Optimization of Mobile MEC Offloading with Energy Harvesting and Dynamic Voltage Scaling
Yuandong Zhuang, Xi Li, Hong Ji, Heli Zhang · 2019
Mobile edge computing (MEC) is an emerging technology to support mobile user equipments (UEs) to implement computing-sensitive applications by offloading the computation tasks to the MEC server. However, due to the random movement of the UEs, the offloading procedure and the return of computation results may be interrupted when the UEs move out of the original MEC server's wireless communication coverage. In this paper, we make use of mobile network entities and powerful mobile terminals to form mobile MEC servers, which could provide computing offloading service for surrounding mobile UEs. The mobile MEC server is powered by battery with limited energy. In order to prolong its life time, we introduce the energy harvesting (EH) and dynamic voltage scaling (DVS) techniques. Then based on the requirements of UEs, we discuss the problem of how to jointly optimize the computation speed of mobile MEC server and corresponding power allocation. It is modeled as a nonconvex problem, aiming to minimize the end-to-end computation offloading latency under energy restriction. Then we adopt variable substitution method to transform it into a convex problem, and further propose a latency-optimal computation offloading algorithm to obtain optimal solution. Simulation results show that our proposed algorithm may significantly reduce the energy consumption and the latency.