A Petri net-based approach to robust deadlock prevention in automated manufacturing systems with unreliable resources

This paper studies the robust deadlock control issues in automated manufacturing systems (AMSs) with unreliable resources. The purpose of this paper is to synthesize a robust prevention supervisor to guarantee that the controlled system can operate smoothly when unreliable resources work well, and to ensure that parts that do not require any failed resources in their remaining routes can continue to be processed even if unreliable resources malfunction. In this paper, Petri nets (PNs) are employed to model the considered AMSs, which allow multi-quantity and multi-type of resource acquisitions. First, to avoid the resource’s underflow, a set of resource constraints denoted as a set of inequalities are developed based on the nominal capacity of shared resources and the minimal resource requirements of processes. Control places (monitors) are derived along with their control variables so as to resolve the blocking states caused by resource failures. Second, to guarantee the system’s liveness, a monitor is designed for each considered siphon in the different subnets derived from a system net. As a result, a robust prevention supervisor is synthesized for AMSs with unreliable resources. The effectiveness and superiority of our method are elucidated by some analyses and discussions.