Towards Soundness Preserving Approximation for ABox Reasoning of OWL2

ABox Reasoning in large scale description logic (DL) knowledge bases, e.g. ontologies, is important for the success of many semantic-enriched systems. Performance of existing approaches, such as the tableau-based approach, and the disjunctive datalog approach, is restricted by their theoretical worst case complexity bound. In this paper, we propose a soundness-preserving approximate reasoning approach to address this issue. We first approximate an ontology in DL RO, a major fragment of OWL2-DL, to DL EL++ , the underpin of OWL2-EL, plus an additional table maintaining the complementary relations between concept names. Then we can perform ABox reasoning either internally, or externally of the TBox with additional completion rules. The approximation and reasoning can be performed in PTIME. Our preliminary evaluation shows that our approach can outperform existing DL reasoners on real world and benchmark ontologies.

• Type: Journal
• Year: 2010
• Where: DLOG
• Authors: Yuan Ren, Jeff Z. Pan, Yuting Zhao
• Keywords: Approximate Reasoning Approach, Automated Reasoning, Disjunctive Datalog Approach, DLOG 2010, Reasoning