Research On Resource Management And Performance Analysis In Heterogeneous Networks

With the ever increasing growth of the number of smart devices and the popularity of diversified mobile multimedia applications,the demands for high data rates explosively increase.The fifth generation(5G)radio access networks need to significantly improve the spectrum utilization to meet the tremendous traffic demand.Network densification through increasing the density of base stations(BSs)to improve spectrum utilization is envisioned as an important solution for the 5G networks.Considering that the high capital expenditure and operating expenditure,heterogeneous network(Het Nets)with dense deployment of low-power access nodes(LPNs),such as femto and pico BSs,are recognized as a cost-effective solution to increase spectrum resource utilization.Due to the scarcity of spectrum resources,heterogeneous networks in co-channel deployment is envisioned as a key solution to the evolution of cellular networks.However,the dominant inter-tier interference and load imbalance significantly hinder the performance of HetNets,which must be addressed by practical resource management techniques.Although academia and industry have proposed relevant resource management techniques to improve network coverage and spectrum efficiency of Het Nets,the impact of parameters on Het Net performance and the optimal resource allocation solutions are still not clearly quantified.Moreover,due to the huge energy consumption caused by the intensive deployment of low-power BSs,it is of great practical significance to study energy-efficient resource management strategies to reduce emissions.In addition,the ideal hexagonal grid model is no longer applicable to Het Nets due to the random deployment of low power BSs,and the performance evaluation by system level simulation is highly complex.Modeling more general and tractable frameworks to analyze the performance of Het Nets is desired.In response to the above questions,in this dissertation,stochastic geometry is applied to model general frameworks to study resource management techniques and the performance of sparse and dense Het Nets,respectively,in terms of the Signalto-Interference-plus-Noise-Ratio(SINR)coverage,data rate and energy efficiency.The main contributions are summarized as follows.Firstly,performance analysis of cell range expansion(CRE)and subchannel allocation in backhaul constrained Het Nets.In order to promote load balancing thereby fully utilizing small cell resources in Orthogonal Frequency Division Multiple Access(OFDMA)Het Nets,the cell range expansion technique is introduced.Meanwhile,frequency domain resource partitioning is used to mitigating the cross-tier interference.To capture the randomness of BS locations and obtain tractable analytical results,we develop a general framework for Het Nets with backhaul constraints and reasonably model network interference for OFDMA Het Nets,while taking cell load and general channel model into account.Stochastic geometry is applied to derive the expressions of SINR coverage,rate coverage and energy efficiency.The analytical results are validated by Monte Carlo simulation.We further investigate the effect of association bias and resource partitioning fraction on Het Net performance.Subsequently,we investigate the optimal association bias and resource partitioning fraction in terms of rate coverage,and the optimal BS density that the maximization of energy efficiency(EE)can be achieved.Secondly,Joint optimization of resource allocation in Het Nets with Fe ICIC.We develop a general model considering jointly the user association,power control and resource partitioning fraction for OFDMA Het Nets with further enhanced inter-cell interference coordination(Fe ICIC)and CRE.Stochastic geometry is applied to derive the expressions of SINR coverage and rate coverage in two-tier Het Nets with limited backhaul capacity.The analytical results are validated through Monte Carlo simulation.We further analyze the effect of each parameters on the performance of Het Nets in terms of rate coverage and investigate the optimal configuration of resource allocation.In addition,we introduce a load-aware spectrum allocation and BS sleeping technique for a given traffic demand,and derive the expressions of spectral and EE.In order to create spectral and energy efficient Het Nets,we solve the multi-objective optimisation problem formulated for the SE-EE tradeoff and obtain the optimal solution through proposed Dinkelbach and Gradient descent algorithms.Thirdly,Performance analysis of Het Nets with location-aware cross-tier cooperation(CTC).In order to address the serious co-channel interference problem for the cellular edge region,we propose a location-aware CTC scheme for Het Nets with sparse deployment of LPNs,and develop a general model of Het Nets with location-aware CTC,while introducing cooperation factors to control the cooperation range.Stochastic geometry is applied to derive expressions of SINR coverage and ergodic capacity,and verified the accuracy of analytical results and the effectiveness of the CTC scheme.We further investigate the effect of network parameters on the performance of Het Nets,and obtained the optimal cooperation factors to provide system design insights for Het Nets with CTC.In addition,user association in Het Nets with cross-tier BS clustering and cooperation.To cope with the cross-tier and co-tier interference in ultra-dense Het Nets(UDHs),we propose a user-centric cross-tier clustering and cooperation scheme.Stochastic geometry is applied to derive expressions of SINR coverage and ergodic capacity(EC)in UDHs with cluster cooperation.Furthermore,we quantify the requirements of backhaul capacity for a typical BS,and provide analytical results of successful serving probability(SSP)and effective ergodic capacity(EEC)for the clustering and cooperation in UDHs.The effect of the deployment density of BSs,backhaul capacity and cluster size on the SSP and EEC are analyzed and further propose the optimal clustering strategy.