Team Based Incentive Mechanisms In Crowdsourcing Systems

Crowdsourcing harnesses the potential of large and open networks of people.It is a new phenomenon which can be used to solve the problems that are hard for machines but relatively easy for the human.It attracts substantial interest in many fields,and has evolved in different kinds of crowdsourcing systems.To stimulate individuals' participation and contribution,incentive mechanisms play a key role in crowdsourcing systems.In this thesis,we focus on team-based incentive mechanism,and investigate team strategies in different crowdsourcing models.In all models,team-based mechanism can increase the efficiency or improve the quality of tasks by teaming workers.The contributions of this thesis are threefold:(1)Considering the low efficiency of the traditional “Pay-Per-Hit” incentive scheme,this thesis proposes a novel mechanism named “ABT”(Ability-balanced Team)to greatly improve the efficiency of “Pay-Per-Hit” by introducing team formation,competition and gamification.Platforms ask workers to report their ability levels,and according to crowdworkers' ability levels,form the competing teams.On one hand,a basic payment scheme is designed to incentivize the crowdworkers to truthfully report her ability level and make a corresponding contribution.On the other hand,according to the chosen ability thresholds,crowdworkers are organized into total ability balanced teams to earn extra team bonus,which can further motivate crowdworkers to exert more efforts.Through the experiment,the advantage of the incentive mechanism is verified.(2)Basically,the existing incentive mechanism in mobile crowdsourcing has a lack in quality control.To obtain high quality sensed data,this thesis proposes a quality-aware incentive mechanism which integrates reverse auction and reputation system.This mechanism can motivate workers to form teams and bid in the formation of teams,and this “team-based bidding” can motivate workers to provide high-quality sensed data.Besides,the winner selections algorithm and payment calculation algorithm applied to the mechanism are given,and we prove these algorithms can make sure workers truthfully bid their cost.(3)The task of mobile crowdsourcing is becoming more and more complicated,requiring the cooperation among workers of heterogeneous abilities.The existing team strategies only aimed at reducing economic cost of team but not take communication cost into consideration.This thesis proposes an incentive mechanism which takes both economic cost and communication cost into account.The platform recommends potential partners for bidders and encourages these bidders to invite them to participate in tasks.When selecting winners,through incorporating social relationships of workers,the mechanism optimizes team according to a bi-objective optimization function in economic cost and communication cost so as to select the best team.The advantages of the mechanism are verified by simulations.