Shao Zihao, Wang Huiqiang, Feng Guangsheng
College of Computer Science and Technology, Harbin Engineering University, Harbin 150001, China.
Sensors (Basel). 2019 Jul 2;19(13):2927. doi: 10.3390/s19132927.
Mobile crowdsensing (MCS) is a way to use social resources to solve high-precision environmental awareness problems in real time. Publishers hope to collect as much sensed data as possible at a relatively low cost, while users want to earn more revenue at a low cost. Low-quality data will reduce the efficiency of MCS and lead to a loss of revenue. However, existing work lacks research on the selection of user revenue under the premise of ensuring data quality. In this paper, we propose a Publisher-User Evolutionary Game Model (PUEGM) and a revenue selection method to solve the evolutionary stable equilibrium problem based on non-cooperative evolutionary game theory. Firstly, the choice of user revenue is modeled as a Publisher-User Evolutionary Game Model. Secondly, based on the error-elimination decision theory, we combine a data quality assessment algorithm in the PUEGM, which aims to remove low-quality data and improve the overall quality of user data. Finally, the optimal user revenue strategy under different conditions is obtained from the evolutionary stability strategy (ESS) solution and stability analysis. In order to verify the efficiency of the proposed solutions, extensive experiments using some real data sets are conducted. The experimental results demonstrate that our proposed method has high accuracy of data quality assessment and a reasonable selection of user revenue.
移动群智感知（MCS）是一种利用社会资源实时解决高精度环境感知问题的方法。发布者希望以相对较低的成本收集尽可能多的感知数据，而用户则希望以低成本获得更多收益。低质量数据会降低MCS的效率并导致收益损失。然而，现有工作缺乏在确保数据质量的前提下对用户收益选择的研究。在本文中，我们提出了一种发布者 - 用户进化博弈模型（PUEGM）和一种收益选择方法，以基于非合作进化博弈理论解决进化稳定均衡问题。首先，将用户收益的选择建模为发布者 - 用户进化博弈模型。其次，基于误差消除决策理论，我们在PUEGM中结合了一种数据质量评估算法，旨在去除低质量数据并提高用户数据的整体质量。最后，从进化稳定策略（ESS）解和稳定性分析中获得不同条件下的最优用户收益策略。为了验证所提解决方案的效率，使用一些真实数据集进行了广泛的实验。实验结果表明，我们提出的方法具有较高的数据质量评估准确性和合理的用户收益选择。
