A framework to support heterogeneous reputation models interoperation in Multi-Agent System

Abstract

The reputation of agents in the Multi-Agent System(MAS) has been focused on distributed artificial intelligence domain in recent years. Many efficient and flexible distributed reputation models have been developed. Nevertheless the heterogeneity of agents (including heterogeneous reputation models) becomes one of the obstacles for them to cooperate. Based on the Functional Ontology of Reputation (FORe) and existing reputation systems, a framework is put forward to support interoperation among heterogeneous reputation models in this paper. The elements of the framework, such as rating, recommendation are formally defined. The issues of recommendation report forms, reputation integration, and witness reputation are discussed. Before depicting the agent architecture, we provide a case about how a present reputation model is implemented in the framework. The work will be beneficial to combine heterogeneous reputation models and to integrate applications with different reputation models.