ABSTRACT
Discrete event systems (DES) are typically man-made reactive systems such as manufacturing, traffic control and embedded systems. DES behaviours are modelled in terms of states and events; states represent certain situations under which specific properties hold, while events represent significant occurrences that change the properties. The main purpose of DES models is the analysis and design of control functions to achieve some desired behaviour. The typical modelling formalism for DES and the supervisory control theory (SCT) is finite automata (FA). This formalism explicitly represents each state and transition, so that modelling complex systems with FA can lead to incompact and intractable models. Though Extended Finite Automatas (EFAs) allow compact representation of large state spaces, the state-space explosion problem does not simply vanish by using EFAs. Furthermore, the SCT the idea of automatically synthesizing a supervisor can be adapted to work directly on EFAs, and implementations based on binary decision diagrams (BDDs) have shown to be efficient.
