Extending an agent-based FMS scheduling approach with parallel genetic algorithms

Flexible manufacturing systems (FMS) aim at efficiently reacting to changing market needs to stand the increasing competitiveness. This imposes efficiency and flexibility requirements on FMS scheduling. Manufacturing scheduling is the process of allocating available manufacturing resources to the set of planned jobs over time. It is an optimization process by which limited manufacturing resources are to be allocated to several jobs of different products efficiently. The agent-based scheduling approach has shown the ability to fulfill the flexibility requirement. Although this approach emphasizes flexibility, it lacks the optimization support. In this paper, an agent-based scheduling approach is extended with parallel genetic algorithms (PGA) to provide the required optimization support. Test results have shown a remarkable enhancement to the optimality of the generated schedules with respect to the predefined set of manufacturing objectives. The extended approach fulfils both flexibility and efficiency requirements on manufacturing scheduling.