Pareto-Optimal Aerial-Ground Energy Minimization for Aerial 3D Mobile Edge Computing Networks

Unmanned aerial vehicles (UAVs) are a crucial integrated component in mobile edge computing (MEC) for extending service coverage, where UAVs' and devices' energy consumption are two fundamental problems. In this paper, we consider a UAV-enabled three-dimensional (3D) MEC network and study the Pareto-optimal aerial and ground energy minimization problem by jointly optimizing offloading time allocation, resource allocation, and UAV 3D trajectory. We first investigate the problem of minimizing energy consumption for the UAV, and then shift our focus to minimize the energy consumption of the devices. To tackle them, we propose two efficient alternating optimization algorithms with different outer and inner structures to address the two optimization problems. To be specific, we split the original problem into two subproblems in the outer structure. The inner structure addresses the subproblems with difference-of-convex (D.C.) framework and successive convex approximation (SCA) technique. Based on the energy consumption lower bounds for the UAV and devices obtained by these two aforementioned original problems, a Pareto-optimal solution is proposed, in which the aerial-ground energy tradeoff can be balanced. Extensive simulation results validate the effectiveness of our proposed scheme and reveal two insightful tradeoffs for 3D UAV-enabled MEC networks, i.e., aerial-ground energy tradeoff and altitude-related angle-distance tradeoff.