Ontology Based Data Access with Expressive Ontology Languages

In this paper, we considered an application of ontologies to facilitate access to various types of datasources that is known as ontology-based data access (OBDA) where an ontology is used to support query answering for distributed and heterogeneous data sources. For this approach to work, one has to carefully choose the ontology language. We considered ontologies formulated in terms of a suitable description logic. The OWL 2 Web Ontology Language (OWL 2) is an ontology language for the Semantic Web recently designed ad standardised by the W3C. There are three profiles of OWL 2: RL, EL and QL, with OWL 2 QL designed specifically for OBDA via query rewriting. Unfortunately, OWL 2 QL does not contain many concept and role constructs that are useful in many OBDA applications. In order to implement OBDA to the data with a complex structure, it is necessary to be able to rewrite the ontology formulated in some expressive language into ontology written in OWL 2 QL which are specified for OBDA. The goal of this work was to present overview of state-of-art researches with respect to different rewriting techniques and approximations of ontologies. We classified existing results according to OWL 2 QL extensions. We consider example of ontology with disjunctive information, query answering to which is intractable in general. We proved that there is no first-order query rewriting for a special case data structure by reducing query answering problem to well-known graph reachability problem.