Research On Key Technologies For Next Generation Mobile Networks

With the rapid development of mobile communication technologies,an immense growth of traffic demands becomes a great challenge for the mobile networks,as well as the consequential increase of network cost and energy consumption.As a result,how to reduce the network cost and mitigate energy consumption while meeting the ever-increasing of traffic demands becomes one of the most important issues for the next generation mobile networks.To address the above-described challenges and sup-port the sustainable development of mobile networks,many key technologies for the next generation mobile networks,including densely deployed heterogeneous networks(HetNets),massive MIMO,device-to-device(D2D)communication,cloud radio ac-cess network(C-RAN),software defined network(SDN)and network function virtu-alization(NFV),have been proposed and received a lot of attention in both academia and industry.In this thesis,we try to lower the network cost,reduce the energy consumption,enhance the user access capability and increase the system capacity in the next genera-tion mobile networks.We mainly focus on the following key technologies for the next generation mobile networks:Approximation algorithms for network planning in Het-Nets,remote radio head(RRH)sleeping and user association in C-RANs,and resource sharing for D2D communication underlaying cellular networks.The major contribu-tions of this thesis are summarized as follows:1.We study the network planning problem in HetNets.Our optimization task is to select a subset of candidate access points(APs),including macro base stations(BSs)and low-power APs,to minimize the total cost of ownership(TCO)and ener-gy consumption of the networks while satisfying necessary constraints.We introduce approximation algorithms to cope with two different network planning scenarios,both of which are NP-hard.First,we discuss the macro-only network planning scenari-o,for which our proposed algorithm achieves an approximation ratio of O(log R),where R is the maximum achievable capacity of the macro BSs.Then we introduce an O(log R)-approximation algorithm to the HetNet planning scenario,where R is the maximum achievable capacity of the macro BSs and the low-power APs overlaid with these macro BSs.Simulation results indicate that the proposed network planning schemes can significantly reduce the TCO and energy consumption of the considered networks.2.We also investigate the RRH sleeping and user association in C-RANs.First,we focus on the minimization of total power consumption of the C-RAN and try to achieve a flexible tradeoff between the transport network power consumption and the transmission power of the RRHs.We study the RRH sleeping problem based on traffic density,which is formulated as a generalized form of the classical capacitated facility location problem(CFLP).An efficient local search algorithm is developed to address the formulated problem.Specifically,we propose three types of local improvemen-t operations,including "add","open " and close”operations,to search the local optimum.Then,we study the user association problem in the C-RAN,where the optimization task is to maximize the number of satisfied users.We propose a 1/2-approximation algorithm to address the problem.Different from the conventional user association strategies,our proposal takes account of channel gain,user's rate re-quirement and available bandwidth.Simulation results show that our user association strategy is promising for practical implementation in different scenarios.3.We study the resource sharing schemes for D2D communication underlay-ing cellular networks.Different from most of existing literature,our resource sharing schemes exploit the spatial reuse of the D2D pairs to further increase the system capaci-ty,where multiple D2D pairs can share same subchannel with single cellular user(CU).We first propose a subchannel sharing protocol to suppress the mutual interference a-mong the D2D pairs to a negligible level.Based on the analysis results,we develop an efficient subchannel allocation scheme by employing a simple greedy strategy,as well as a power distribution algorithm.Then,we propose an alternating optimization method to maximize the system capacity by alternately optimizing the subchannel al-location and power distribution.Finally,simulation results verify that our proposed resource sharing schemes can suppress the mutual interference among the D2D pairs and enhance the system capacity efficiently and effectively.In summary,several key technologies for the next generation mobile networks are deeply studied in this thesis.With careful theoretical analysis and research,the operat-ing cost is reduced,the energy consumption is mitigated and the system performance is also enhanced.Our proposed schemes are promising for practical implementation in the next generation mobile networks.