Energy efficiency optimisation in cellular networks.

Abstract

It is highly expected that soon there will be environmental and economic negative implications from the amount of energy consumed by wireless network devices. Therefore, many researchers have paid attention towards addressing these challenges to investigate the impact of these wireless networks on both environment and the economy. The continuous operation of base stations even during low peak hours contributes to this issue. This research investigated sleep mode management based on q learning a Reinforcement Learning model free algorithm to find optimal tradeoff between energy consumption and quality of service constraint in a heterogeneous network architecture where small cells can switch to different sleep mode levels to save energy without impacting quality of service. In this regard, the algorithm intelligently learns from the environment based on the co-channel interference, the cell buffer size and the cell traffic level in order to decide the best sleep mode policy. A power consumption model for small base stations was developed incorporating multi sleep modes with varying deactivation levels. All simulations were done in Matlab. The simulation results showed that intelligent base station switching mechanism significantly reduced un necessary power consumption during off peak period by about 60.35% , high user throughput and low transmit delay were achieved. In conclusion leveraging AI based technologies in cellular networks can address the optimal tradeoff between energy consumption and Qos aligning with emerging 5G technological requirements which include high data rate, low latency and energy efficient networks. Extending the model with actor-critic reinforcement learning can be adopted for better adaptability.