Exhaustive Diagnosis of Discrete Event Systems through Exploration of the Hypothesis Space
Alban Grastien, Patrik Haslum, Sylvie Thiébaux · ANU Open Research (Australian National University) · 2011
We formulate DES diagnosis as the problem of finding preferred candidates in a space of hypotheses. This formulation allows us to compute an exhaustive diagnosis by solving a series of simple problems, called “diagnosis tests”, which consist in finding a candidate in a set of hypotheses. In this way, we can avoid enumerating all explanations of the observation. Experiments on alarm data generated by power networks show that this approach is promising to monitor large