基于遗传算法的可重入半导体生产线的调度  

Scheduling Re-entrant Manufacturing Lines of Semiconductor Based on GA

作　　者：王永[1] 吴智铭[1] 隋义[1] 

机构地区：[1]上海交通大学自动化系,上海200240

出　　处：《计算机仿真》2007年第12期247-251,共5页Computer Simulation

基　　金：国家自然基金(65074049)

摘　　要：半导体生产线的可重入性使得半导体生产车间的调度成为NP难题。基于遗传算法的调度是将遗传算法应用到半导体生产线的投料策略和调度策略上,通过合理的编码来寻找一组投料策略和调度策略的组合,来解决可重入半导体生产线的调度问题。和以往的方法不同,这种方法可以从全局上来调度半导体生产线,充分发挥出各种策略的优势。仿真结果表明这种方法对于产出量、制造周期偏差、瓶颈设备利用率等关键生产指标都有明显改善。Manufacturing line of semiconductor has the property of the re-entrant.It＇s NP hard to schedule the re-entrant lines for its complexity.Scheduling re-entrant manufacturing lines of semiconductor based on GA is a method of using GA in release rules and dispatching rules to find an optimal combination.Different with the existing methods,this method can schedule the whole manufacturing line and make good use of every rule.The simulation result proves that this method can improve many important criteria,such as Mean Cycle Time,Variance of Cycle Time,Mean Throughput and so on.

关 键 词：半导体 遗传算法 调度方法 投料控制 仿真 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]