A Privacy-Preserving Computation Offloading Method Based on Privacy Entropy in Multi-access Edge Computation
Xing Zhao, Jianhua Peng, Yingle Li, Haitao Li · 2020
Computation offloading decision in Multi-access Edge Computing (MEC) may expose users' characteristics and reveal users' privacy. In this paper a privacy-preserving computation offloading method based on privacy entropy is proposed. Firstly, the privacy exposure risk caused by the characteristics of tasks' offloading frequencies is studied, and privacy entropy based on the deviation of tasks' offloading frequencies is proposed as a quantitative analysis metric. Then, the privacy restriction is introduced into the offloading decision model with optimal energy consumption under delay constraint, ensuring that the privacy entropy does not exceed the privacy threshold, and a privacy-preserving computation offloading method based on privacy entropy is established. Finally, based on the genetic algorithm the offloading decisions that satisfy the privacy constraint and the optimal energy consumption target are solved, and the optimal decisions are used to train the artificial neural network's parameters to quickly solve the optimal decision in each timeslot. Simulation results show that the method can effectively reduce the average energy consumption of the terminal and make the offloading frequencies of tasks always meet the privacy constraint.