• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

Abstract. A growing interest in the practical application of the autonomous agent paradigm in open and distributed environments has been observed in recent years. A host of standards are being developed in this respect, aiming at easing and promoting interoperability between heterogeneous agents and multiagent systems. The work of the FIPA is prominent in this field. The effort of this organization centres on defining an abstract architecture that identifies all the elements involved in achieving the above-mentioned interoperability and establishes the relationships between these elements. Nevertheless, this model does not completely address the inherent social-ability of any multi-agent system and which is especially important for achieving a higher level of interoperability. This paper lays the foundations for an extended role model and an organizational model that will serve as a framework for establishing and specifying relationships between heterogeneous agents interacting in an open environment. The agent and agent information models proposed by the FIPA are adapted and extended to accommodate this social viewpoint and social requirements on agent conversations and communication languages are discussed. We then propose a policy framework that constitutes the basis for agent social infrastructure management. Finally, the most relevant platform service infrastructure needed to support the policy framework is presented and defined. Our aim is to be able to model how agents are going to interoperate with one another within an open environment by means of normative policies, and how this environment will be organized and supported. 1

Citations