A Study on Multi-Objective Particle Swarm Optimization in Solving Job-Shop Scheduling Problems

Abstract

Particle Swarm Optimization (PSO) is a population-based metaheuristic that was modelled based on the social interaction and communication of organisms, such as a flock of birds or a school of fishes. It is widely applied to solve a single-objective function in existing research, but this is not suitable for cases in the real world, which normally consist of multiple-objective criteria. Such cases encompass the Job-shop Scheduling Problem (JSP), where it is a typical production scheduling problem and belongs to one of the most difficult problems of combinatorial optimization. Subsequently, the multi-objective Particle Swarm Optimization (MOPSO) was established to accommodate the requirement of multiple-objective cases encountered in real-world production systems. Nevertheless, research works on solving JSP with multiple objectives using MOPSO are still limited compared to the single objective. In this study, comparison and discussion of existing works, in terms of objective functions, test problems, multi-objective optimization methods, scheduling constraints, strategies and performances are conducted. This study also highlights current MOPSO improvement strategies and the aims of their implementation in solving JSP. Finally, this study proposes a MOPSO model in solving JSP that consolidates these aspects of improvement strategies, which would set the path for future directions of research provided in the final part of the paper.