Investigations of Model-Preference Defaults

    Abstract

    In this paper, we explore the expressive and computational properties of the logic of model-preference defaults (MPD). We extend the set of sublanguages of MPD known to be tractable. We compare MPD with the approaches to nonmonotonic reasoning represented by default logic and circumscription. Finally, we discuss MPD’’s suitability for various applications of nonmonotonic reasoning, including inheritance, temporal reasoning, and knowledge representation. In doing so, we show that MPD provides tractable solutions to the problems of temporal projection and mixing strict and defeasible inheritance.

    Assets

    BibTeX