Agent association group negotiation within Intelligent Environments

Agents working within Intelligent Environments often need to associate themselves with others whether this is to gather data or use services. In some circumstances such as those within machine condition monitoring these associations must be formed using noisy data and a mechanism needs to put in place to allow agents the freedom to negotiate over those they wish to be associated with. In this paper we describe two methods of negotiation based on voting techniques that allow agents to create groups of associations by passing simple messages to each other in a distributed and scalable fashion, a majority voting system and a refinement voting system. We show that our refinement method based on agent persuasion outperforms the majority voting method and show that one of the most important influences on the accuracy of these techniques is average group size.