A teaching-learning-based optimization with feedback for L-R fuzzy flexible assembly job shop scheduling problem with batch splitting

Mengke Sun,Zongyan Cai,Haonan Zhang

DOI: https://doi.org/10.1016/j.eswa.2023.120043

IF: 8.5

2023-04-13

Expert Systems with Applications

Abstract:Batch splitting is an effective means to improve production efficiency. This paper integrates batch splitting into flexible assembly job shop scheduling problem. Complex manufacturing flexible system has high uncertainty. A L-R fuzzy number with nonlinear membership function is used to describe the time uncertainty of scheduling problem. Therefore, this paper proposes the L-R fuzzy flexible assembly job shop scheduling problem with batch splitting (BFFAJSP). BFFAJSP is a two-stage scheduling problem, namely the processing stage and the assembly stage. A multi-objective optimization model is formulated to optimize makespan, energy consumption and quality simultaneously. A teaching-learning-based optimization with feedback (FTLBO) is proposed to solve the proposed BFFAJSP. First, three initialization heuristics are designed to improve the quality and diversity of solutions. Next, the classes are divided based on non-dominated sorting and crowding distance. Then, the quality of the solution is improved based on feedback teaching. In addition, the communication between classes further improves the convergence speed of the algorithm. Finally, in the self-learning stage of teachers, the integrated five local search strategies improve the exploration and exploitation ability of the algorithm. The results of extensive numerical tests show the effectiveness of the proposed algorithm for solving BFFAJSP.

computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science