile, a self-adaptive mutation operator was introduced to enhance the diversity of solutions. Finally, comparative experiments were conducted with several groups of FJSP instances. The experimental results show the feasibility of the algorithm in solving multi-objective FJSP.

关 键 词：多目标 粒子群优化 动态概率 变异 柔性工作车间调度 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]