Leveraging LEO Assisted Cloud-Edge Collaboration for Energy Efficient Computation Offloading
Zhixuan Tang, Haibo Zhou, Ting Ma, Kai Yu, Xuemin Shen
Abstract
Mobile edge computing (MEC) has been widely considered as an effective technology to handle computationally intensive tasks generated by mobile devices. However, the computation resources at an edge node is usually several orders of magnitude smaller than that of a cloud. Thus, it is rather vital to take an investigation into the collaboration between the cloud and the edge. In this paper, to fully exploit the computation power of the cloud server and achieve energy efficient task offloading, we propose an LEO-assisted terrestrial-satellite network (TSN) architecture for cloud-edge collaborative computation offloading. We formulate the collaborative cloud-edge computing problem that minimizes the energy consumption of the whole TSN under the quality-of-service (QoS) constraints. The optimization problem is further decomposed into two subproblems which are solved by deep neural networks (DNN) and successive convex approximation (SCA) algorithm, respectively. Simulation results show the effectiveness of our proposed cloud-edge collaborative computation offloading architecture on achieving a lower energy cost.