A digital twin-driven flexible scheduling method in a human–machine collaborative workshop based on hierarchical reinforcement learning

Abstract

Under the influence of the global COVID-19 pandemic, the demand for medical equipment and epidemic prevention materials has increased significantly, but the existing production lines are not flexible and efficient enough to dynamically adapt to market demand. The human–machine collaboration system combines the advantages of humans and machines, and provides feasibility for implementing different manufacturing tasks. With dynamic adjustment of robots and operators in the production line, the flexibility of the human–machine collaborative production line can be further improved. Therefore, a parallel production line is set up as a parallel community, and the digital twin community model of the intelligent workshop is constructed. The fusion and interaction between the production communities enhance the production flexibility of the manufacturing shop. Aiming at the overall production efficiency and load balancing state, a digital twin-driven intra-community process optimization algorithm based on hierarchical reinforcement learning is proposed, and as a key framework to improve the production performance of production communities, which is used to optimize the proportion of human and machine involvement in work. Finally, taking the assembly process of ventilators as an example, it is proved that the intelligent scheduling strategy proposed in this paper shows stronger adjustment ability in response to dynamic demand as well as production line changes.