QoS-Based Budget Constrained Stable Task Assignment in Mobile Crowdsensing

Abstract

One of the key problems in mobile crowdsensing (MCS) systems is the assignment of tasks to users. Most of the existing work aim to maximize a predefined system utility (e.g., quality of service or sensing), however, users (i.e., task requesters and performers/workers) may value different parameters and hence find an assignment unsatisfying if it is produced disregarding these parameters that define their preferences. While several studies utilize incentive mechanisms to motivate user participation in different ways, they do not take individual user preferences into account either. To address this issue, we leverage <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Stable Matching Theory</i> which can help obtain a satisfying matching between two groups of entities based on their preferences. However, the existing approaches to find stable matchings do not work in MCS systems due to the many-to-one nature of task assignments and the budget constraints of task requesters. Thus, we first define two different stability conditions for user happiness in MCS systems. Then, we propose three efficient stable task assignment algorithms and discuss their stability guarantees in four different MCS scenarios. Finally, we evaluate the performance of the proposed algorithms through extensive simulations using a real dataset, and show that they outperform the state-of-the-art solutions.