Ultra-Dense Networks for 5G and Beyond

9Big Data Methods for Ultra‐dense Network Deployment

Weisi Guo¹, Maria Liakata², Guillem Mosquera³, Weijie Qi⁴, Jie Deng⁵ and Jie Zhang⁴

¹School of Engineering, University of Warwick, UK

²Department of Computer Science, University of Warwick, UK

³Mathematics Institute, University of Warwick, UK

⁴Department of Electronic and Electrical Engineering, University of Sheffield, UK

⁵School of Electronic Engineering and Computer Science, Queen Mary University of London, UK

9.1 Introduction

As we accelerate into the twenty‐first century, we are seeing increased human digital activity through the compounded effects of urbanization and the proliferation of smart devices (see Figure 9.1). Data demand has risen at a super‐linear rate, with a growth from 2011 to 2017. Beyond these statistical trends, networks are also experiencing increased data demand complexity from people and machines. Enabling understanding and exploitation of these complexities is important. This is especially so for the deployment of ultra‐dense network (UDN) nodes, which include small cells, relays, distributed antennas/reflectors, and other heterogeneous network elements deployed in a spatially dense formation [1]. Besides the wireless notions of cell planning, economically efficient UDN deployment requires precise knowledge of traffic patterns and usage contexts, as well as the shortfalls in existing deployment to reduce ...