A Theory of Dynamic Diagnosis

Diagnosis is, in general, more than a mere passive reasoning task. It often requires to actively produce observations by performing a test series on a faulty system. We present a theory of diagnosis which captures this dynamic aspect by appealing to Action Theory. The reactions of a system under healthy condition are modeled as indirect effects, so-called ramifications, of actions performed by the diagnostician. Under abnormal circumstances - i.e., if certain aspects or components of the system are faulty-one or more of these ramifications fail to materialize. Ramifications admitting exceptions is shown to giving rise to a hitherto unnoticed challenge - a challenge much like the one raised by the famous Yale Shooting counter-example in the context of the Frame Problem. Meeting this challenge is inevitable when searching for "good" diagnoses. As a solution, we adapt from a recent causality-based solution to the Qualification Problem the key principle of initial minimization. In this way, when suggesting a diagnosis our theory of dynamic diagnosis exploits causal information, in addition to possibly available, qualitative knowledge of the a priori likelihood of components to fail.Some of the results in this paper have been preliminarily reported in (Thielscher, 1997a)