Delay-Aware Energy Minimization Offloading Scheme for Mobile Edge Computing

Offloading is regarded as a promising technology to reduce the delay and energy consumption of computation application in Mobile Edge Computing (MEC) network. By considering the demand of requesting user for low energy consumption and dynamic offloading environment, this paper proposes an offloading strategy by achieving the tradeoff between the energy consumption and time delay of computation application. Specifically, the computation offloading decision is first formulated as the finite horizon Markov decision process. Then, based on dynamic programming method, a Delay-Aware joint (Device-to-Device) D2D, MEC and Local Offloading (DADMLO) algorithm is proposed to get the optimal offloading policy which aims at minimizing the energy consumption of requesting user before deadline. Simulation results demonstrate that compared with heuristic schemes, the proposed strategy can complete the computation application with a higher completion probability and lower energy consumption.