移动群智感知中基于群组的参与者招募机制

Group-based Participant Recruitment Mechanism in Mobile Crowd Sensing

在移动群智感知中,平台需要招募大量的参与者来完成一项包含众多感知类型的复杂任务.本文研究有限预算内的移动群智感知中,如何招募合适的参与者完成感知任务这一问题.在此挑战下,平台希望招募到的参与者完成感知任务所带来的总收益最大化,同时,招募总花费不超过给定的预算.不同于以往的研究,本文提出了一种新型招募机制,以群组的形式代替个人的形式进行招募.该机制综合考虑了3种类型的特征(覆盖率、信誉和积极性)衡量群组的感知能力,并设计了一种基于遗传算法的群组招募算法最大化群组感知能力.经过实验评估,本文提出的参与者群组招募算法在任务执行效率、平均任务质量、任务完成率和招募人数方面均优于其他个人招募算法.

In mobile crowdsensing(MCS),the platform needs to recruit a large number of participants to complete a complex task involving various types of sensing.In this paper,we focus on the problem of how to recruit appropriate participants to complete the task under budget constraint in MCS.Under this challenge,with a limited budget,the platform hopes to maximize the total profitsof sensing tasks completed by the recruited participants.Different from previous studies,we propose a new recruitment mechanism,which uses the form of group instead of individual.The mechanism combines three types of characteristics(coverage,reputation,and positivity)to measure the group sensing ability,and a group recruitment algorithm based on genetic algorithm is designed to maximize group sensing ability.The experimental evaluations show thatthe proposed group recruitment algorithm is superior to other individual recruitment algorithms in terms ofthe task execution efficiency,average task quality,task completion rate and number of participants recruited.