Assertional-based Removed Set Revision of DL-LiteR Belief Bases

DL-Lite is a tractable fragment of Description Logics, which provides a powerful framework to compactly encode available knowledge with a low computational complexity of the reasoning process. Belief revision is an important problem that appears in many applications and especially in the semantic web area. This paper investigates the so-called “Removed Sets Revision” for revising DL-Lite knowledge bases when a new sure piece of information, called an input, is added. Such strategy of revision is based on inconsistency minimization and consists in determining smallest subsets of axioms that should be dropped from the current DL-Lite knowledge base in order to restore its consistency and accept the input. We consider different forms of the input: a positive inclusion axiom, a negative inclusion axiom and a membership assertion. We then show how to use the notion of hitting sets for computing removed sets. We show in particular, for some revision strategies and some kinds of inputs, that the removed sets revision can be achieved in polynomial time. Finally, we rephrase the Hansson’s postulates for belief bases revision within a DL-Lite framework in order to give logical properties of removed sets revision operators.