信息物理系统环境下面向多层次资源协同的制造智能方法

Real Time Manufacturing Intelligence Method Oriented Multilevel Resource Collaboration in Cyber Physical System Environment

信息物理系统(Cyber-physical system,CPS)在制造过程中的应用,能够将企业不同层次的制造资源连接起来,实现资源的按需使用和高效的过程监控。在CPS环境下,可通过整合多粒度社会化制造资源,构建基于动态服务联盟的灵活多Agent制造系统。随着系统的运营和联盟演化,实时制造智能方法成为解决任务分配、实时决策等问题的关键技术。在对多Agent制造系统服务能力和协同关系进行分析的基础上,提出了动态的协同任务分配模型。考虑制造任务的关系、时空等因素,构建了基于Petri-net和多领域矩阵的协同任务节拍控制模型。根据智能服务单元的特点,提出了多层次实时决策体系,分析了本地Agent的决策模型和不同企业或同一企业多Agent的协同决策机制。最后,在对实时制造智能一般框架进行阐述的基础上,提出了制造智能方法的未来研究方向和挑战。

Cyber--physical system（CPS） applied in the manufacturing process,can connect different levels of the enterprises＇ manufacturing resources,and achieve high efficient process monitoring. A flexible manufacturing system based on dynamie service alliance is built by integrating multiple granularity resources. Many problems appear in the operation stage, such as task allocation,real time decision. Hence,it is urgent to study real time manufacturing intelligence method to improve the performance of service alliance. According to the service capacity and collaboration relations,a dynamic collaborative task allocation model is firstly proposed. Considering the task correlations, time,physical location and related influencing factors, we construct a collaborative task rhythm control model based on Petri--net and MDM to enhance collaborative productivity. Moreover, a multilevel real time decision making system, describing the decision model of local agent and collaboration decision in the same enterprise or across different enterprises,is presented to adapt the state change of task,object,human and environment. The generalized framework and illustration of manufacturing intelligence in real time manufacturing system is given here, future challenges in the field should be widely discussed and further researched.