A DRL Framework for Data Offloading and Energy Management in MEC Networks with Distributed Power Source

As the number of Internet of Things (IoT) devices rises rapidly, the data produced by these devices has dramatically surged. Deploying distributed power sources integrated with edge servers in IoT networks is an efficient approach to address the intrinsic bottlenecks of traditional IoT devices, limited computing ability and energy capacity. In this paper, we consider a scenario consisting of mobile edge computing (MEC), a distributed power source and smart grid, where the distributed power source can provide energy for devices and trade with the power grid. We aim to maximize the computing rate and weighted energy trading benefits by optimizing offloading decisions and energy resource allocation. We formulate this problem as a mixed integer nonlinear programming (MINLP) problem and decompose it into an equivalent master problem and sub-problem. We propose a Deep Reinforcement Learning (DRL) method to obtain the offloading decision and utilize convex optimization to derive the optimal energy resource allocation. The experimental results indicate that the method proposed in this paper outperforms existing algorithms in terms of convergence and performance.