Geometric approach and taboo search for scheduling flexible manufacturing systems

This paper addresses the scheduling and deadlock avoidance of a class of automated manufacturing systems. In such systems, a set of jobs is to be performed on a set of resources and each job requires several operations. An operation may require several types of resources with several units of each type. Further, upon the completion of an operation, its related resources cannot be released until resources needed for the next operation become available. One important characteristic of such systems is the possibility of deadlock. The scheduling problem deals with the allocation of resources such that jobs are completed within a minimal makespan and deadlocks are avoided. We extend the classical geometric approach to solve the two-job case of our model. A greedy algorithm based on this result and the taboo search heuristic are then developed for the general case. Numerical results show that the proposed algorithm is fast and provides good schedules.