Horn clause contraction functions

摘要: In classical, AGM-style belief change, it is assumed that the underlying logic contains classical propositional logic. This clearly a limiting assumption, particularly in Artificial Intelligence. Consequently there has been recent interest studying change approaches where full expressivity of not obtained. this paper we investigate contraction Horn knowledge bases. We point out obvious extension to case, involving remainder sets as starting point, problematic. Not only do have undesirable properties, but also some desirable functions are captured by approach. For set contraction, develop an account terms model-theoretic characterisation weak sets. Maxichoice and partial meet specified, show problems arising with earlier work resolved these approaches. As well, constructions specific operators postulates provided, representation results examine package or formulas. Again, give construction postulate set, linking them via result. Last, closely-related notion forgetting clauses. arguably interesting since clauses found widespread use AI; given here may potentially be extended other areas which make Horn-like reasoning, such programming, rule-based systems, description logics. Finally, reasoning weaker than sheds light on foundations change.