Game-theoretic Design Of Optimal Rating Protocols For Crowdsensing

With the rapid development of internet of things,cloud computing,big data and the ubiquity of mobile devices that are equipped with multiple sensors,crowdsensing has become a new paradigm for pervasive sensory data collection,analysis and exploration,and has been widely applied to many areas of people' daily lives.Over the past decade,despite the emergence of many successful research results on crowdsensing,the problem of the insufficient number of participants and low-quality sensing data prevent the development of crowdsensing.The main reason are the lack of appropriate punishments for such malicious behaviors.Therefore,incentive mechanisms are of great importance to compel users to participate in crowdsensing and submit high quality sensing data.For crowdsourcing contest dilemma problem,we first achieve extortion and cooperation simultaneously,aiming to maximize their utilities of all tasks and provide workers sufficient incentives of contributing good behaviors.Workers' heterogeneity is taken into consideration when modeling the two-stage asymmetric crowdsourcing contest dilemma game.The rating protocol integrates binary rating labels with pricing mechanism to incentivize workers to contribute good behaviors.Differential punishments are used to transfer payoffs from low-rating workers to high-rating workers,which can reduce performance loss in the presence of imperfect monitoring.Obtain locally optimal design parameters according to Linear programming solution and obtain the global optimal design parameters according to several locally optimal design parameters.Finally,a low-complexity algorithm is proposed to select optimal design parameters.Simulation results show the validity and effectiveness of our proposed algorithm for crowdsourcing contest dilemma.A service exchange dilemma arises when there is non-cooperation among self-interested users.The specific features of crowdsensing are taken into consideration,such as a large number of anonymous users having asymmetric services requirements,different service capabilities,and dynamically joining/leaving a crowdsensing platform with imperfect monitoring.We develop the first game-theoretic design of the two-sided rating protocol to stimulate cooperation among self-interested users,which consists of a recommended strategy and a rating update rule.The recommended strategy recommends a desirable behavior according to intrinsic parameters,while the rating update rule involves the update of ratings of both users,and uses differential punishments that punish users with different ratings differently according to users' historical strategy.By quantifying necessary and sufficient conditions for a sustainable social norm,we formulate the problem of designing an optimal two-sided rating protocol that maximizes the social welfare among all sustainable protocols.We verify that the problem is a convex optimization problem and provide a low-complexity algorithm to select optimal design parameters in an alternate manner according to Alternating multiplier optimization.Finally,evaluation results show the impact of intrinsic parameters on optimal recommended strategy,design parameters,and we can find that the performance of the optimal design two-sided rating protocol is to be as close as possible to the social optimum which is impossible to be achieved.In order to solve incentive users to submit high quality sensing data in online environment,real-time user capability assessment and real-time user selection problem,we propose the online rating protocol.An online rating protocol is first developed for practical crowdsensing to deal with adverse selection and moral hazard simultaneously,which integrates the quality of sensing,rating update,user selection,and payment determination.An endogenous learning approach is designed for quantifying heterogeneous users' ratings based on their historical contribution.Under this approach,the users are motivated to exert high efforts regardless of their ability,and thus increase their ratings and the probabilities of being selected by the platform.An incremental learning approach is exploited to compute the density threshold as the user selection criterion based on previous users' information,which becomes more accurate and efficient as the sample increases.The proposed protocol is theoretically proven to achieve rating coverage,computationally efficiency,budget feasibility,individual rationality,truthfulness and constant competitiveness.Moreover,experimental results support the performance and validate these theoretical properties of our protocol.