Cognitive Green Backhaul Deployments for Future 5G Networks

This paper introduces a cognitive green topology management scheme for future 5G networks, which can be used to reduce energy consumption in low traffic scenarios. The scheme is based on a backhaul link selection algorithm which aims to concentrate distributed traffic on fewer backhaul links by exploiting backhaul link diversity from other cells. A reinforcement learning based resource assignment algorithm has been introduced to work in conjunction with the topology management scheme. It is shown that total energy consumption can be reduced by up to 35% with marginal Quality of Service compromises. In addition, the tradeoff between energy saving and control overhead is also explored in this paper.