A cooperative water wave optimization algorithm with reinforcement learning for the distributed assembly no-idle flowshop scheduling problem

The distributed assembly no-idle flow-shop scheduling problem (DANIFSP) is a novel and considerable model, which is suitable for the modern supply chains and manufacturing systems. In this study, a cooperative water wave optimization algorithm, named CWWO, is proposed to address the DANIFSP with the goal of minimizing the maximum assembly completion time. In the propagation phase, a reinforcement learning mechanism based on the framework of the VNS is designed to balance the exploration and exploitation of the CWWO algorithm. Afterwards, the path-relinking combined with the VNS method as the modified breaking operator is introduced to enhance the capability of local search. Furthermore, a multi-neighborhood perturbation strategy in the refraction phase is applied to extract knowledge information to increase the probability of escaping the local optimal. Moreover, the comprehensive experimental program is executed to calibrate the control parameters of the CWWO algorithm and illustrate the cooperative effect of the three modified operations. The performance of the CWWO algorithm is verified on the benchmark set, and the experimental results demonstrated the stability and validity of the CWWO algorithm.

Automated summary

This study introduces a cooperative water wave optimization algorithm CWWO to solve the distributed assembly no-idle flow-shop scheduling problem DANIFSP, aiming to minimize the maximum assembly completion time. The algorithm uses a reinforcement learning mechanism in the propagation phase, introduces a combination of path-relinking and VNS method as the modified breaking operator in the improvement phase, and applies a multi-neighborhood perturbation strategy in the refraction phase to increase the probability of escaping the local optimal.