基于多智能体的工业数字孪生系统云边端架构与关键技术  

Cloud-edge-device system architecture and key technologies of industrial digital twin system based on multi-agent

作　　者：李浩 邢志远 李琳利 鲁效平 杨文超 王昊琪 刘根 王晓丛 刘晓军[4] 文笑雨 武春龙 邢宏文 董李扬 王志强[7] LI Hao;XING Zhiyuan;LI Linli;LU Xiaoping;YANG Wenchao;WANG Haoqi;LIU Gen;WANG Xiaocong;LIU Xiaojun;WEN Xiaoyu;WU Chunlong;XING Hongwen;DONG Liyang;WANG Zhiqiang(Henan Provincial Key Laboratory of Intelligent Manufacturing of Mechanical Equipment,Zhengzhou University of Light Industry,Zhengzhou 450002,China;State Key Laboratory of Massive Personalized Customization System and Technology,Qingdao 266101,China;School of Mechanical Engineering,Dalian University of Technology,Dalian 116000,China;School of Mechanical Engineering,Southeast University,Nanjing 210096,China;National Engineering Research Center for Technological Innovation Method and Tool,Hebei University of Technology,Tianjin 300401,China;Shanghai Aircraft Manufacturing Limited Company,Shanghai 200436,China;High-Tech Standardization Research Institute,China National Institute of Standardization,Beijing 100191,China)

机构地区：[1]郑州轻工业大学河南省机械装备智能制造重点实验室,河南郑州450002 [2]大规模个性化定制系统与技术全国重点实验室,山东青岛266101 [3]大连理工大学机械工程学院,辽宁大连116000 [4]东南大学机械工程学院,江苏南京210096 [5]河北工业大学国家技术创新方法与实施工具工程技术研究中心,天津300401 [6]上海飞机制造有限公司,上海200436 [7]中国标准化研究院高新技术标准化研究所,北京100191

出　　处：《计算机集成制造系统》2024年第11期3755-3770,共16页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金面上项目(52175256);河南省科技研发计划联合基金资助项目(225200810029);大规模个性化定制系统与技术全国重点实验室开放课题(H&C-MPC-2023-02-02)。

摘　　要：基于多智能体的工业数字孪生系统(iDTS)由人、设备、环境、孪生仿真系统和计算系统等多智能体构成,是一个多智能体、AI算法和数字孪生融合驱动的新一代iDTS系统。针对工业数字孪生系统在实时性、智能性和负载均衡等方面的需求,提出了基于多智能体的iDTS系统的典型特征、云边端架构和关键技术体系,主要包括模型驱动智能体自学习优化方法、多智能体交互的同步与共享方法、多智能体协同的分布式控制方法、iDTS系统制造资源协同调度方法和iDTS系统的云边端负载均衡方法等。最后,通过典型案例展示了基于多智能体的iDTS系统的应用效果。所提出的系统架构与关键技术为推动人工智能和数字孪生技术在智能制造过程中的深入应用提供了重要参考。The multi-agent-based industrial Digital Twin System(iDTS)is composed of multiple agents,including humans,machines,environments,digital twin simulation systems and computing systems.It is a new generation of iDTS system driven by the integration of multiple agents,Artificial Intelligence(AI)algorithms and digital twin systems.In view of the real-time,intelligence and load balancing needs of the iDTS system,the typical characteristics,cloud-edge-device system architecture and key technologies of the iDTS system were presented based on multiple agents,which mainly included the methods of agents'self-learning and optimization driven by Large Language Models(LLMs),the methods of clock synchronization and data sharing for the interaction of iDTS system based on multiple agents,the distributed control methods for the collaboration of multiple agents in the iDTS system,the collaborative scheduling methods for multiple manufacturing resources in the iDTS system and the load balancing methods for the iDTS system with a cloud-edge-device architecture.Finally,the application effect of multi-agent-based iDTS system was demonstrated through a typical case.The proposed system architecture and key technologies provide important reference for the in-depth application of AI and digital twin technology in intelligent manufacturing processes.

关 键 词：工业数字孪生 多智能体 人工智能 云边端协同 分布式计算 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]