Online task dispatching and pricing for quality-of-service-aware sensing data collection for mobile edge clouds

Abstract

The proliferation of mobile devices equipped with rich sensing and computing resources has pushed the emergence of a new cloud paradigm, mobile edge clouds, where tasks are dispatched from the centralized cloud to the network edge. By taking the advantage of widely-distributed mobile devices, urban monitoring-oriented crowdsourcing services can be provided by a mobile edge cloud, where fine-grained monitoring data over time are crowdsourced by mobile devices and then useful information is extracted. However, as considerable costs are incurred on mobile devices, there exists a major problem that a high financial budget is required to guarantee the quality of service. Fortunately, we observe that real-world sensing data exhibit strong spatial and temporal correlations, and advanced inference methods can be employed to efficiently recover missing data. Motivated by the observation, we provide a near-optimal online task dispatching approach for crowdsourcing services provided by a mobile edge cloud, aiming to minimize the total cost incurred on devices while guarantee the quality of service in the meantime. Besides, considering strategic device users with private cost information, we also propose a truthful pricing policy. Extensive simulations based on real datasets show that our approach outperforms other competing schemes, producing a high quality of service with a much lower budget.