有限理性下群智感知的激励机制研究

Mobile Crowd Sensing (MCS) has emerged as a promising data sensing paradigm, through leveraging a large number of users and their handheld mobile devices. Proper incentive mechanisms are indispensable for ensuring the participation of users in MCS. However, the existing incentive mechanisms for MCS rarely considered the bounded rationality of user decision-making. In practice, users usually exhibit bounded rationality due to the limitations in cognition and available information for decision-making, and such a feature has a significant impact on the MCS system performance. In this project we design incentive mechanisms for MCS with bounded rational users:.1) We focus on the bounded rationality due to users' cognitive limitations. Based on cognitive heuristics in cognitive psychology, we first construct a quantifiable model to characterize the bounded rationality of MCS users. Then, combined with the MCS task features, we design incentive mechanisms to jointly improve user participation and optimize MCS system performance..2) We focus on the bounded rationality due to users' available information limitations. Based on prospect theory in behavioral economics, we theoretically model and characterize users' risk preferences with bounded rationality. We further model the strategic interactions between MCS tasks and users, and design effective incentive mechanisms for MCS with bounded rationality..3) We further consider the bounded rationality due to both users' cognitive and information limitations. By combining the cognitive decision-making theory and behavioral economics, we characterize users' cognitive decisions and risk preferences with bounded rationality, as well as their impacts on the MCS system performance..The achievements of our research have the potential to establish an effective incentive framework for MCS with bounded rationality, which will be of great importance for practical applications of MCS.

群智感知是一种基于用户参与和智能移动设备的新兴的数据感知范式。用户的参与度需要合适的激励机制来保障，然而现有的群智感知激励机制研究很少考虑由于用户认知及可用信息受限而导致的有限理性特征，因而在实际应用中效果并不理想。本课题拟研究有限理性下的群智感知激励机制:.1) 针对认知受限的有限理性，构建基于认知启发式决策的有限理性量化表征，并结合感知任务的特征设计激励机制提高用户的参与度及优化群智系统的性能;.2) 针对可用信息受限的有限理性，构建基于前景理论的用户风险决策量化表征，并分析感知任务与用户的交互设计适用于有限理性群智感知的激励机制;.3) 研究认知和信息同时受限的有限理性，提出基于认知决策理论和行为经济学的组合模型来刻画用户的有限理性决策及其对群智系统性能的影响。.本课题有望形成有限理性下群智感知激励机制的一套有效理论框架，并为群智感知在实际有限理性用户群中的部署和激励奠定基础。

群智感知是一种基于用户参与和智能移动设备的新兴的数据感知范式。用户的参与度需要 合适的激励机制来保障，然而现有的群智感知激励机制研究很少考虑由于用户认知及可用信息 受限而导致的有限理性特征，因而在实际应用中效果并不理想。本课题拟研究有限理性下的群 智感知激励机制：1) 针对认知受限的有限理性，构建基于认知启发式决策的有限理性量化表征，并结合感知任务的特征设计激励机制提高用户的参与度及优化群智系统的性能; 2) 针对可用信息受限的有限理性，构建基于前景理论的用户风险决策量化表征，并分析感知任务与用户的交互设计适用于有限理性群智感知的激励机制; 3) 研究认知和信息同时受限的有限理性，提出基于认知决策理论和行为经济学的组合模型来刻画用户的有限理性决策及其对群智系统性能的影响。项目按照拟定的研究内容和研究目标开展工作，在IEEE TNSE、IEEE IoTJ、IEEE TPDS等重要国际期刊和IEEE INFOCOM等重要国际会议上发表论文14篇，其中IEEE系列SCI期刊论文8篇，CCF推荐列表中会议论文6篇（2篇A类，1篇B类，3篇C类）。申请发明专利1项。本项目为有限理性群智感知激励机制建立了一套有效理论框架，并为群智感知在实际有限理性用户群中的部署和激励奠定基础，具有重要的理论与现实意义。