Multiple-objective optimization of a reconfigurable assembly system via equipment selection and sequence planning

Abstract

With the development of society and technology, the modern manufacturing system is required to handle unpredictable and rapidly changing market demands during various demand periods. This is a new challenge to the production scheduling of the production line, which requires a faster and smarter response to changes in production demand. To meet the challenge, the reconfigurable manufacturing system (RMS) has emerged as a competitive and promising manufacturing model. The reconfigurable assembly system (RAS), a critical feature of RMS, can be seamlessly incorporated into multi-variety, variable-batch, multi-functional, and fast-delivery manufacturing modes, opening up the possibility of rapid changes for production scheduling. In this paper, we first build a RAS model that can more accurately represent scheduling problems that may occur in real-world industrial operations. More specifically, we consider the assembly process of products could be divided into a number of major steps which are performed in different stations and optimize the assembly sequence of products and the equipment selection within individual assembly stations. After that, three objective functions corresponding to the workload balance, reconfiguration/assembly cost, and lead time are proposed to jointly optimize the assembly sequence and the equipment selection in individual assembly stations. Finally, a hybrid particle swarm optimization (MOHPSO) method is utilized to solve the multiple-objective optimization problem, which combines the advantages of particle swarm optimization and the hybrid method. Experimental results show the superiority of the proposed MOHPSO method over other alternatives including MOPSO and NSGA-II, achieving lower Spacing Metric (SM) and Mean Ideal Distance (MID) indexes and is suitable for solving production scheduling problems.