等待时间受限的并行批处理调度问题研究  

A Heuristic Algorithm for Parallel Batch Processing Scheduling with Limited Waiting Time Constraints

作　　者：刘志[1] 董明[1] 李永林[1] 

机构地区：[1]上海交通大学安泰经济与管理学院,上海200030

出　　处：《工业工程与管理》2014年第6期117-123,共7页Industrial Engineering and Management

基　　金：教育部博士点基金资助项目(20120073110029);国家自然科学基金资助项目(71371123);国家自然科学基金重点项目(71131005);上海交通大学文理交叉重点项目(11JCZ02);Europe-China High Value Engineering Network(EC-HVEN:295130)

摘　　要：研究了任务等待时间受限的并行批处理调度问题,建立它的约束规划模型,设计调度任务的"小s大S"规则,该调度规则最重要的参数为临界队长参数和考虑受限时间的紧急时间参数。基于该规则进行模拟实验,用RSM响应面方法进行参数优化,将模拟结果同约束规划模型(CP方法)在CPLEX环境下求得的解进行比较,分析"小s大S"规则的性能和效果。A parallel batch processing scheduling problem with limited waiting time constraints is studied to minimize the total weighted processing time. This paper builds the constraint programming（CP）model and designs the（s, S）dispatching rule derived from inventory management to solve the batch processing problem. There are two important parameters, queue length threshold and urgent time thresholds, in this rule and coefficients a and ~ respectively determine the values of the two parameters. Then, simulation experiments based on the（s, S）rule are conducted and response surface method（RSM）is utilized to optimize the two coefficients in（s, S）rule. After optimizing the parameters, many cases are used to compare the results of （s, S） dispatching rule with results of CP method in CPLEX environment and study the property and value of（s, S） rule.

关 键 词：并行机 批处理 等待时间受限 半导体制造 小s大S规则 

分 类 号：F253[经济管理—国民经济]