Meta-modeling Extension of Horn-SROIQ and Query Answering

We investigate reasoning and query answering in expressive domain knowledge bases (KBs) that are in OWL 2 Full and contain large individual data sets. For this, we introduce Hi(Horn-SROIQ) and meta-queries based on HiLog semantics. Hi(Horn-SROIQ) is extended from Horn-SROIQ, the horn fragment of the most expressive description logic (DL) for OWL 2 with well known low data complexity, by allowing classes and roles to be used as individuals. Accordingly, meta-queries are obtained from conjunctive queries by allowing variables to appear in the class and role positions. For reasoning, we first provide a method of reducing satisfiability checking and conjunctive query answering in Hi(Horn-SROIQ) to the corresponding reasoning tasks in Horn-SROIQ soundly and completely, then we show that meta-query answering in Hi(Horn-SROIQ) can be captured by conjunctive query answering. Based on this, we obtain that adding meta-modeling capability to Hi(HornSROIQ) has no impact on the complexity of the considered reasoning tasks. These results make Hi(Horn-SROIQ) appealing for practical usage.