It is a common belief that ontologies are the necessary building blocks for the creation of efficient knowledge repositories that will lead to the realization of the Semantic Web. However, real-world applications of ontologies have shown that even medium-sized domains are very hard, if not impossible, to be described by large, monolithic ontologies. It is therefore necessary to pursue the creation of small, focused ontologies that capture and model a very specific knowledge domain.
Nonetheless, users usually need to access information stored in other knowledge domains and hence under a different ontology. In this case, the problem of conflicting and overlapping ontology semantics arises, during information exchange. Varying terminology and/or different conception and design even for concepts similar between the two domains causes this problem.
Creation of mappings and semantic bridges between the different ontologies can provide a solution to this problem. In practice, though, performing this task manually can prove an expensive and error-prone process. The task of ORA (Ontology Reconciliation Agent) is to speed-up the conflict reconciliation process by allowing a number of dedicated agents in a MAS to negotiate their differences. This way, domain experts and knowledge engineers will reach the optimal mapping scheme in shorter time.