Sustainable Scheduling of Distributed Flow Shop Group: A Collaborative Multi-Objective Evolutionary Algorithm Driven by Indicators
Yuhang Wang,Yuyan Han,Yuting Wang,Quan-ke Pan,Ling Wang
DOI: https://doi.org/10.1109/tevc.2023.3339558
IF: 16.497
2023-01-01
IEEE Transactions on Evolutionary Computation
Abstract:Sustainable scheduling within the manufacturing field has garnered substantial attention from both academia and industry. The escalating market demands have heightened requirements on the flexibility of production modes, multi-zone, and multi-objective. In this context, our study explores the intricacies of the multi-objective distributed flow shop group scheduling problem with sequence-dependent setup times, aiming to concurrently optimize makespan and total energy consumption (DFm|group, sdst|#(Cmax, TEC) ). Firstly, a mathematical model is constructed to analyze problem characteristics. Subsequently, we introduce a collaborative multi-objective evolutionary algorithm driven by indicators (CMOEA/I). In CMOEA/I, an indicator-driven approach is proposed for solution selection, which approximates the Pareto front based on the convergence indicator, while screening potential solutions based on the spread indicator. Furthermore, a collaborative model and local search are developed by incorporating the intrinsic linkages of factories, groups, and jobs. Additionally, to further explore the potential non-dominated solutions, a speed variation strategy is devised based on the pivots of decreasing speed to save energy and increasing speed to reduce makespan. An extensive set of simulation experiments is conducted on a diverse range of test instances. Through meticulous statistical analysis, the outcomes demonstrate that the CMOEA/I exhibits efficacy when contrasted with other advanced algorithms.
computer science, artificial intelligence, theory & methods