Scheduling of Customized Tasks in Cloud Manufacturing with Deep Reinforcement Learning

With the increase of user demand for customized products, it leads to the rapid growth of customized production task orders. The traditional large-scale production mode has been transformed into multi-variety and small-batch production mode. It also introduces more uncertainty into the cloud manufacturing scheduling problem, causing the dynamic environment to become more complex. In order to solve the problem of customized task scheduling under multi-variety and small-batch production mode in cloud manufacturing, a customized constrained task scheduling model for cloud manufacturing was proposed by exploring customized task modeling and scheduling target modeling. Considering the constraints of time, cost and quality, the scheduling objective model is constructed. On this basis, the deep reinforcement learning algorithm Dueling-DQN (DDQN) is used to schedule customized production tasks. The experiment respectively analyzes the impact that customizing tasks under different QoS weights and different constraints when scheduling. The classical deep reinforcement learning algorithm is used as a comparison algorithm to verify the effectiveness of the algorithm in cloud manufacturing customized task scheduling. Experiments show that the model can fully reflect the characteristics of customized task scheduling, and the algorithm can effectively solve the problem of customized task scheduling considering constraints. The algorithm designed in this paper has stronger adaptability and faster convergence speed.