OntoDaWa, un système d'intégration à base ontologique de sources de données autonomes et évolutives

More and more integration systems of data sources use ontologies in order to solve the traditional conflicts. The principal limitations of these systems are: (1) the absence of source autonomy and (2) the lack of support for asynchronous evolution of sources and ontologies. In this paper, first, we propose an integration technique allowing a fully automatic integration of sources, while offering to each source a significant autonomy. It supposes that each source contains its own ontology and the semantic relations that articulate it with a shared ontology. Second, we propose a model for managing ontology evolutions. Finally, we present a method for managing the life cycle of instances avoiding any duplication of data.