LAUSR

In large-scale mobile crowd sensing systems, multi-task-oriented worker selection has shown an increased efficiency in workers’ allocation. However, existing solutions for multi-task selection mainly focus on meeting the requirements of the available tasks and fall short in considering workers’ preferences. Assigning workers to their preferred tasks should substantially improve the possibility that they will perform the tasks assigned to them, which will improve the quality of the sensing outcome. In this paper, we propose to use the Gale–Shapley matching game selection to allocate multiple workers to multiple tasks based on the preferences of both the tasks and workers. It aims at maximizing the level of satisfaction for the workers, by assigning them to their most preferred tasks, as well as maximizing the Quality of Service (QoS) and the completion confidence of the tasks. Simulations based on the real-life dataset show that the proposed approach outperforms other multi-task allocation benchmark in terms of the completion confidence of the tasks, the QoS of the sensing outcome, and the workers’ satisfaction level without compromising on their traveling distance.