Resource Allocation and Service Management in Next Generation 5G Wireless Networks

Abstract

The accelerated evolution towards next generation networks is expected to dramatically increase mobile data traffic, posing challenging requirements for future radio cellular communications. User connections are multiplying, whilst data hungry content is dominating wireless services putting significant pressure on network's available spectrum. Ensuring energy-efficient and low latency transmissions, while maintaining advanced Quality of Service (QoS) and high standards of user experience are of profound importance in order to address diversifying user prerequisites and ensure superior and sustainable network performance. At the same time, the rise of 5G networks and the Internet of Things (IoT) evolution is transforming wireless infrastructure towards enhanced heterogeneity, multi-tier architectures and standards, as well as new disruptive telecommunication technologies. The above developments require a rethinking of how wireless networks are designed and operate, in conjunction with the need to understand more holistically how users interact with the network and with each other. In this dissertation, we tackle the problem of efficient resource allocation and service management in various network topologies under a user-centric approach. In the direction of ad-hoc and self-organizing networks where the decision making process lies at the user level, we develop a novel and generic enough framework capable of solving a wide array of problems with regards to resource distribution in an adaptable and multi-disciplinary manner. Aiming at maximizing user satisfaction and also achieve high performance - low power resource utilization, the theory of network utility maximization is adopted, with the examined problems being formulated as non-cooperative games. The considered games are solved via the principles of Game Theory and Optimization, while iterative and low complexity algorithms establish their convergence to steady operational outcomes, i.e., Nash Equilibrium points. This thesis consists a meaningful contribution to the current state of the art research in the field of wireless network optimization, by allowing users to control multiple degrees of freedom with regards to their transmission, considering mobile customers and their strategies as the key elements for the amelioration of network's performance, while also adopting novel technologies in the resource management problems. First, multi-variable resource allocation problems are studied for multi-tier architectures with the use of femtocells, addressing the topic of efficient power and/or rate control, while also the topic is examined in Visible Light Communication (VLC) networks under various access technologies. Next, the problem of customized resource pricing is considered as a separate and bounded resource to be optimized under distinct scenarios, which expresses users' willingness to pay instead of being commonly implemented by a central administrator in the form of penalties. The investigation is further expanded by examining the case of service provider selection in competitive telecommunication markets which aim to increase their market share by applying different pricing policies, while the users model the selection process by behaving as learning automata under a Machine Learning framework. Additionally, the problem of resource allocation is examined for heterogeneous services where users are enabled to dynamically pick the modules needed for their transmission based on their preferences, via the concept of Service Bundling. Moreover, in this thesis we examine the correlation of users' energy requirements with their transmission needs, by allowing the adaptive energy harvesting to reflect the consumed power in the subsequent information transmission in Wireless Powered Communication Networks (WPCNs). Furthermore, in this thesis a fresh perspective with respect to resource allocation is provided assuming real life conditions, by modeling user behavior under Prospect Theory. Subjectivity in decisions of users is introduced in situations of high uncertainty in a more pragmatic manner compared to the literature, where they behave as blind utility maximizers. In addition, network spectrum is considered as a fragile resource which might collapse if over-exploited under the principles of the Tragedy of the Commons, allowing hence users to sense risk and redefine their strategies accordingly. The above framework is applied in different cases where users have to select between a safe and a common pool of resources (CPR) i.e., licensed and unlicensed bands, different access technologies, etc., while also the impact of pricing in protecting resource fragility is studied. Additionally, the above resource allocation problems are expanded in Public Safety Networks (PSNs) assisted by Unmanned Aerial Vehicles (UAVs), while also aspects related to network security against malign user behaviors are examined. Finally, all the above problems are thoroughly evaluated and tested via a series of arithmetic simulations with regards to the main characteristics of their operation, as well as against other approaches from the literature. In each case, important performance gains are identified with respect to the overall energy savings and increased spectrum utilization, while also the advantages of the proposed framework are mirrored in the improvement of the satisfaction and the superior Quality of Service of each user within the network. Lastly, the flexibility and scalability of this work allow for interesting applications in other domains related to resource allocation in wireless networks and beyond.