Arborescent Architecture for Decentralized Prognosis of Discrete Event Systems
Ahmed Khoumsi · 2019
We study failure prediction (or prognosis) in discrete event systems using decentralized architectures. We first identify three basic prognoses as the simplest language-based decentralized prognosis architectures. On the other side, inference-based prognosis generalizes several decentralized prognoses, among which the three basic prognoses. Then we propose arborescent prognosis, a framework that generates and uses a tree-like decentralized prognosis architecture. Each node n of the tree is a conjunction or disjunction of the decisions of the two children of n. We show that if inference-based prognosis is applicable to the prognosis objective, then each leaf of the obtained tree is a basic prognoser. This means that by combining basic prognoses, we can realize every prognosis objective that is realizable by inference-based prognosis.