Modeling and Control of Networked Discrete‐Event Systems

Abstract

Modern engineering systems and cyber‐physical systems often consist of many distributed and networked physical plants, control units, and other devices with wired or wireless communication networks to exchange information between controllers and plants. These systems are called networked systems. In this article, we review supervisory control of networked discrete‐event systems. We first discuss how to model networked discrete‐event systems and how to obtain state estimates using a networked observer. We then introduce and solve the following supervisory control problems of networked discrete‐event systems: (i) (nondeterministic) supervisory control problem, (ii) deterministic supervisory control problem, (iii) nonblocking supervisory control problem, (iv) robust supervisory control problem, (v) robust deterministic supervisory control problem, (vi) deterministic decentralized supervisory control problem, and (vii) nonblocking decentralized supervisory control problem. We present and prove necessary and sufficient conditions for solving the above problems. Finally, we list the open problems in supervisory control of networked discrete‐event systems.

Networked supervisory control with imperfect communication channels

“…The first point is that the existing conditions for predictive supervisors are weaker than those for nonpredictive supervisors, and any specification that can be achieved by any supervisor can also be achieved by a predictive supervisor. One more good point is network controllability is no longer needed, and the detailed discussion will be explained in [28]. The following characterization result is obtained.…”

“…The discrete-event system (DES) community, in particular, the supervisory control community, has also been actively involved in this popular research trend, as illustrated in a large number of publications on modular/distributed control and In review paper [28], the author first discusses how to model networked discreteevent systems and how to obtain state estimates using a networked observer, following the seminal modelling framework introduced in [29], [30], [31]. Then, the authors introduce and solve several supervisory control problems of networked DES, including (i) (nondeterministic) supervisory control problem, (ii) deterministic supervisory control problem, (iii) nonblocking supervisory control problem, (iv) robust supervisory control problem, (v) robust deterministic supervisory control problem, (vi) deterministic decentralized supervisory control problem, and (vii) nonblocking decentralized supervisory control problem.…”

“…Finally, the authors list some open problems in supervisory control of networked discrete-event systems. In contrast to [28], we also provide a detailed review on different modelling frameworks, in particular, including those that are based on modelling channels as automata. We…”