Tableau-based revision over SHIQ TBoxes

Abstract

Semantics-based applications encapsulate commonly a set of ontologies which represent knowledge formalized from different data sources. Some of these ontologies may change over time since, not only data would be updated but also our understanding on application domain would evolve. To ensure that ontologies remain usable, it is needed to revise ontologies in such a way that takes into account new knowledge and guarantees the principle of minimal change. In this paper, we propose an ontology revision approach which uses finite structures equipped with a total pre-order to characterize a set of models of an ontology. This allows us to introduce a revision operation that satisfies all revision postulates. Moreover, we propose a procedure for revising an ontology expressed in an expressive description logic, namely SHIQ, and show that the resulting ontology remains expressible in SHIQ.