Abstract
In this paper, we propose an application of Petri Net decomposition approach for the simultaneous optimization of dispatching and conflict-free routing for automated guided vehicles in the dynamic situation where transport requests are given in real time. The objective is to maximize the total throughput of the AGV transport system during the time horizon. In order to solve the dynamic problem, static problems are periodically solved when the transport requests are given to the AGV system. The dispatching and conflict-free routing are simultaneously optimized by the Petri Net decomposition approach. In the proposed method, the Petri Net is decomposed into several subnets for task subproblems and AGV subproblems that can be solved by the shortest path algorithm on the reachability graph. The local solutions for the subnets are coordinated by a penalty function algorithm. To ensure the generation of conflict-free routing, a new deadlock avoidance strategy is incorporated in the optimization algorithm. The effects of simultaneous optimization of dispatching and conflict-free routing are investigated for routing problems in dynamic environments.
