Collaborative memetic agents for enabling semantic interoperability

Abstract

Semantic interoperability represents the ability of two or more systems to automatically interpret the information exchanged meaningfully in order to produce useful results. Currently, the best recognized technology for achieving a specification of meaning is represented by ontologies. However, the variety of ways that a domain can be conceptualized results in the creation of different ontologies with discrepancies and heterogeneities. As a consequence, an ontology alignment process is necessary for bridging this gap and achieving a full communication understanding across different software components. This paper uses a synergetic approach, based on the integration of collaborative agents and parallel memetic algorithms, for efficiently aligning ontologies and, consequently, solving the semantic heterogeneity problem. As shown by a statistical procedure, our approach yields high performance in terms of the ratio between alignment quality and computational effort with respect to conventional evolutionary approaches for ontology alignment.