Resumen de Design of static intercell interference coordination schemes for realistic lte-based cellular networks
David González González
Today, 3.5 and 4G systems including Long Term Evolution (LTE) and LTE-Advanced (LTE-A) support packet-based services and provide mobile broadband access for bandwidth-hungry applications. In this context of fast evolution, new and challenging technical issues must be e ectively addressed. The nal target is to achieve a signi cant step forward toward the improvement of the Quality of Experience (QoE). To that end, interference management has been recognized by the industry as a key enabler for cellular technologies based on OFDMA. Indeed, with a low frequency reuse factor, intercell interference (ICI) becomes a major concern since the Quality of Service (QoS) is not uniformly delivered across the network, it remarkably depends on user position. Hence, cell edge performance is an important issue in LTE and LTE-A. Intercell Interference Coordination (ICIC) encompasses strategies whose goal is to keep ICI at cell edges as low as possible. This alleviates the aforementioned situation. For this reason, the novelties presented in this Ph.D. thesis include not only developments of static ICIC mechanisms for data and control channels, but also e orts towards further improvements of the energy e ciency perspective. Based on a comprehensive review of the state of the art, a set of research opportunities were identi ed. To be precise, the need for exible performance evaluation methods and optimization frameworks for static ICIC strategies. These mechanisms are grouped in two families: the schemes that de ne constraints on the frequency domain and the ones that propose adjustments on the power levels. Thus, Soft- and Fractional Frequency Reuse (SFR and FFR, respectively) are identi ed as the base of the vast majority of static ICIC proposals. Consequently, during the rst part of this Ph.D. thesis, interesting insights into the operation of SFR and FFR were identi ed beyond well-known facts. These studies allow for the development of a novel statistical framework to evaluate the performance of these schemes in realistic deployments. As a result of the analysis, the poor performance of classic con gurations of SFR and FFR in real-world contexts is shown, and hence, the need for optimization is established. In addition, the importance of the interworking between static ICIC schemes and other network functionalities such as CSI feedback has also been identi ed. Therefore, novel CSI feedback schemes, suitable to operate in conjunction with SFR and FFR, have been developed. These mechanisms exploit the resource allocation pattern of these static ICIC techniques in order to improve the accuracy of the CSI feedback process. The second part is focused on the optimization of SFR and FFR. The use of multiobjective techniques is investigated as a tool to achieve e ective network-speci c optimization. The approach o ers interesting advantages. On the one hand, it allows for simultaneous optimization of several con icting criteria. On the other hand, the multiobjective nature results in outputs composed of several high quality (Pareto e cient) network con gurations, all of them featuring a near-optimal tradeo between the performance criteria. Multiobjective evolutionary algorithms allow employing complex mathematical structures without the need for relaxation, thus capturing accurately the system behavior in terms of ICI. The multiobjective optimization formulation of the problem aims at achieving e ective adjustment of the operational parameters of SFR and FFR both at cell level and network-wide. Moreover, the research was successfully extended to the control channels, both the PDCCH and ePDCCH. Finally, in an e ort to further improve the network energy e ciency (an aspect always considered throughout the thesis), the framework of Cell Switch O (CSO), having close connections with ICIC, is also introduced. By means of the proposed method, signi cant improvements with respect to traditional approaches, baseline con gurations, and previous proposals can be achieved. The gains are obtained in terms of energy consumption, network capacity, and cell edge performance.
