Joint Uplink and Downlink Resource Allocation for UAV-enabled MEC Networks under User Mobility

This paper investigates a unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) system with stochastic user mobility, time-varying wireless channel, and computational task arrival. An optimization problem is formulated to minimize the weighted-sum energy consumption of the ground user and the UAV, subjected to constraints on the limited processing capability and transmit power of both sides, the total system bandwidth, and the stability of task queues. Local computation at the user and the volume of offloaded tasks are jointly optimized with the UAV's uplink/downlink bandwidth and remote processing power allocation to minimize the system's energy. Due to the time coupling of optimization variables, the Lyapunov framework is adopted to transform the original multistage problem into deterministic problems that can be solved in each time slot. The per-time-slot problem is then decomposed into three manageable sub-problems that are designed for different variables. An online algorithm is proposed to iteratively solve the problems. Numerical results show that the proposed approach outperforms benchmark schemes in terms of energy efficiency for the whole system while satisfying the queue stability.