基于效用最大化的实时分布式移动群智感知任务分配方法  

Real-time Distributed Mobile Crowdsensing Task Assignment Method Based on Utility Maximization

作　　者：杨桂松[1] 姚秋言 YANG Guisong;YAO Qiuyan(School of Optical-Electrical&Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《计算机与数字工程》2024年第12期3601-3609,共9页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61802257)资助。

摘　　要：为了解决移动群智感知系统中大规模实时任务高响应的要求,研究了参与者动态到达和离开时的分布式的任务分配方案。首先基于任务距离与参与者可达距离进行全局分簇,设计基于参与者信誉的玻尔兹曼分布参与者意愿模型,用于表征参与者传感数据的质量。其次为了实现有效的激励,考虑到参与者的信誉以及平台的预算,贪婪挑选参与者完成任务分配并根据不同簇任务完成率进行剩余预算重分配。最后,提出参与者效用函数,感知平台在预算限制下最大化参与者的效用。使用真实交通轨迹数据集验证了提出的算法在任务分配人数、任务完成率、效用以及剩余预算上的优势。In order to solve the high-response requirement of large-scale real-time tasks in mobile crowd-sensing systems,a distributed task assignment scheme for dynamic arrival and departure of participants is studied.Firstly,global clustering is per-formed based on task distance and participant reachable distance,and a participant willingness model of Boltzmann distribution based on participant reputation is designed to characterize the quality of sensing data of participants.Secondly,in order to achieve effective incentives,considering the reputation of the participants and the budget of the platform,the participants are greedily se-lected to complete the task assignment and the remaining budget is redistributed according to the completion rate of different cluster.Finally,a participant utility function is proposed,and the sensing platform maximizes the participant's utility under budget con-straints.The advantages of the proposed algorithm on task assignment,task completion rate,utility,and remaining budget are veri-fied with a real traffic trajectory dataset.

关 键 词：移动群智感知 大规模任务分配 实时分布式 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]