Mixed-production flexible assembly job shop scheduling considering parallel assembly sequence variations under dual-resource constraints using multi-objective hybrid memetic algorithm

Abstract

In this study, a mixed-production flexible assembly job shop scheduling considering parallel assembly sequence variations under dual-resource constraints (MFAJSS-PASV-DRC) is proposed, to achieve simultaneous optimization of part processing sequence and assembly sequence in mixed-production. By analyzing the MFAJSS-PASV-DRC problem, an integrated mathematical model that considers the interactive effects between part processing sequence and assembly sequence under dual-resource constraints in mixed-production is established, with the optimization objectives to minimize the total production completion time, the total inventory time, and the total labor cost during the production process. Based on the above, a multi-objective hybrid memetic algorithm (MoHMA) is proposed to solve the MFAJSS-PASV-DRC. In MoHMA, a four-layer segmented hybrid chromosome encoding structure is designed, then a mixed initialization strategy (MIX3) is applied to obtain a population of high quality, and two evolutionary methods are used to generate offspring chromosomes. Meanwhile, a variable neighborhood search (VNS) incorporating five local search methods is designed to prevent MoHMA from being stuck into a local optimum. The effectiveness of MIX3 and VNS are verified by an ablation experiment. Then an elite retention strategy is used to improve the quality of non-dominated solutions. In the case study, the Taguchi method is applied to obtain the best combination of parameters for the MoHMA algorithm. After that, the superiority of the proposed MFAJSS-PASV-DRC in improving production efficiency is verified, and the effectiveness of MoHMA is verified in solving the MFAJSS-PASV-DRC problem with different scales by comparing with other algorithms.