Multi-Task Sensing for Multiple Crowdsourcers: A Dynamic Game Based Pricing Model

摘要

Mobile Crowdsourcing is an emerging paradigm for sensing and collecting data over a large area for numerous applications. In metropolitan areas, as an increasing number of applications need to work with multi-source sensing to enhance the data diversity, developing an economic model for such a crowdsourcing system is required to support the multiple concurrent tasks while satisfying certain constraints. In this paper, we address the issue of sensing service pricing where several mobile contributors compete with each other to provide secondary sensing services to the crowdsourcers to fulfill the multi-task sensing demand. The objective of the mobile contributors is to maximize its utility using an equilibrium pricing strategy under the quality of service (QoS) constraints. This secondary sensing approach augments the capabilities of existing mobile contributors without introducing additional costs, resulting in a win-win strategy for both mobile contributors and crowdsourcing systems. We formulate this as an oligopoly market and apply a non-cooperative dynamic game, which is based on the Bertrand model to analyze the impact of several parameters at the Nash equilibrium. For this dynamic game, distributed dynamic learning algorithm is proposed. The stability of the proposed distributed algorithm is studied in terms of the convergence to the Nash equilibrium. All the results are supported by both theoretical analysis and simulations.