An automatic multi-objective evolutionary algorithm for the hybrid flowshop scheduling problem with consistent sublots
Biao Zhang,Quan-ke Pan,Lei-lei Meng,Chao Lu,Jian-hui Mou,Jun-qing Li
DOI: https://doi.org/10.1016/j.knosys.2021.107819
2022-02-01
Abstract:Lot streaming is the most widely used technique to facilitate overlapping of successive operations. Inspired by real-world scenarios, this paper studies a multi-objective hybrid flowshop scheduling problem with consistent sublots, aiming to simultaneously optimize two conflicting objectives: the makespan and total number of sublots. Considering the setup and transportation operations, a multi-objective mixed integer programming model is developed and the trade-off between the two objectives is evaluated. Because of the NP-hard property of the addressed problem, metaheuristics are suggested. It is well known that the performance of metaheuristics is highly dependent on the setting of algorithmic parameters, referred to as numerical and categorical parameters. However, the traditional design process might be biased by previous experience. To eliminate these issues, an automated algorithm design (AAD) methodology is introduced to conceive a multi-objective evolutionary algorithm (MOEA) in a promising framework. The AAD enables designing the algorithm by automatically determining parameters and their combinations with minimal user intervention. With regard to the problem-specific characteristics and the employed algorithm framework, for the categorical parameters, including decomposition, solution encoding and decoding, solution initialization and neighborhood structures, several operators are designed specifically. Along with the numerical parameters, these categorical parameters are determined and combined using the designed iterated racing procedure. Comprehensive computational results demonstrate that the automated MOEA outperforms other state-of-the-art MOEAs for the addressed problem.
computer science, artificial intelligence