Rewriting Conjunctive Queries over Description Logic Knowledge Bases

We consider the problems of conjunctive query answering and rewriting for information integration systems in which a Description Logic ontology is used to provide a global view of the data. We present a resolution-based query rewriting algorithm for DL-Lite ontologies, and use it to show that query answering in this setting is NLogSpacecomplete with respect to data complexity. We also show that our algorithm produces an optimal rewriting when the input ontology is expressed in the language DL-Lite. Finally, we sketch an extended version of the algorithm that would, we are confident, be optimal for several DL languages with data complexity of query answering ranging from LogSpace to PTime-complete.