Multi-decision decentralized prognosis of failures in discrete event systems

An architecture for the decentralized prognosis of discrete event systems (DES) was recently proposed in, where several local prognosers cooperate for predicting failures in a system modeled as a DES. In this paper, we first formulate the proposition of as a disjunctive architecture, and we develop a conjunctive architecture which is dual and complementary to the disjunctive one. Then, we develop a so-called multi-decision prognosis which generalizes the conjunctive and disjunctive prognosis architectures. The basic principle of multi-decision prognosis consists in using several decentralized prognosis architectures working in parallel. We finally show that our work can be easily extended for predicting a failure at least k steps before its occurrence, for a given k ges 1.