P2-TaskMP: Privacy-Preserving Task Allocation Optimization Based on Mobility Prediction

Abstract

The emergence of Mobile Crowd Sensing (MCS) has provided a new paradigm for data sensing. An effective task allocation can ensure the stability and efficiency of the system in MCS. In this paper, we propose a privacy-preserving multi-objective, multi-task allocation optimization scheme, P2-TaskMP (Privacy-Preserving Task Allocation Optimization based on Mobility Prediction), to solve the multi-objective optimization task assignment problem while preserving users’ privacy. The scheme evaluates participants’ task completion capabilities by introducing mobility prediction based on fuzzy logic, which then informs task pre-allocation to form the initial population, unlike traditional methods that initialize populations randomly. To address potential privacy leaks of participants’ spatiotemporal information during mobility prediction, we employ differential privacy techniques to add Laplace noise to participants’ historical trajectory records, achieving adequate privacy protection. P2-TaskMP achieves Pareto-optimal solutions using the NSGA-II-DE (Non-dominated Sorting Genetic Algorithm II with Differential Evolution) algorithm and realizes satisfactory results with fast solution speed for large-scale task allocation problems. Simulations conducted on two real-world datasets demonstrate that our proposed method achieves higher accuracy, and the task allocation algorithm performs better than the compared algorithms in maximizing task completion rate and minimizing cost.