Experimenting with Certified Reputation in a Competitive Multi-Agent Scenario

In hostile environments, there's a risk of low efficiency of reputation due to possible attacks coming from malicious agents. Indeed, it's widely accepted that in competitive scenarios, the use of direct knowledge about the environment (reliability) is more effective than its combination with indirect knowledge (reputation). In the open multi-agent system research field, the notion of certified reputation has been proposed to improve reputation effectiveness, but there's no supporting practical experience with the introduction of certified reputation in competitive environments available. Even though it's obvious that the effectiveness of the reputation mechanism is generally improved when certification is adopted, it isn't clear when this improvement lets us use the reputation mechanism to increase individual gain, thus resuming the role of reputation. This article deals with this problem by assessing the use of certified reputation in the context of competitive agents