On Designing And Optimizing Of Rating Protocol For Crowdsourcing Contest Dilemma

With the development of knowledge economy and information technology,it is more convenient for users to acquire content,while enterprises also pay more attention to the interaction with users.Users no longer merely act as consumers,but as creators transform the company's innovation model into a platform,open,and collaborative style.Thus a new and original pattern has come into being,that crowdsourcing.Crowdsourcing means the task that a company or organization will be originally performed by the staff,outsourced to non-specific web users in a free and open manner.This novel problem-solving model can bring together talents from all walks of life and give full play to the wisdom of the group.It is of great significance to reduce innovation costs and improve work efficiency.The number of crowdsourcing applications and the openness of participants have gradually been expanding in recent years,while giving selfish users the opportunity to perform malicious behaviors(such as free-riding and false-reporting),leading to unfavorable social equilibrium.To curb the adverse effects caused by selfish behavior,more and more scholars devote themselves to related research in this field,mainly based on pricing mechanism and reputation mechanism.But it is not feasible to use traditional pricing mechanism or reputation mechanism to solve the crowdsourcing contest dilemma.On the one hand,the behavior of the participants depends on his costs,rewards and long-term benefits,so the pricing mechanism can not uniquely determine the strategy chosen by the participants;On the other hand,using a single reputation mechanism in crowdsourcing system can only reflect the past historical behavior of the participants,and can not reduce the occurrence of malicious behaviors,it is necessary to provide the difference rewards and punishments mechanism for participants with different reputations.This paper based on game theory to design anindiect reciprocal rating protocol combined the pricing mechanism and reputation mechanism.The practical research is as follows:Due to the malicious behavior of the participants,attacking opponents can not improve their profits,but reduce the overall social benefit,which leads to crowdsourcing contest dilemma.Considering the imperfect monitoring and the cost of crowdsourcing contest,this paper models the crowdsourcing contest as a two-stage sequential game,use the reverse induction method to carry on the equilibrium analysis of the two stages,and get the conclusions that the crowdsourcing contest dilemma still exists in the real environment.To solve this problem,this paper proposes a rating protocol combining reputation mechanism and pricing mechanism based on repeated game theory.According to the principle of one-shot deviation,this paper analyzed selfish individual's compliance with social norms,and verified the necessary and sufficient conditions for the sustainability of rating protocol.Based on this,the reward factors,penalty factors and optimal pricing schemes that satisfy incentive compatibility are calculated.The simulation experiment analyzes the influence of system parameters on the design variables and system benefits in the rating protocol.The participants opted for a free ride because of short-sighted pursuit of immediate gains in the crowdsourcing system;In addition,participants log out from the system and re-register in a new identity to re-enjoy new user preferences in order to evade punishment(whitewashing).To solve the above two problems,this paper uses the piece-counting method to find the sufficient and necessary conditions for the rating protocol to be sustainable based on the one-shot deviation principle,that the benefits that participants perform a recommendation strategy in N-time tasks are always greater than the benefits of executing at least one betrayal strategy at a time in N-time tasks.This paper verifies the sustainability of the protocol and gives the social optimization design method.The optimal number of pieces and the penalty are calculated to implement the principle of incentive compatibility.The simulation results verify the validity of the system parameters to the designed rating protocol and the impact of maximizing system benefits.