Multi-Level Two-Sided Rating Protocol Design for Service Exchange Contest Dilemma in Crowdsensing

Abstract

Strategic users in a service exchange application of crowdsensing are apt to exhibit malicious behaviors such as greed, free-ride, and attack, resulting in the phenomenon that no user is willing to serve others and low social utility is obtained in myopic equilibrium, which is considered as a service exchange contest dilemma. To address this issue, we propose a game-theoretic framework of multi-level two-sided rating protocol using all-pay contests to balance service request and service provision between users, in which a user is tagged with a multi-level rating to represent her social status, and she is encouraged to take the initiative to be a server and provide high-quality services to increase her rating. The two-sided rating update rule updates the ratings of both service requesters and service providers, and thus no one can always get services without providing services. By quantifying necessary and sufficient conditions for a sustainable multi-level two-sided rating protocol, we formulate the problem of selecting the optimal design parameters to maximize the social utility among all sustainable multi-level two-sided rating protocols, and design a low-complexity algorithm to select optimal design parameters via a two-stage procedure in an alternate manner. Finally, the extensive evaluation results demonstrate how intrinsic parameters impact on recommended strategies, design parameters, as well as the performance gain of the proposed rating protocol.