Sandip Sen and Neelima Sajja
We consider the problem of user agents selecting service providers to process tasks. We assume that service providers are drawn from two populations: high and low-performing service providers with different averages but similar variance in performance. For selecting a service provider an user agent queries other user agents for their high/low rating of different service providers. We assume that there are a known percentage of "liar" users, who give false estimates of service providers. We develop a trust mechanism that determines the number of users to query given a target guarantee threshold likelihood of choosing high-performance service providers in the face of such "noisy" reputations. We evaluate the robustness of this reputation-based trusting mechanism over varying environmental parameters like percentage of liars, performance difference and variances for high and low-performing agents, learning rates, etc.