Abstract
Mobile communications are growing and the number of users is constantly increasing at an accelerated rate, as well as the demand for the services they request. In the last few years, many efforts have focused on the design and deployment of the new fifth generation (5G) cellular networks. However, novel highly demanding applications, which are already emerging, need to go beyond 5G in order to meet the requirements in terms of network performance. But, at the same time, as the Earth does not allow us to increase the carbon footprint anymore, the energy consumption of the communication networks has to be critically taken into consideration. A multi-objective approach for addressing all these issues is therefore required. This work develops a cellular network framework that allows the evaluation of different system parameters over dynamic traffic patterns, as well as optimizing the different conflicting objectives simultaneously. The novelty relies on that the optimization process integrates key performance indicators from different layers of the network, namely the radio and the network layers, aiming at reaching solutions that account for the power consumption of the base stations, the total capacity provided to mobile users and also the signaling cost generated by handovers. Moreover, new metrics are needed to evaluate different solutions. Starting from the well-known energy efficiency merit factor (bits/Joule), three new merit factors are proposed to classify the network performance since they take into account several network parameters at the same time. These indicators show us the ideal working point that can be used to plan the point of operation of the network. These operation points are a medium-high power and capacity load and a low signaling load.
Highlights
In 2019, the monthly mobile traffic reached 38 Exabytes, while this figure is estimated to be 160 Exabytes by 2025, at a 30 percent compound annual growth rate
In order to provide the results with confidence, 30 independent runs of the multi-objective algorithms (MOEAs) have been carried out
The results of this procedure have shown that all the differences are statistically significant, pointing out that the hypervolume indicator (HV) of SMPSO is statistically greater than that of NSGA-II and the MatLab algorithms
Summary
In 2019, the monthly mobile traffic reached 38 Exabytes, while this figure is estimated to be 160 Exabytes by 2025, at a 30 percent compound annual growth rate. This work targets these latter scenarios using a novel approach that jointly optimizes several performance criteria from both the radio and the network layer, and considering energy efficiency issues [11]. Several previous works related to the physical layer are, for example, the work of Zhang et al [18] They have proposed a joint power allocation, mode selection, and channel assignment scheme for optimizing energy efficiency in D2D (Device-to-Device) communications. Works related to the network layer are, for example, the work of Keshavarzian et al [24] They have introduced several algorithms to minimize the energy consumption taking into account the mobility-aware capability.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
