A Method for simultaneously Handling Uncertainty Associated with default Reasoning and Ignorance about Knowledge

In this paper, we develop a model for uncertain default reasoning. It has the following characteristics: 1) basic uncertainty is estimated in terms of an interval number, and so the ignorance about knowledge can be coped with; 2) it has some capability of handling commonsense knowledge in forms of default rules with uncertainty, in which Reiter's system is a special case; and 3) the propagation for uncertainties is built completely on fuzzy mathematics. Hence, this paper forms an alternative approach that allows us to handle ignorance, uncertainty and incompleteness simultaneously. Keywords: Uncertain Reasoning, Default Reasoning, Interval Number, Fuzzy Systems. 1 Introduction When developing knowledge-based systems in the real world, commonsense knowledge needs to be represented and manipulate[13]. In order to model this kind of knowledge, several nonmonotonic logics have been proposed since the 70's. One of them is Reiter's default logic [18]. Default reasonings are performed under..