Timed Petri net modeling and optimization with heuristic search for flexible manufacturing workstations

This paper presents an approach to model and optimize the activities of a manufacturing workstation based on a Petri structure. To model the workstation activities, a comprehensive hierarchical task decomposition consistent with manufacturing control requirements is followed. We also introduce the concept of time-state equations for tracking and optimization purposes. In order to optimize the flow of parts through the physical workstation, a method that combines the heuristic A* search algorithm with limited marking expansion, a new heuristic function and a "non-delay" branching scheme is presented. The output of the algorithm is a near-optimal solution to the part scheduling problem under the makespan criterion. A comparison against previous work shows promising results.