| Ontologies are becoming increasingly important in various large-scale information systems such as health care systems. Ontologies give a coherent user-centric view of application domains. A major obstacle in developing ontology-based applications is the poor capability of current techniques to handle large ontologies. The web ontology language (OWL) is a semantic markup language for ontologies. Description logic (DL)-based OWL is used for developing ontologies where automated reasoning services are required. Classical DL reasoners such as FaCT++, HermiT, Pellet and Racer are main memory oriented and are not scalable. One of the approaches to improve the scalability is the direct manipulation of ontologies into databases by reasoning systems. However, using this method, DL-Lite is the maximal fragment that can be used for reasoning over ontologies stored into databases by query rewriting algorithms. We propose a hybrid approach which combines a logic-based reasoning strategy with materialization of knowledge into a relational database to enable reasoning over large ontologies based on the OWL 2 RL profile, a sublanguage of OWL 2, which is more expressive than DL-Lite and amenable to rule-based implementation. We also develop a restriction checker to check the ABox consistency of OWL 2 RL ontologies. |
