Bounded rationality for BitTorrent networks

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] As peer-to-peer networking has becoming a focus in the global networking realm, peer-to-peer distributed systems have been through rapid developments these days. One typical example of these systems is the BitTorrent-based network. In this thesis, a new incentive mechanism is proposed for BitTorrent networks. Compared with the original BitTorrent mechanism, this proposed mechanism is based on the theory of bounded rationality, which is a branch in game theories. In this thesis, we also prove the feasibility of reducing the unfairness among peers in BitTorrent networks by incorporating the idea of bounded rationality into the system. We will present an Epsilon-equilibrium-formed model specific to the situations in BitTorrent networks. The discussions and evaluations of the proposed incentive mechanism over the original one are given as well.