Abstract: Owing to lack of accurate and real-time data for the service-oriented intelligent manufacturing and few mechanisms exist to tackle the dynamic and collaborative decision-making,a dynamic decision-making method of manufacturing resource based on digital twin was proposed.The framework of dynamic decision-making of manufacturing resource based on digital twin was established.Moreover,device digital twin including Cloud-Deployed Digital Twin and Edge-Deployed Digital Twin and product digital twin were proposed to provide the real-time and accurate decision-making data.A multi-party collaborative mechanism based on real-time multi-agent was designed,and the hierarchical timed colored Petri Net was employed to model and analyze some complex manufacturing tasks.Furthermore,the rank correlation analysis was introduced to obtain the comprehensive weight,and the reference ideal method was introduced to perform a comprehensive evaluation of manufacturing resource.The proposed approach of dynamic decision-making was applied to some aero-engine blade manufacturing factory.The feasibility and validity of the proposed decision-making method were verified by the experimental decision results.

Key words: dynamic selection, multi-criteria decision making, digital twin, multi-Agent, reference ideal method, Petri nets

摘要： 针对面向服务的智能制造环境下决策数据的准确性和实时性不足,以及协同和动态决策机制欠缺等问题,提出一种基于数字孪生的制造资源动态优选决策方法。设计了基于数字孪生的制造资源动态优选决策架构,并提出设备数字孪生(云/边数字孪生)和产品数字孪生,为决策者提供准确、实时的评价数据;构建基于实时Multi-Agent的资源优选决策多方协同机制,采用层次赋时有色Petri网仿真分析复杂制造任务中的瓶颈设备资源;采用序关系分析法计算评价指标权重,引入参考理想法对候选制造资源进行综合评估。将该方法应用于某航空发动机叶片制造企业设备资源优选中,验证了所提方法的可行性和有效性。

关键词: 动态优选, 多准则决策, 数字孪生, 多代理, 参考理想法, Petri网