移动群智感知多任务参与者优选方法研究  

Multitask-oriented Participant Selection in Mobile Crowd Sensing

作　　者：牛宏英 刘文菊[1] 王赜[1] 孙士民 NIU Hong-ying;LIU Wen-ju;WANG Ze;SUN Shi-min(Tianjin University of Technology,School of Computer Science and Technology,Tianjin 300000,China)

机构地区：[1]天津工业大学计算机科学与技术学院,天津300000

出　　处：《小型微型计算机系统》2020年第8期1707-1712,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61702366,61802281)资助;天津市科技重大专项与工程项目(15ZXHLGX003901)资助;天津市教委科研计划项目(2017KJ090)资助;天津市自然科学基金项目(19JCYBJC15800)资助。

摘　　要：本文针对参与者优选方法提出基于多任务分发的研究算法,目的在于实现参与者优选过程中的最小化参与者人数的基础上最小化用户移动的总距离.在规定的时间内通过完成多个任务以实现参与者最大化奖励值.而平台需要根据候选者完成任务个数以及移动总距离来选择参与者,在平台信息最小化的情况子下实现成本最小化.本文设计了VT-MOST、VPT-MOST算法,分别以任务、参与者为中心的参与者优选方法.不同于T-MOST、PT-MOST的情况,这里我们考虑速度对于参与者选择的影响.选用参与者近期速度的平均值来估算该用户的速度值,通过对三种算法进行实验对比,分析各种因素对参与者的选择产生的影响,如任务个数、时间、候选者人数等.In this paper,a research algorithm based on multitask distribution is proposed for participant optimization method,which aims to minimize the total distance of user movement on the basis of minimizing the number of participants in the process of participant optimization.Participants can maximize personal incentives by completing multiple tasks within a specified time frame.The platform selects the participants according to the number of tasks completed and the total distance of the participants,in order to minimize the cost under the condition of minimizing the platform information.In this paper,VT-MOST and VPT-MOST algorithm is designed,one is task-centered and the other is participant-centered participant optimization method.Unlike T-MOST and PT-MOST,we consider the effect of speed on participant selection.The average speed of the participant is selected to estimate the speed value of the user.The influence of various factors on the selection of the participants is analyzed by comparing the three algorithms,such as the number of tasks,time,the number of candidates and so on,is analyzed by the experimental evaluation of the three algorithms.

关 键 词：移动群智感知 任务分发 多任务优选 参与者选择 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]