An ontology-mapping service for agent-based automated negotiation

Abstract

The automation of B2B processes requires a high level of interoperability between potentially disparate systems. We model such systems using software agents (as enterprises representative agents), which interact using ACL protocols. This paper presents, in the context of an Electronic Institution, an ontology-mapping service that enables the automation of negotiation protocols where the involved agents may use different ontologies to represent their domain knowledge. The ontology-mapping service employs two approaches used for lexical and semantic similarity, namely N-Grams and WordNet, and poses few requirements on the ontologies representation format. Examples are provided that illustrate the integration of ontology-mapping with automated negotiation.