Crowdsensing Data Trading Mechanism Based on Personalized Local Differential Privacy

To address the problem that only the same level of privacy protection is provided for all users in crowdsensing data trading, we propose a personalized local differential privacy-based crowdsensing data trading mechanism, which consists of two parts: privacy protection model and user selection. In the privacy protection model, we design a data perturbation algorithm to implement personalized local differential privacy, which can perturb any numerical data according to the user’s privacy budget. In terms of user selection, we establish a user selection optimization problem with the objectives of minimizing data aggregation error and minimizing payment cost, and propose an Enhanced Non-dominated Sorting Genetic Algorithm-II (ENSGA-II) to solve it. Finally, experimental results show that the ENSGA-II has better performance than other algorithms.