Systematic Assessment of Temporal Reasoning Methode for Use in Autonomous Agnents

Non-monotonic logics for reasoning about time and action have traditionally been verified through their informal plausibility and their application to a small number of test examples. I present a systematic method using an underlying semantics, where a preferential nonmonotonic logic is correct if the maximally preferred models equals the set of histories that can arise in a “scenario simulation” of the underlying semantics. Several old and new logics have been analyzed using the systematic method.