A polynomial-time algorithm for diagnosability verification of discrete event systems

This paper concerns the problem of diagnosing the occurrence of permanent fault events in partially observed discrete event systems modelled as finite-state automata in a event-based framework. The first step is the verification of the existence of diagnosers (the diagnosability problem) for such systems. We define a necessary and sufficient condition of diagnosability of such systems. The results proposed in this paper allow testing the diagnosability of discrete event systems in an efficient way, i.e. in polynomial time. The diagnosability test is stated in terms of existence of similar cycles of observable events both in the faultless model and in the faulty model.