Abstract The increasing demands for customized products have brought a lot of inevitable operational dynamics to the production logistics system. How to systematically monitor and evaluate its overall real-time operation status, and to invoke the most appropriate decision-making level, use the most economical computational resource, and conduct a synchronized decision-making and control to the most accurate operation scope to address a randomly occurred dynamics has remained a long-term challenge for researchers. This paper proposes a multi-level cloud computing enabled digital twin system for the real-time monitor, decision and control of a synchronized production logistics system. In the IoT-driven production logistics synchronization (PLS) system with complete real-time information, the dynamics that occurred in the physical layer could be accurately and real-timely captured and its negative effects on the system’s overall operation state could be effectively evaluated in the digital layer. For slight, moderate and severe dynamics, edge computing, fog computing and cloud computing and their dynamically formed multi-level distributed decision-making system will be used to deal with the dynamics in the most effective and economical mode. Finally, the PLS optimization model of production and storage is presented with an industrial case, and the effectiveness is also demonstrated and analyzed.

中文翻译：

---

多级计算架构下基于数字孪生的实时生产物流同步系统

摘要 对定制产品日益增长的需求给生产物流系统带来了许多不可避免的操作动态。如何系统地监控和评估其整体实时运行状态，并调用最合适的决策层，使用最经济的计算资源，同步决策和控制到最准确的运行范围，以解决随机发生的动力学一直是研究人员面临的长期挑战。本文提出了一种支持多级云计算的数字孪生系统，用于同步生产物流系统的实时监控、决策和控制。在物联网驱动的具有完整实时信息的生产物流同步（PLS）系统中，物理层发生的动态变化可以被准确、实时地捕获，其对系统整体运行状态的负面影响可以在数字层得到有效评估。对于轻度、中度和重度动态，将使用边缘计算、雾计算和云计算及其动态形成的多级分布式决策系统，以最有效、最经济的方式处理动态。最后，结合工业案例提出了生产和存储的PLS优化模型，并对其有效性进行了论证和分析。雾计算和云计算及其动态形成的多级分布式决策系统将被用于以最有效和最经济的方式处理动态。最后，结合工业案例提出了生产和存储的PLS优化模型，并对其有效性进行了论证和分析。雾计算和云计算及其动态形成的多级分布式决策系统将被用于以最有效和最经济的方式处理动态。最后，结合工业案例提出了生产和存储的PLS优化模型，并对其有效性进行了论证和分析。