Research On Energy Efficient Service Control Strategy Based On Large-scale User Behavior In Wireless Networks

With the rapid development of information and communication technology (ICT), the wireless networks satisfy the explosive increase of user traffic demands while their power consumptions are rising sharply as well. Green communication technologies have been researched extensively and intensively recently. Energy efficient service control strategies which considered as the most promising method for improving the network energy efficiency is becoming the hot issue in wireless communications. In this dissertation, we want to solve the mismatch between wireless resource allocation and traffic demands in the time dimension, the imperfection of large-scale user behavior model in spatial dimension and how to use the multiformity of user requirement to solve the insufficient of wireless resource in next generation wireless networks.Through the employ of the existing large-scale user behavior model and make improvements, we want to propose energy efficient service control theory and methods based on the heterogeneity of traffic demands in the time/spatial/ content dimensions and thus provides necessary theoretical support to real networks and next generation wireless networks. The main research work and contributions of this dissertation are summarized as follow:1. Research on wireless resource allocation for time heterogeneity of traffic demandsConsidering the time heterogeneity of traffic demands and large-scale user behavior model, we research the resource allocation technology in wireless networks. User traffic demands have fluctuation law and we want to propose some energy efficient base station (BS)sleeping strategy and coverage control strategy to improve network energy efficiency. The traditional resource allocation schemes do not consider the influence of time heterogeneity of traffic demands. Therefore,we propose a novel BS sleeping strategy based on the usage rate of subchannels and a coverage control strategy which joint downlink and uplink communications. The closed-form expressions of transmission success probability (TSP) and energy efficiency (EE) are derived. There is an optimal sleeping threshold and optimal coverage bias can maximize the EE, respectively. The proposed sleeping strategy can improve more than 15.5% EE than traditional random sleeping strategy and the proposed coverage control strategy can improve 5% EE than no coverage control strategy. On this basis, we propose three service control strategies based on user behaviors. The optimal spectrum allocation and service control schemes under different convergence coefficient are obtained.This research provides new insights into the mismatch between wireless resource allocation and traffic demands in the time dimension.2. Research on service control strategy for spatial heterogeneity of traffic demandsNetworks are divided into the hotspot and non-hotspot areas in traditional spatial heterogeneity model. This model cannot depict the aggregation when there are many hotspots in the network. A spatial heterogeneity model is defined to quantitative characterize the heterogeneity of traffic demands in this dissertation and the expression of spatial aggregation coefficient is derived. This research is the supplement and perfection of traditional study of spatial heterogeneity of traffic demands. Then, an edge-aware cross-tier cooperation scheme is proposed to improve the performance of edge hotspot users. Stochastic geometry is utilized to derive the TSP and EE of the proposed scheme. Numerical results show that there has a tradeoff between TSP and EE. The proposed scheme can maximize the network performance by an optimal cooperation threshold. Two user behaviors are considered and we find that user behavior has enormous influence on EE when users very gathered.3. Research on joint optimal association and resource allocation for heterogeneity of traffic demandsUsers in different areas have various traffic demands. Wide area with low data requirement scenario and hotspot area with high data requirement scenario are typical scenarios in next generation networks.This dissertation research how to jointly control the user association and resource allocation can maximize EE. Firstly, the upper bound and lower bound of TSP in two-tier heterogeneous networks is derived when using the orthogonal spectrum allocation strategy. Secondly,a network energy consumption minimization problem which considering the TSP constraints and jointly determines the optimal density of micro BS and the optimal association bias is formulated. The closed-form solution of the optimal micro BS density and association bias are derived for the first time. Finally,a rate coverage maximization problem by adjusting the spectrum allocation and user association is investigated in hotspot area scenario. A dynamic gradient iterative algorithm is proposed to solve the maximization problem. Simulation results show that optimal associationbias is a negative exponential function of user density (i.e.?opt?O(?u-a/2)and the optimal micro BS is a linearly monotone increasing function of the user density (i.e.?2opt?O(?u))in the wide area. In hotspot area,the optimal association bias is inversely proportional to user traffic demand when the density of micro BS very small but the optimal association bias is proportional to user traffic demand when micro BS density very large.The optimal spectrum allocation ratio decrease with the increase of micro BS density and user rate demands have little effect on the optimal spectrum allocation ratio.