An Efficient Partially Correlated Task Assignment Algorithm for Mobile Crowdsensing

• Published in: IEEE Global Communications Conference (Online) pp. 253 - 258
• Main Authors: Liu, Peng, Liu, Kun, Peng, Shuo, Zhang, Baoxian, Li, Cheng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.12.2024
• Subjects: Computational complexity; Correlation; Crowdsensing; Global communication; Greedy algorithms; Integer programming; Mobile crowdsensing and correlated task assignment
• ISSN: 2576-6813
• DOI: 10.1109/GLOBECOM52923.2024.10901615

Task assignment is a critical issue in mobile crowd-sensing, which is aimed to maximize the number of completed tasks subject to budget constraints. However, existing work in this aspect did not consider the correlation between the tasks submitted by the same task requester. That is, tasks in the same subset from the same task requester are often correlated such that they are considered completed only when all of them are completed, and partial completion of them are useless. This requirement largely affects the performance of existing algorithms for the assignment of such partially correlated tasks. In this paper, we formulate the problem of maximizing the total number of completed tasks subject to such correlation and also budget constraints as an integer programming problem. We propose two greedy algorithms, one is requester happiness utility based algorithm and the other is minimum task remaining subset first algorithm. We present design details of both algorithms and deduce their computational complexities. Numerical results demonstrate that these two algorithms can significantly outperform the existing work.