Robust distributed control in discrete-event systems

In the context of discrete-event systems, control is defined as the enforcement of legal behavior in the propagation of and reaction to random discrete events. Conditions for the existence of supervisory control and its efficiency for decentralized discrete-event systems are presented. Monitoring state machines (MSMs) are introduced for robust and efficient supervisory control for a class of distributed discrete-event systems. It is shown that the MSMs can significantly improve performance, diagnosis, and fault recovery. The author states three necessary and sufficient conditions for the existence of MSMs and describes measures of efficiency to compute the efficacy of MSMs in distributed control. An illustrative example of a manufacturing system is given to show the application of MSMs in distributed control.<>