Privacy-Preserving Multitask Online Matching in Mobile Crowdsensing: A Snapshot-Based Approach

Kaimin Wei, Shiting Zhao, Jinpeng Chen, Tingrui Pei, Dezhi Sun

Abstract: With the growing popularity of mobile crowdsensing (MCS), online matching has recently attracted considerable attention. However, most previous schemes focused on single-task matching, which limits their practicality in new MCS applications that require multitask matching. Moreover, most MCS tasks require workers to share locations with the platform, which poses serious privacy concerns. To address this issue, we propose a privacy-preserving multitask online matching algorithm in a snapshot-based mode (PMS). Specifically, the entire time period is divided into snapshots to reduce the waiting time for newly arrived tasks to be matched. In each snapshot, the planar Laplace-based privacy mechanism is applied to protect worker locations and ensure $\varepsilon $ -geo-indistinguishability. Meanwhile, the minimum-cost maximum-flow (MCMF)-based multitask matching mechanism is presented to maximize the task completion rate while minimizing the total travel cost. Experiments on real-world datasets demonstrate that PMS achieves superior task completion, reduced travel costs, and improved privacy preservation compared to existing algorithms.

External IDs:dblp:journals/iotj/WeiZCPS25