An expert system for scheduling in a flexible manufacturing system

Abstract The use of artificial intelligence techniques has been investigated as a contribution to a solution to the following problems: daily estimated scheduling in a FMS, and real time control of production disturbances. We present our investigation in object oriented languages to represent the relevant information of a scheduling problem, particularly knowledge about constraints and flexibility factors in a shop. We detail the off-line daily scheduling prototype SOJA and the reactive capabilities we are incorporating in it to repair previous schedules with respect to real-time situations. The integration of automatic generation and modification of schedule plans leads us to introduce a coordination level which generates an admissible solution and a local, cooperation level which deals with local modifications of the plan.