An improved multi-agent particle swarm optimization to solve flexible job-shop scheduling problem.
Résumé
Many efficient meta-heuristics methods are developed to solve Flexible Job Scheduling Problem (FJSP) to get nearly optimal solutions that optimize an objective function. However, in real-world manufacturing systems, schedules are often confronted with unexpected factors such as random machine breakdown. Therefore, centralized approaches are in general inflexible, expensive, and slow to satisfy scheduling problems under disruptions. For a better efficiency targeting such situations, distributed approaches based on Multi-Agent System methods and using metaheuristics have attracted more attention. In a previous work, we have proposed a Multi-Agent Particle swarm optimization named MAPSO2 to solve FJSP. Despite the promising results obtained in the proposed model in term of solution quality, some limitations have been detected. In this paper, we present an improved MAPSO2 named to solve FJSP with better flexibility.