Comparing Ontologies Using Multi-agent System and Knowledge Base

This paper presents an approach for handling several ontologies in a domain by integrating a knowledge base in a multi-agent system. For some online facilities, like e-business, several ontologies in different languages are needed to match the users' request. Finding the ontologies is one of the tasks, comparing and combining these ontologies is the other. The accomplishment of these tasks depends on the content in the ontologies like tags and structure but, in some cases, also language and matching techniques. Matching the contents is difficult due to differences between ontologies, often resulting from the lack of explicit and exact standards and development guidelines. This complication increases with the ontologies diverged languages. These problems are tackled by applying a multi-agent system wherein the agents, i.e., software agents and meta-agents, use the users' request to search for ontologies and the knowledge base to compare and combine the contents of the ontologies to create an overall solution. The software agents search for ontologies; the meta-agents keep track of the software agents, ontologies and the knowledge base that reasons with the contents of the ontologies.