基于多目标粒子群算法的混合流水车间调度方法研究  

A Hybrid Flow-Shop Scheduling Approach Based on Multi-Objective Particle Swarm Optimization

作　　者：欧微[1] 邹逢兴[1] 高政[1] 徐晓红[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,湖南长沙410073

出　　处：《计算机工程与科学》2009年第8期52-56,共5页Computer Engineering & Science

基　　金：国家自然科学基金资助项目(60634020)

摘　　要：混合流水车间调度问题HFSP是一种具有很强应用背景的生产调度问题。本文给出了一种HFSP多目标调度模型,提出了一种针对该类问题的多目标粒子群算法。该算法采用基于Pareto支配关系的极值更新策略;采取对自适应惯性权重递减和对种群变异的方法以保持种群多样性;设置Pareto解池保存计算中出现的Pareto最优解,并提出了一种基于适应度拥挤度的聚类算法优化解的分布特性。实验结果表明,本文算法是求解HFSP问题的一种有效方法。The hybrid flow-shop scheduling problem（ HFSP） is a scheduling problem with strong application background. A multi-objective scheduling model of HFSP is put forward in this paper, and a multi-objective particle swarm optimization （MOPSO） algorithm is proposed. This algorithm adopts an updating strategy based on the Pareto dominance relationship, and introduces a self-adaptive inertia weight and mutation approach to keep the diversity of the population, the also designs a Pareto optimal set pool to preserve the dominance solutions found in the evolution process. Then a clustering algorithm based on the fitness crowd degree is presented. Finally, the effectiveness of MOPSO is demonstrated by numerous simulation experiments.

关 键 词：生产调度 混合流水车间 多目标优化 粒子群算法 PARETO最优解 进化计算 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]