Fairness Emergence in Reputation Systems

Reputation systems have been used to support users in making decisions under uncertainty or risk that is due to the autonomous behavior of others. Research results support the conclusion that reputation systems can protect against exploitation by unfair users, and that they have an impact on the prices and income of users. This observation leads to another question: can reputation systems be used to assure or increase the fairness of resource distribution? This question has a high relevance in social situations where, due to the absence of established authorities or institutions, agents need to rely on mutual trust relations in order to increase fairness of distribution. This question can be formulated as a hypothesis: in reputation (or trust management) systems, fairness should be an emergent property. The notion of fairness can be precisely defined and investigated based on the theory of equity. In this paper, we investigate the Fairness Emergence hypothesis in reputation systems and prove that , under certain conditions, the hypothesis is valid for open and closed systems, even in unstable system states and in the presence of adversaries. Moreover, we investigate the sensitivity of Fairness Emergence and show that an improvement of the reputation system strengthens the emergence of fairness. Our results are confirmed using a trace-driven simulation from a large Internet auction site.