基于马尔可夫决策过程的MES系统动态调度方法

针对以事件为驱动的网络化制造环境下生产系统的特征,扩展马尔可夫决策过程,根据生产制造的特点进行动态生产调度方法的研究.以面向用户标准为基础,用制造系统响应时间来评价调度算法在不同规则下具有的性能特点;根据分析结果分别确定各种调度规则在预期时间内能够达到生产优化调度或满意调度的可能程度.将调度规则与实际生产相结合,设计网络化制造环境下动态调度指令单管理系统,实现制造管理的信息化.本系统可以按照用户需求实时反映生产制造信息;根据企业情况确定调度规则,生成调度决策方案,最大程度保证企业利益.

Research and Analysis About Dynamic Job Scheduling on MES System Toward Markov Decision Processes

Aiming at the characteristics of production system in an event-driven networked-manufacturing environment,the Markov decision process is extended such that the dynamic scheduling is available to be studied further in view of these characteristics.Based on the user-oriented criteria,the performance of the scheduling algorithm is appraised according to the response time of manufacture system in which there are some different rules.Then,the possibilities of optimal or satisfactory scheduling decision available to achieve within expected time are determined according to different rules on appraisal basis.Combining these rules with the actualities in production processes,a management system of dynamic scheduling command notices is designed in the networked-manufacturing environment to make the manufacturing management more informative.The system designed can reflect the real-time production information to meet users' demands,and the scheduling rules can be readjusted according to manufacturing practice to offer a scheme for scheduling decision making,thus ensuring the maximized economic benefits.