G.: Efficient processing of huge ontologies in logic and relational databases

Abstract. Today ontologies are heavily used in the sematic web. As they grow in size reasoning systems can’t work without secondary storage anymore. Thus database technology is required for storing and process-ing huge ontologies. In this paper we present an efficient technique for representing and reasoning with ontologies in databases. We also present some benchmarking results in comparison with previous approaches. 1 Mapping Ontologies into Logic Programs Converting ontologies into description logic programs (DLP; DLP covers most of OWL1 Lite plus a portion of OWL DL, namely general concept inclusions with disjunction and qualified existential qualification on the left hand side) would allow to use deductive database systems as reasoners. A straightforward “Direct Mapping ” approach (DiMA) to convert ontologies into a DLP was suggested in [1]. This approach maps every class or property definition into a rule and every class-instance or instance-property-instance relationship into a fact. The following gives a brief example of some frequently used statements