A decomposition-based multi-objective evolutionary algorithm for hybrid flowshop rescheduling problem with consistent sublots
Biao Zhang,Quan-ke Pan,Lei-lei Meng,Xin-li Zhang,Xu-chu Jiang
DOI: https://doi.org/10.1080/00207543.2022.2093680
IF: 9.018
2022-07-06
International Journal of Production Research
Abstract:Lot streaming is the most widely used technique to facilitate the overlap of successive operations. Considering the consistent sublots and machine breakdown, this study investigates the multi-objective hybrid flowshop rescheduling problem with consistent sublots (MOHFRP_CS), which aims at optimising the total completion time, starting time deviations of operations, and average adjustment of sublot sizes simultaneously. By introducing the decomposition strategy and effective migrating birds optimisation framework, this paper develops a multi-objective migrating birds optimisation algorithm based on decomposition (MMBO/D). In MMBO/D, the problem is decomposed into a series of sub-problems, and its solutions are initialised by the Glover operator and further optimised by the variable neighbourhood descent strategy. The weights assigned to the sub-problems are adapted dynamically according to a variable weight strategy, and a global update strategy is employed to update the solutions. A novel sharing and benefiting mechanism is proposed to implement coevolution among different sub-problems. Competitive mechanisms are modified by considering similar sub-problems to improve population quality. A criterion is designed to check whether a subproblem is stuck in the local optima. The comprehensive computational results demonstrate that MMBO/D outperforms other state-of-the-art multi-objective evolutionary algorithms (MOEAs) for the addressed problem.
engineering, manufacturing, industrial,operations research & management science