摘要
Abstract
The mobile crowdsourcing data transaction in the industrial field faces a series of challenges,such as the vulnerability of the indus-trial Internet of Things network environment to malicious attacks,unreasonable distribution of crowdsourcing task rewards,low participation enthusiasm of crowdsourcing workers,insufficient transparency of task execution processes,and easy tampering of task records,which will re-duce the reliability of crowdsourcing systems.To this end,research is being conducted on optimizing the quality of crowdsourcing services us-ing game theory,reinforcement learning,blockchain,and digital twin technologies.Firstly,based on the task allocation scheme of evolution-ary game and digital twin,the interactive behavior is analyzed by modeling the four party evolutionary game,and the digital twin model and re-inforcement learning algorithm are used to prevent fraudulent behavior;Secondly,a crowdsourcing data trading scheme based on Stackelberg game and smart contracts is proposed,which introduces blockchain and smart contracts to enhance transaction transparency and reliability,and optimizes profit distribution through a three-stage Stackelberg game;Finally,based on intelligent agent modeling for game decision analy-sis,multi-agent reinforcement learning is used to simulate the evolution of game decisions in the digital twin environment,in order to improve prediction accuracy and explore the influence of hyperparameters on learning results in the DDPG algorithm model.Theoretical analysis and simulation verification show that the proposed model has certain effectiveness and provides a new solution for mobile crowdsourcing data trad-ing schemes in the industrial field.关键词
移动众包/博弈论/强化学习/区块链/多智能体/数字孪生Key words
mobile crowdsourcing/game theory/reinforcement learning/blockchain/multi-agent/digital twin分类
信息技术与安全科学
