A colored Petri net-based approach to the design of controllers

In this paper, we extend the supervisory control theory and a supervisor synthesis problem to a class of colored Petri nets. More specifically, we investigate the forbidden state control problem with full observation in which the discrete-event system is modeled as a colored Petri net with a symmetry specification. This problem is decidable if the colored Petri net has finite color sets and bounded places. A new algorithm for deriving a controller is presented in detail with a proof of its correctness. Unlike conventional algorithms that explore the entire reachable set of states, our algorithm avoids an exhaustive search of the state space by exploiting a symmetry specification. It performs particularly well when applied to large but structured processes with similar components. Furthermore, this approach allows to represent a controller in a compact form.