Efficient reasoning about rich temporal domains (Q1113891)

Much of the discussion of this article has been extensively presented elsewhere [the author, Reasoning about change (MIT Press, Boston, MA, 1987)]. The approach concerns the construction of the logic of chronological ignorance which is intended to face the problem of nonmonotonic temporal inferences. In developing the theory the author gives a very general framework for nonmonotonic logics: the preference logics. The notions of preferential satisfiability, validity and entailment are defined. The syntax and semantics of a monotonic logic of temporal knowledge (TK) are first introduced. The nonmonotonic version of the logic of TK follows as the chronological ignorance (CI). The syntax of the two is identical and the semantics of CI follows by associating with TK a preference criterion on Kripke structures. The notions of latest time point of a base formula and chronologically maximally ignorant model of a formula are defined. The framework is used to offer a solution to an instance of the qualification problem: the shooting scenario. Finally, causal theories which are supported by the CI are discussed and a solution to the extended prediction problem is suggested.