A multi-agent system for FJSP with setup and transportation times

Abstract

Flexible job shops are quite common in manufacturing, where the machines can perform a variety of operations on a job. The setup and transportation times are quite significant in flexible job shops, which recently have attracted attention. This paper deals with the problem of scheduling of flexible job shops, termed FJSP, with setup and transportation times. Several metaheuristics approaches have been proposed for the problem assuming centralized decision-making. The centralized approaches, however, may not be suitable for large complex problems due to higher computational effort and slower convergence. The decentralized approaches are more suitable for such problems. To the best of authors’ knowledge, no work has been reported towards developing a decentralized approach for FJSP with setup and transportation times. To fill this gap a multi-agent system, a popular decentralized approach, is developed for FJSP with setup and transportation time. The performance of the proposed approach is compared with the three state-of-the-art (centralized) approaches by solving 20 problem instances. The proposed approach has been found to produce schedules with makespan lower (maximum of 35% and an average of 10.27%) than the comparison approaches. The algorithm has also been found to converge faster than the comparison algorithms.