Managing networks and services based on artificial intelligence in heterogeneous broadband environments for 5th generation and beyond
Μπελικαΐδης, Ιωάννης - Πρόδρομος
Belikaidis, Ioannis - Prodromos
KeywordsΔιαχείριση δικτύων και υπηρεσιών ; Διαχείριση πόρων ραδιοσυχνοτήτων ; Τεχνητή νοημοσύνη ; Νοητός διαχωρισμός δικτύων ; Περιβάλλοντα 5ης γενιάς και ακόλουθων γενεών ; Δίκτυα πλέγματος ; Ποιότητα υπηρεσιών ; Ποιότητα εμπειρίας χρήστη ; Υπερ-πυκνά δίκτυα
The motivation for this research activity comes from the exponential increase in data demand observed in recent years as well as the approximation of the theoretical limits of network capacity. At the same time the process of licensing and acquiring new frequencies is expensive, time consuming and in some cases licensing is impossible for many providers. In combination with the support of the large volume of data caused by the new services and applications, it becomes necessary to research, implement and test the various solutions proposed for the optimal management of networks. The dissertation is structured in chapters and each chapter provides a detailed description and results of the conducted research activities for achieving the overall goal of managing networks and services based on artificial intelligence in heterogeneous broadband environments for 5th generation and beyond. Chapter 1 provides the main introduction and motivation of our work and sets the requirements of the necessary research for managing networks and services based on artificial intelligence. Chapter 2 investigates and evaluates mesh networks for achieving enhanced performance, even when connectivity is challenging. Chapter 3 focuses on the modeling and analysis of management for heterogeneous infrastructure. Chapter 4 elaborates on radio frequency resource management in a multi-provider environment with emphasis on hierarchical radio resource management scheme. Chapter 5 discusses the notion of network slicing and how slicing can lead to better service provisioning in demanding environments by blending different traffic types (e.g. URLLC, eMBB etc.). Chapter 6 provides useful insights on sharing and allocation of resources with emphasis on dynamic channel assignment for the new 5G services such as URLLC, eMBB and mMTC. Chapter 7 elaborates on RRM issues in a multi-connection environment with emphasis on 5G component carrier management and is being evaluated with system level simulations. In addition, the requirements of ultra-high connection density of IoT devices in mMTC environments are evaluated. The research of the proposed solutions resulted in various publications, conference papers, journals, books and in standards such as ITU for a study regarding evaluation of 5G. The last chapter provides the conclusions of this dissertation and suggestions for future research. Specifically, this thesis presents an in-depth analysis of solutions for better management of networks and services based on intelligent algorithms and techniques in heterogeneous broadband environments of 5th and next generations, taking into account the current situation and new challenges of networks. The aim was to analyze, propose and improve the state of the art techniques to bring 5G networks one step closer to what we think it could be. D2D / M2M, Mesh Networks, Ultra-densification, Dynamic selection channel, Network sharing, Network slicing, Multi-connectivity / Multi link, Carrier Aggregation and the utilization of new frequencies such as 3.5 GHz and narrow band are some of the techniques that were analyzed, improved and presented, achieving good results in all tested cases. These proposals could be further tested in a real environment and formulated by the research community and standards to be integrated into the new versions of 3GPP for 5G networks and future generations.