A Hybrid Multi-Threaded Parallel Iterated Greedy Algorithm for Distributed Flowshop Group Scheduling Problems with Preventive Maintenance

The distributed flowshop group scheduling abstracted from the production process of printed circuit boards is a recent active research topic. The distributed flow shop group scheduling problem is based on the assumption that processing machines can run continuously. However, in actual production, factories not only pursue productivity, but also pay attention to the reliability and stability of the production process. In this paper, we study a distributed flowshop group scheduling problem with preventive maintenance (DFGSP/PM) to minimize the makespan. Based on the characteristics of DFGSP/PM, we proposed a hybrid multi-threaded parallel iterated greedy algorithm (HMPIG). The greedy NEH algorithm based on the LPT rule is used in the initialization phase of the algorithm to generate the initial solution. Different destruction and reconstruction rules for groups and jobs are designed to update the solution in the destruction and reconstruction phase. In the local search phase, a hybrid multithreaded parallel search strategy is introduced to improve the search efficiency of the IG algorithm so that the optimal insertion position of the group in the sequence can be searched more quickly. The numerous experimental results show that HMPIG has better solution quality as well as solution stability.