Reasoning under inconsistency: A forgetting-based approach

In this paper, a fairly general framework for reasoning from inconsistent propositional bases is defined. Variable forgetting is used as a basic operation for weakening pieces of information so as to restore consistency. The key notion is that of recoveries, which are sets of variables whose forgetting enables restoring consistency. Several criteria for defining preferred recoveries are proposed, depending on whether the focus is laid on the relative relevance of the atoms or the relative entrenchment of the pieces of information (or both). Our framework encompasses several previous approaches as specific cases, including reasoning from preferred consistent subsets, and some forms of information merging. Interestingly, the gain in flexibility and generality offered by our framework does not imply a complexity shift compared to these specific cases. Key words: Knowledge Representation, Reasoning under Inconsistency, Forgetting.

• Type: Journal
• Year: 2010
• Where: AI
• Authors: Jérôme Lang, Pierre Marquis
• Keywords: AI 2010, Artificial Intelligence, General Framework, Inconsistent Propositional Bases, Reasoning