Intelligent Energy Efficiency Algorithm for the 5G Dense Heterogeneous Cellular Networks

摘要

Over the past years, telecom network operators have seen an endlessly growing demand for ubiquitous highspeed wireless access and an exceptional increase in connected wireless devices. As a result, we have seen a high growth in traffic volumes. The fifth-generation (5G) heterogeneous cellular networks (HetNets) have been developed by telecom network operators to meet the growing mass data capacity. The 5G dense HetNets is made up of multiple radio base stations (RBSs)/small base stations (SBSs) to increase the coverage and system capacity which led to a high number of network elements. Hence, significantly increases power consumption and lessening the energy efficiency for the telecom network operators. Energy efficiency algorithms have been developed for the 5G dense HetNets, however, these existing energy efficiency algorithms do not satisfy the throughput QoS requirements such as minimalized packet loss, longer battery lifetime, reliability, and high data rates. In addition, real-time traffic types such as voice and video requires high computational load at the terminal side, which have an undesirable impact on energy/battery lifetime which further affects the throughput QoS performance. As a result, this paper proposed an Intelligent Energy Efficiency (IEE) algorithm for throughput QoS and energy efficiency enhancement in 5G dense HetNets. In the proposed IEE algorithm, a deep neural network (DNN) was used to determine the cell capacity ratio for the SBSs. Hence, the SBSs cell capacity ratio was employed as decision criteria to put the SBSs into a sleep state. In addition, transferable payoff coalitional game theory was used in order to ensure real-time applications have a higher priority over non-real time applications. Numerous computer simulation results illustrated that the proposed IEE algorithm experienced an average packet loss of 2.6%, the energy consumption of 3.4 joules, and produced a network throughput of 97.4%.