Supervisory Controller Optimization for Deadlock Resolution in Automated Manufacturing Systems With Timed Petri Nets

For automated manufacturing systems (AMS), few the existing approaches can incorporate the optimization of some prominent quantitative performance measures during their controller synthesis process for liveness-enforcing supervision. This causes unnecessary control limitations that often lead to higher implementation cost and lower performance. We propose a method in the paradigm of a class of timed Petri nets to minimize the cycle time of such systems. This problem is characterized in the form of mixed integer programming (MIP) formulation which can properly balance the trade-off between the control implementation cost and system efficiency. An example highlights the effectiveness and efficiency of the proposed approach.