Maximally Permissive Robustness Control of Automated Manufacturing Systems Using Petri Nets and Inhibitor Arcs

For automated manufacturing systems, resource failures are very common. So, robustness is a primary concern. Although there are many robust control strategies, most of them are deficient in achieving maximally permissive robustness control. This paper design maximally permissive robustness controller using Petri nets (PNs) and inhibitor arcs. We show that unreliable resource failures can be formalized by using a new PN structure with inhibitor arcs, called failure-recovery subnets. The input transitions of unreliable resource places, denoted as unreliable transitions, are prevented from firing by the corresponding failure-recovery subnets. We further show that there exist reliable transitions which are "dangerous"; namely, they are eventually forbidden to fire by unreliable transitions after they occupy the reliable resources which are necessarily required by the reliable subnet systems to proceed repeatedly. This damages the robustness of the system. Therefore, we identify the dangerous transitions by solving a set of integer linear programming problems and prevented them from firing by designing a robust controller. A lots of example are presented to illustrate the proposed approach.