A REPUTATION BASED NEGOTIATION MODEL FOR BARTER TRANSACTIONS BETWEEN SOFTWARE AGENTS

Abstract

The interactions between parties within large open systems are driven by trust. A major component in building trust is represented by reputation. When the trust value of a target agent, computed by its partner based on previous experience, and the reputation it has been assigned are different, an agreement is reached by negotiation. This paper describes a normative multiagent system and proposes a negotiation model where reputation represents the leverage that ensures norm enforcement. The negotiation framework is comprised of a marketplace using barter exchanges. Based on this model, the paper shows how self-interested agents manage to establish cooperation relationships in order to accomplishing their goals, being aware of the influence that reputation has on the costs of future transactions.