Design of Adaptive Supervisors for Discrete Event Systems via Learning
Diana F. Gordon, Kiriakos Kiriakidis · 2000
Abstract In many practical applications, modeling using event-driven dynamics leads to interconnected discrete event systems. Often, these tend to be large-scale event-varying structures, which need to possess certain event-invariant properties. Although supervisory control theory offers some methods for the synthesis of such discrete event systems, adaptive supervision is, by and large, an open problem. This paper proposes an approach to the design of adaptive supervisors based on a systematic revision of the desirable language, via a learning mechanism, so that the system’s properties are safe.