A multi-agent system simulation based approach for collision avoidance in integrated Job-Shop Scheduling Problem with transportation tasks

avril 2023
Ingénierie & Outils numériques
Articles dans des revues internationales ou nationales avec comité de lecture
Auteurs : Kader Sanogo (LINEACT), Abdelkader Mekhalef Behafssa (LINEACT), M'hammed Sahnoun (LINEACT), Belgacem Bettayeb (LINEACT), Abderrahim Moussa (University of Relizane), Abdelghani Bekrar (LAMIH)
Journal : Journal of Manufacturing Systems, 31 mars 2023

At the operational level of Flexible Manufacturing Systems (FMS), several research works have developed solutions not fully realizable considering the operational constraints, which can be difficult to model. One of the well-known problems in FMS is the Job-Shop Scheduling Problem (JSSP) with transportation tasks. Indeed, this problem is addressed by many researchers, but most of their studies do not consider operational constraints such as collision avoidance between transporters. Moreover, optimized scheduling may contain deadlock situations due to often neglected geometric constraints. Therefore, in this paper, we propose a simulation approach based on a multi-agent system to test the optimized results known in the literature in more realistic conditions, where collision avoidance between transporters is considered. We propose as well a sim-optimization approach to explore suitable solutions considering collision avoidance constraints. In addition, an algorithm is proposed to solve deadlocks. The obtained results highlight the impact of collision avoidance on the performance of the literature solutions while being more realistic. Furthermore, these results demonstrate the efficiency of the proposed algorithms, since all identified collisions are avoided and all deadlocks are also solved. The proposed approach is expected to be more attractive to the industries as it ensures a small gap between the theoretical and actual results.