Mappings Between Domain Models in Answer Set Programming.
Stefania Costantini, Andrea Formisano, Eugenio Giovanni Omodeo · 2003
Abstract. Integration of data is required when accessing multiple databases within an organization or on the WWW. Schema integration is required for database interoperability, but it is currently, a manual process. The so-called Global-as-View approach constitutes an effective proposal for the design of data integration systems. The combination of diverse database schemata into an integrated view is achieved by means of a global conceptual data model. Consequently, the user is provided with a unified view of the data and (s)he can query such a global schema ignoring the location and structure of the data sources. However, data integration involves resolving conflicts. Schema conflicts include naming, structural, and semantic conflicts. Then, when trying to answer a query in a data integration system, various forms of incompleteness and/or uncertainty may arise. This work aims at demonstrating that Answer Set Programming is a suitable paradigm for defining and implementing data integration systems. In particular, we define a formalization of the Global-as-View approach which is able to cope with forms of incompleteness in data correspondence. We present a working inference engine for effectively answering queries, where each answer set corresponds to one of the possible consistent answers to the given query.