Equilibrium Analysis Of Game Decision-making In Macro-task Crowdsourcing Systems

Crowdsourcing is a paradigm in which the company outsources complex and tough tasks to the online public in a free and voluntary way.Crowdsourcing system has been widely used in indoor localization,environmental monitoring,data sensing,and other fields because of its advantages of reducing enterprise costs,using wisdom of crowd,and faster problems solving,and has been widely concerned by scholars in mathematics,control theory,computer science,and other related disciplines.However,many studies have shown that workers’ free-riding behavior can reduce the efficiency of macro-task crowdsourcing systems,and incentive mechanism is a significant measure to solve the free-riding problem.Nevertheless,the incentive budget of the requester is usually limited.Therefore,it is worth exploring how to reach the strategic equilibrium between the workers’ effort level and the incentive budget of the requester in the macro-task crowdsourcing system.Thus far,most of works study this problem by considering that involved workers can obtain the total reward provided by the requester,which collides with many realistic situations.Hence,this thesis considers an S-shaped reward function to more faithfully depict the dependence of obtaining the reward from the requester on the collective efforts of workers and studies how the collective-effort-dependent reward function affects the strategic equilibrium between the workers’ effort level and the incentive budget of the requester.The specific research content and results of this thesis are as follows:Firstly,this thesis studies the strategic equilibrium between the worker’s effort level and the incentive budget of the requester in a macro-task crowdsourcing system under the framework of classical game theory.This thesis applies the Stackelberg game to model the interactions between the requester and workers in the macro-task crowdsourcing system,in which the former is the leader,while the latter group are followers.Accordingly,in this thesis,the problem of solving the strategic equilibrium in the macro-task crowdsourcing system is transformed into the problem of analyzing the Nash equilibrium of the Stackelberg game under the framework of classical game theory.By using the backward induction,this thesis theoretically obtains the condition in which the Stackelberg game has a unique Nash equilibrium under the framework of classical game theory and provides the corresponding solution algorithm.Finally,numerical calculations verify the results of theoretical analysis and the effectiveness of the algorithm.Secondly,this thesis further considers the dynamic interactions between the requester and workers and explores the strategic equilibrium between the worker’s effort level and the requester’s incentive budget in the macro-task crowdsourcing system under the framework of evolutionary game theory.Accordingly,this thesis supposes that the size of worker is large and all the workers choose the same effort level.Based on this,in this thesis,the problem of the solving the strategic equilibrium in the macro-task crowdsourcing system is transformed into the problem of analyzing the Nash equilibrium of Stackelberg game under the framework of evolutionary game theory.By using the backward induction and adaptive dynamics,this thesis theoretically obtains the condition where the second stage game has a unique evolutionarily stable strategy and the Stackelberg game has a unique Nash equilibrium under the framework of evolutionary game theory.Finally,numerical calculations verify the results of theoretical analysis and the effectiveness of the algorithm.