A Large-Scale Multi-Objective Flexible Job-Shop Scheduling Algorithm and Its Application Research

This article proposes an improved NSGA-II algorithm for optimizing high-dimensional multi-objective job-shop scheduling problems. NSGA-II algorithm is widely used in multi-objective optimization problems. However, due to the drawbacks of NSGA-II algorithm, such as being prone to falling into local optima, long solving time, low population diversity, high computational cost, and poor performance in solving large-scale multi-objective optimization problems. Therefore, this paper presents a large-scale multi-objective flexible job-shop scheduling algorithm named Improved NSGA-II. The algorithm combines two initialization strategies, maximum completion time and total energy consumption, and variable neighborhood search strategies to schedule operation sorting and machine allocation. Moreover, this method is not only applicable to flexible job-shop problem that consider transportation waiting time between different processes, but also has good performance for flow job-shop with tight process connections and no transportation waiting time. Through simulation experiments based on MK01MK10 and actual production data, it is shown that this algorithm significantly improves its ability to escape local optima, greatly improves the quality of scheduling schemes, and reduces the maximum completion time and total energy consumption. Furthermore, this method has better solutions and more Pareto frontier solution sets than traditional NSGA-II algorithm in largescale scheduling problems, which means that a variety of scheduling schemes can be selected. It has important decision-making significance for the practical intelligent manufacturing scenarios.