Information Gain Based Maximum Task Matching in Spatial Crowdsourcing

Jiantong Zhang, Feilong Tang, Leonard Barolli, Yanqin Yang, Wenchao Xu

Abstract: Along with the popularization of smart mobile devices and the rapid development of wireless networks, a new class of crowdsourcing, termed with spatial crowdsourcing, is drawing much attention, which enables workers to perform spatial tasks based on their positions. In this paper, we study an important spatial crowdsourcing problem, namely information based maximum task matching (IG-MTM), in which each spatial task needs to be performed before its expiration time and workers are dynamically moving. The goal of IG-MTM problem is to maximize the number of spatial tasks that are assigned to workers while satisfying the quality requirement of collected answers. We first define this problem, and then two approximation approaches are proposed, namely greedy and extremum algorithms. Subsequently, in order to improve time efficiency, we propose an optimization methodology. Through extensive experiments on both real-world and synthetic datasets, we evaluate the performance of our proposed approaches.