Supervisor Synthesis and Performance Improvement for Automated Manufacturing Systems by Using Petri Nets

For event-driven systems like automated manufacturing systems (AMSs), intelligent part dispatching, i.e., to assign various parts to proper processing routes, is crucial to gain high resource utilization and production efficiency. This work proposes an innovative dispatching mechanism, which can better these metrics by minimizing the cycle time, i.e., the mean time to produce a product, while avoiding any deadlock, i.e., no part can proceed due to circuit wait. Its prerequisite is liveness supervision; however, temporal information is associated to certain operation stages for the sake of performance evaluation and throughput maximization, i.e., to maximize the finished-part quantity per time unit. In the paradigm of Petri nets, we show the criticality to prioritize certain activities locally. Our method can well design ratio parameters among distinct processes to improve system productivity.