Integration of ontologies with programs based on rules

The combination of description logics and rule-based reasoning systems has been widely studied in last years, with the proposal of several different systems that achieve this goal. Description logic programs (dl-programs) were introduced a few years ago as a mechanism to combine a description logic knowledge base with a logic program that can access and dynamically change its view of the knowledge base. The original definition of a dl-program was later extended with the capability of combining several description logic knowledge bases. In Multi description logic programs (Mdl-programs) the logic programming represents the “conductor” that “coordinates” the several knowledge bases, which can be physically separated or independently maintained. In this dissertation we show how a current implementation of dl-programs, the DLplugin for dlvhex, can be extended to work with Mdl-programs, where one can work with more than one ontology, keeping them completely separated. Mdl-programs can be extended by new syntactic constructions – observers – allowing to extend concepts or roles from a knowledge base (in the program’s view of that knowledge base) automatically with all instances of a predicate in the logic program or reciprocally. This syntactic construction also was implemented in the DL-plugin as annotations in the logic program. We provide a performance analysis from which we can conclude that an Mdl-program with observers has practically the same performance as a similar Mdl-program. One Mdlprogram with observers is shorter and more legible just because all global observations are clearly marked. This dissertation also provides some case-studies to illustrate the constructions detailed above, using real-life ontologies freely available on the internet.