Observability with respect to a language in discrete event systems modeled by IPN

Abstract

In real DES applications, the system is usually confined into a particular behavior in order to perform a specific task. In this case, the observability problem is turned into the problem of determining whether or not it is possible to recover the actual system state when the system is confined into a specific sublanguage. In this context, this paper addresses the observability problem for interpreted Petri nets (IPNs) whose behavior is confined into a cyclic sublanguage of the firing language. The concept of L-observability is introduced. A characterization of this kind of observability is derived based on the notions of L-event-detectability, and L-marking-detectability is also introduced in this paper.