Green hybrid energy harvesting for intelligent mobile edge computing in internet of things

Abstract

Mobile device (MD) is usually energy constrained and the computation task will be interrupted when the battery power is run out. Intelligent mobile edge computing (MEC), as a promising technology in internet of things (IoT), can effectively save the computational energy of MD. Meanwhile, the hybrid energy harvesting (HEH) technology can enable MD to harvest available energy from the surrounding environment, so as to achieve the goal of green computing. According to the above mentioned, we integrate HEH technology into MEC to solve the limited energy problem of MD in this paper. Besides, to maximize the system utility (SU), a SU model with three measurable indicators of the latency, remaining energy and task success rate is proposed. Then, we improve the deep deterministic policy gradients (DDPG) algorithm and propose twin delayed deep deterministic policy gradient (TDDDPG) algorithm to obtain the suboptimal solution of proposed model. Eventually, the simulation results show that the TDDDPG can obtain the optimal SU compared with local computing (LC) algorithm, edge server computing (ESC) algorithm, random offloading (RO) algorithm and deep deterministic policy gradient (DDPG) algorithm.