A dynamic scheduling system for flexible manufacturing systems (FMSs) is developed. The system consists of an artificial intelligence scheduler (AIS), a simulation model, a database, and a user interface. The AIS is used to generate candidate schedules according to specified dispatching rules and to control the search depth based on the status of an FMS. The simulation model is utilized to evaluate candidate schedules. The user interface is used to manually generate candidate schedules and to facilitate communication between different modules so as to realize automatic scheduling. Schedule related data are stored in the database. A dispatching-rule-based approximation search method is employed, which can quickly converge on good areas of the solution space, and the search method is suitable for dynamic scheduling in FMSs. A hypothetic FMS is designed and an sample scheduling problem is used to demonstrate the mechanism of the scheduling system.
Dynamic Scheduling of Flexible Manufacturing Systems with Learning Agents
