Resource Allocation Strategy for UAV-assisted Non-linear Energy Harvesting MEC System

Abstract

Aiming at the problem that the user devices (UDs) far from the base station (BS) suffer from the doubly near-far effect, a resource allocation strategy for unmanned aerial vehicle (UAV)-assisted non-linear energy harvesting mobile edge computing (MEC) system is investigated in this paper. By deploying a UAV, remote user devices can harvest energy from both the BS and the UAV in the non-linear energy harvesting way, but only offload computing tasks to the MEC server on the UAV to reduce the influence of the doubly near-far problem. The nearby user devices only harvest energy from the BS in the non-linear energy harvesting way and offload computing tasks to the MEC server on the BS. The resource allocation problem is modeled as a non-linear programming problem by jointly optimizing the UDs’ local computing capability, UDs’ transmitting power, and the flight trajectory of the UAV. The objective is to maximize the sum of computation bits of all UDs under the constraints of energy consumption of UDs and UAV and the speed of the UAV. The suboptimal solution is obtained by introducing the differential evolution (DE) algorithm. Simulation results show that compared with the partial optimization methods of transmitting power, local computing capability, and the flight trajectory of the UAV based on the DE algorithm, the proposed method has different degrees of increase in the sum of computation bits of all UDs.