Optimal Control for Discrete-Event Systems using Simulated Annealing

We propose a formulation and computational algorithm for the design of feedback controls for Discrete-Event Dynamical Systems (DEDS). Models for the process and controller are expressed using timed and untimed Petri nets, respectively. Performance specifications are given in terms Of forbidden states. Numerical solution for an optimal feedback controller is found by simulated annealing. These procedures are demonstrated for a literature benchmark problem, and for supervisory control of a pipeless batch plant.