Digital-model-driven distributed control for production and planning in smart manufacturing systems
提出一种利用数字工具的分布式生产控制方法，解决多作业车间调度问题，在文献案例和实际场景中验证了其有效性、鲁棒性和可扩展性，优于集中式遗传算法。
With the aim of improving Smart Manufacturing System performance, novel Production Planning and Control (PPC) tools, combining the new Industry 4.0 technologies with traditional scheduling methods, are necessary. Among the different PPC challenges, herein, we focus on Multi-Job-Shop Scheduling Problems (MJSSP), which, due to their high complexity, are critical for ensuring an optimal production management. Planning, organising and assigning a set of multiple jobs on different groups of machines, especially in the presence of limited resources and with predetermined processing time, is an NP-hard problem. The problem becomes challenging when also considering unexpected events altering the system production flow. To address this issue, this work proposes a Distributed Production Control Approach which, by leveraging digital tools, properly manages the MJSSP, both in a literature case study and in a real-world scenario, while ensuring resilience in Smart Manufacturing Systems against unforeseen events. Simulation results confirm the effectiveness, robustness, and scalability in terms of makespan, total flowtime and computational time of the proposed approach, showing advantages compared to the centralised solutions obtained via benchmarking Genetic Algorithm. Most notably, the obtained results highlight the potential of this control to optimise production in complex manufacturing systems, paving the way for a full integration in real-world manufacturing companies.