An Adaptive Multi-Objective Evolutionary Algorithm with Two-Stage Local Search for Flexible Job-Shop Scheduling

• Published in: International journal of computational intelligence systems Vol. 14; no. 1; p. 54
• Main Authors: Li, Yingli, Wang, Jiahai, Liu, Zhengwei
• Format: Journal Article
• Language: English
• Published: Springer 01.01.2021
• Subjects: Evolutionary algorithm; Flexible job-shop scheduling; Local search; Multi-objective optimization
• ISSN: 1875-6883; 1875-6883
• DOI: 10.2991/ijcis.d.201104.001

An adaptive evolutionary algorithm with two-stage local search is proposed to solve the multi-objective flexible job-shop scheduling problem (MOFJSP). Adaptivity and efficient solving ability are the two main features. An autonomous selection mechanism of crossover operator is designed, which divides individuals into different levels and selects the appropriate one according to the both sides' levels to improve the self-adaptation. In parameter setting, the autonomous determination and adjustment mechanism is proposed, and parameters are adjusted autonomously according to the job scale and iteration number, so as to reduce the complexity of parameter setting and further improve the adaptivity. For improving solving ability, two-stage local search mechanism is designed. The first stage is performed before the evolution operation, so that each individual has more good genes to participate in the following operation. The second stage is performed after the evolution operation to further search the optimal solutions. Finally, a large number of comparative numerical tests are carried out, compared with other excellent algorithms, the proposed algorithm has fewer parameters to be set and stronger solving ability.