Predicates and predicate transformers for supervisory control of discrete event dynamical systems

Discrete-event systems are studied, treating the state space as the fundamental modeling concept. The control of discrete-event systems using predicates and predicate transformers is treated. Predicates have the advantage that they can concisely characterize an infinite state space. The notion of controllability of a predicate is defined, and the supervisory predicate control problem is solved. A closed-form expression for the weakest controllable predicate is obtained. The problem of controlling discrete event systems under incomplete state observation is also considered, and observability of predicates is defined. Techniques for finding extremal solutions of Boolean equations are used to derive minimally restrictive supervisors