The Research Of Trust Mechanism In P2P E-commerce

With the rapid development and popularization of the Internet, P2P e-commerce is developing step by step as a new model of e-commerce. However, the characteristics of dynamic, anonymity and open in P2P network make e-commerce activities to become more convenient and flexible, but lead to greater degree risk and security threats. Therefore, how to establish a trust mechanism in P2P network, which suppress and eliminate malicious behavior of peers, faith cooperation between the pumping peer, improve the security and stability of the network is becoming a hot topic. On the foundation of concluding recent trust, this article makes a deep research on trust model applied in P2P e-commerce. The main works are as the following:(1) Aimed at the detect and disadvantage of traditional trust model, the trust evaluation model based on peer-trading factors is proposed. The model is established based on many factors that affect the trading behavior, and the calculation method of the trust value is given.The reward and penalty factors are introduced to solve the problem, which effectively suppressed the malicious behavior and encourage integrity of transactions. We give the calculation methods of weighting function about the recommended peers to improve the recommendation accuracy, thus effectively avoided the malicious recommendations and collusive behaviors. The model sets the amount of dynamic trading restrictions based on transaction history and discusses its relationship with the trust threshold, in order to prevent malicious peers to raise their reputation by the means of small amount of sucessful transactions money, then take fraudulent behavior in large transactions. Finally, the accuracy, validity and practicality of the trust model is verified by experiments.(2) Combined with the characteristics of the peers’ transactions in P2P network, the transaction model based on "prisoner’s dilemma" game theory is proposed. The peers in the transaction process select the cooperation strategy with different probability. In order to maximize the interests of both sides, the article obtained the Nash equilibrium of the game model by solving the answer. The theoretical analysis shows that the optimal strategy of peers is not the optimal strategy of the whole network for a single transaction, the model must introduce some necessary measures and mechanisms to promote the cooperation strategy to become the only equilibrium strategy.(3) According to the conclusion of "prisoner’s dilemma" game model, this paper presents a trust model based on infinitely repeated game. The discount rate and incentive-punishment mechanism are introduced to restrain the strategy choices of peers in the model. The different types of deviations are endowed with different penalty periods to ensure that the peers return to the cooperative strategy under the pressure of punishment. Mathematical theory proves that the cooperation strategy is Pareto optimal strategy of infinitely repeated game model, and the simulation experiments show that the game model can effectively against the malicious attacks, reduce transaction risk, and improve the overall performance of the network.