A Spatial Coverage-Based Participant Recruitment Scheme for Mobile Crowdsourcing

A revolutionary distributed problem-solving model is being enabled by the increasing proliferation of mobile smart devices, namely mobile crowdsourcing, which employs ubiquitous mobile users to gather and analyze data beyond what was traditionally possible, while becoming a leading area for study in both academia and industry. However, how to recruit participants to provide high quality data with a limited budget is challenging for a crowdsourcing system. In contrast to traditional fixed sensors, participants are the general population with smart devices, which boast the unique advantage of predictable mobility. To address this problem, this article first studies the relationship between predictable mobility and spatial coverage, and proposes a new strategy to recruit the optimal subset of participants by jointly considering their current and future locations with the aim of ensuring that the recruited participants can provide high coverage sensory data. Secondly, participant recruitment proves to be a NP-hard problem, while a harmony search algorithm with annealing randomness (HSA-AR) is proposed as a trade-off between the quality of problem-solving and the computation time. Finally, through extensive simulations utilizing real-world and synthetic datasets, we demonstrate that HSA-AR outperforms other baseline methods under various settings.