Optimization of Energy-Efficient Cloud Radio Access Networks for 5G using Neural Networks

Abstract

Since proposed, Cloud Radio Access Network (Cloud-RAN) gives a committed architecture suitable for fulfilling 5G networks' applications. Cloud-RAN can solve challenges related to ever-evolving networks' mobile operators and an ever-growing number of end-users. Cloud-RAN architecture maintains both profitability and quality of service (QoS) . In this paper, power consumption is jointly formulated as power minimization beamforming and RRHs selection problem. Using the conventional convex or heuristic optimization approaches to find optimal solutions is highly complex; hence, we introduce an Artificial Neural Network (ANN) - based optimization model that aims to optimize the active Remote Radio Heads (RRHs) numbers in remote network sites and the consumed power. The proposed model considers various signal to interference plus noise ratios per client and power consumption models. Specifically, the model uses an adopted Bi-Section Group Sparse Beamforming (GSBF) optimization algorithm to reach near optimum solutions. Obtained validated results encourage machine learning techniques to reduce both the complexity and power consumption in such an emerging area.