A Negotiation Model for Large Scale Multi-Agent Systems

Modeling agent negotiation is of key importance in building multiagent systems. Negotiation provides the basis for managing the expectations of the individual negotiating agents, and it enables selecting solutions that satisfy the agents as much as possible. Thus far, most of the negotiation models have limitations when employed in large scale multiagent systems. This paper presents a negotiation model for large scale multiagents systems that is based on qualitative reasoning (QR) and game theory (GT). In the model, each agent classifies its negotiation opponents according to the similarity of their preference model. The agents use QR components of the model to estimate the preference models of their opponents, and to determine the amount of tradeoff associated with the various solution options. Moreover, they use the GT component of the model to determine the social-acceptance of each of the solution options. The output of the QR and GT components of the negotiation model is used to determine the rationale for accepting or rejecting offers made by the opponents