异常事件监控驱动的动态再调度研究
|
摘要
针对实际生产中由于各种异常事件导致初始调度计划不能适应动态生产环境的问题,提出了一种基于异常事件监控的动态再调度优化策略。建立异常事件监控系统与动态再调度之间的关联框架,通过监控获知异常事件的发生并以此来驱动再调度。再调度是运用滚动窗口技术把动态变化的过程转换为连续的静态调度区间,然后用Pareto的多目标遗传算法对每个区间进行调度优化。根据异常事件的不同,设计了相应再调度机制(滚动窗口的更新、算法参数的初始化),又根据动态调度问题的特性,设计了有效的多目标选择操作和遗传操作,最后通过实例验证这种动态调度的有效性和可行性。
For the problem that initial scheduling plan can not adapt to the dynamic production environment due to various kinds of abnormal events,a dynamic re-scheduling optimization strategy based on abnormal event monitoring was proposed. Abnormal events monitoring system associated with the dynamic scheduling framework wad established,the occurrence of abnormal events was informed by monitoring in order to drive re-scheduling again. The re-scheduling used a rolling window technique converts dynamic process into continuous static scheduling interval,then the pareto multi-objective genetic algorithm were applied to optimize each interval. According to the different abnormal events,a corresponding scheduling mechanism( rolling windows update and the algorithm parameters initialization) was designed. According to the characteristics of dynamic scheduling problem,an effective multi-objective selection operation and genetic manipulation was designed,finally,this dynamic re-scheduling effectiveness and feasibility was proved by an example.
For the problem that initial scheduling plan can not adapt to the dynamic production environment due to various kinds of abnormal events,a dynamic re-scheduling optimization strategy based on abnormal event monitoring was proposed. Abnormal events monitoring system associated with the dynamic scheduling framework wad established,the occurrence of abnormal events was informed by monitoring in order to drive re-scheduling again. The re-scheduling used a rolling window technique converts dynamic process into continuous static scheduling interval,then the pareto multi-objective genetic algorithm were applied to optimize each interval. According to the different abnormal events,a corresponding scheduling mechanism( rolling windows update and the algorithm parameters initialization) was designed. According to the characteristics of dynamic scheduling problem,an effective multi-objective selection operation and genetic manipulation was designed,finally,this dynamic re-scheduling effectiveness and feasibility was proved by an example.
引文
[1]JACKSON J R.Simulation research on job shop production[J].Naval Res Log Quart,1957,4(3):287-295.
[2]钱晓龙,唐立新,刘文新.动态调度的研究方法综述[J].控制与决策,2001,16(2):141-145.
[3]包琳.基于事件驱动的动态调度研究[D].济南:山东大学,2010.
[4]王建,杨建军,杨达玲.基于约束满足的Job-shop车间生产再调度系统[J].现代制造工程,2006(5):16-18.
[5]HOLLOWAY C A,NELSON R T.Job shop scheduling with due datesand variable processing times[J].Management Science,1974,20(9):1264-1275.
[6]DEB K,SRINIVAS N.Multiobjective Optimization Using Nondominated Sorting in Genetic Algorithms[J].Evolutionary Computation,2000,2(3):221-248.
[7]王万良,王磊,王海燕,等.基于混合差分进化算法的作业车间动态调度[J].计算机集成制造系统,2012,18(3):531-539.
[8]李丹,张宏国.资源约束下多项目再调度算法[J].信息技术,2007(4):16-21.
[9]张超勇,董星,王晓娟,等.基于改进非支配排序遗传算法的多目标柔性作业车间调度[J].机械工程学报,2010,46(11):156-164.
[2]钱晓龙,唐立新,刘文新.动态调度的研究方法综述[J].控制与决策,2001,16(2):141-145.
[3]包琳.基于事件驱动的动态调度研究[D].济南:山东大学,2010.
[4]王建,杨建军,杨达玲.基于约束满足的Job-shop车间生产再调度系统[J].现代制造工程,2006(5):16-18.
[5]HOLLOWAY C A,NELSON R T.Job shop scheduling with due datesand variable processing times[J].Management Science,1974,20(9):1264-1275.
[6]DEB K,SRINIVAS N.Multiobjective Optimization Using Nondominated Sorting in Genetic Algorithms[J].Evolutionary Computation,2000,2(3):221-248.
[7]王万良,王磊,王海燕,等.基于混合差分进化算法的作业车间动态调度[J].计算机集成制造系统,2012,18(3):531-539.
[8]李丹,张宏国.资源约束下多项目再调度算法[J].信息技术,2007(4):16-21.
[9]张超勇,董星,王晓娟,等.基于改进非支配排序遗传算法的多目标柔性作业车间调度[J].机械工程学报,2010,46(11):156-164.