Current Research Trends in Possibilistic Logic: Multiple Agent Reasoning, Preference Representation, and Uncertain Databases

摘要: Possibilistic logic is a weighted that handles uncertainty, or preferences, in qualitative way by associating certainty, priority levels, to classical formulas. Moreover, possibilistic copes with inconsistency taking advantage of the stratification set formulas induced associated levels. Since its introduction mid-eighties, multiple facets have been laid bare and various applications addressed: handling exceptions default reasoning, modeling belief revision, providing graphical Bayesian-like network representation counterpart base, representing positive negative information bipolar setting preferences fusion version space learning, extending for dealing time, agents mutual beliefs, developing symbolic treatment priorities partial orders between levels also improving computational efficiency, learning stratified hypotheses coping exceptions. The chapter aims primarily at offering an introductory survey developments. Still, it outlines new research trends are relevant preference representation, reasoning about epistemic states.