Resource Management Of Multi-layer Heterogeneous Integrated Network For 5G

With the rapid development of new services,such as mobile Internet and Internet of Things,the global mobile data traffic experiences an exponential growth.The spatial-temporal distribution of the traffic is extremely non-uniform,and new services with different quality of service(QoS)requirements are emerg-ing,such as ultra-high definition video,virtual reality,which poses a great challenge to the capacity of wireless network.To improve the bearing capac-ity of wireless network,the number and type of base stations are intensively deployed.And to meet the non-uniform and diversified traffic demands,low-power heterogeneous wireless access points,such as small cell base stations and Wi-Fi,are further deployed within the coverage of macro base stations,which is forming overlapping,multi-layer and high-density heterogeneous in-tegrated network.Integrating these multiple radio access network can improve the resource utilization and satisfy the usersJ demand.However,ultra-dense networking results in smaller cell coverage and more complex and intensive inter-cell interference,which limits the improvement of the network capacity.On the other hand,the traditional multi-layer heterogeneous network has verti-cal independence and ineffective coordination between layers,which is simple networking mode.It results in serious inter-system interference,low resource utilization,and cannot guarantee end-to-end seamless QoS requirements.These have brought great challenges to the fifth generation mobile communication(5G)network.This dissertation makes an exploratory study on the resource management of multi-layer heterogeneous integrated network for 5G,and mainly deal with the following challenges:(I)The complex dynamic interference caused by the dense deployment severely restricts the corntinuous improvement of network capacity;(2)Heterogeneous multi-networks severely restrict the efficient uti-lization of wireless resources;(3)Existing networks are difficult to support the differentiated QoS requirements of mass devices.The main contributions and innovations of the dissertation are summarized as follows:(1)For the ultra-dense network(UDN)in the licensed spectrum,consider-ing the unbalanced traffic load caused by the dense deployment,the randomness of network interference caused by random arrival traffic is increasing.Thus,it is difficult to accurately characterize UDN performance due to the lack of consideration of random traffic interference and QoS.We construct a random interference model of UDN.The interference distribution under multiple in-tegrals is solved.Thus we derive the novel effective capacity of ultra-dense network.The effective capacity reveals the rule between capacity,bandwidth,network density and QoS requirements.Based on the non-cooperative game theory and the proposed effective capacity,a traffic control algorithm is fur-ther proposed to maximize total average arrival traffic with QoS guaranteeing in UDN.Compared with saturated traffic scheme,the supported traffic arrival rate of the proposed algorithm has been improved by 3 3%.(2)For the UDN in the unlicensed spectrum,considering the random ac-cess mechanism of unlicensed spectrum,the problem of packet collision in unli-censed spectrum under the ultra-dense network is more serious,which makes it more difficult to guarantee QoS.We develop a new four-state semi-Markovian model to capture the random access mechanism of Licensed-Assisted Access(LAA),and propose the effective capacity.Two power control algorithms are proposed to maximize effective capacity and effective energy efficiency,re-spectively.Finally,numerical results and Monte Carlo simulation results are carried out to verify the accuracy of the proposed theory.The capacity-delay domain of ultra-dense distributed networks is given.The simulation results show that compared with existing schemes,the effective capacity and effective energy efficiency can be improved by 62.7%and 171.4%,respectively.(3)In heterogeneous integrated networks,different wireless access net-works have great differences in terms of transmission capability,networking mode,and the ability to guarantee the quality of service requirements,such as LTE in the licensed spectrum,Wi-Fi and LAA in the unlicensed spectrum,macro base stations with high transmission power and small cell base stations with low transmission power.How to quantify the capacity of heterogeneous resources under QoS guaranteeing in multiple access networks is still difficult in the current 5G network.For the LTE and WiFi Aggregation(LWA)system,we propose a theoretical framework based on effective capacity,which makes an original stream with the QoS requirement decouple into multiple sub-streams with different QoS in adaptation to heterogeneous air interfaces.Then,a multi-band resource allocation with the QoS guarantee between licensed and unli-censed spectrum is proposed based on Block Coordinate Descent(BCD)and difference of two convex functions(D.C.)programming.Simulation results shoxw that the proposed algorithm saves 16.89%and 30.2%of the bandwidth in the licensed spectrum under the QoS guaranteeing,respectively,compared with the sequential allocation scheme and the static mapping scheme.As for LAA system,we develop an adaptive spectrum access and power allocation in the framework of Lyapunov optimization that enables the minimization of the sys-tem average power consumption under a certain queue stability constraint.The theoretical analysis and simulation results show that tuning the control param-eter V can quantitatively achieve a tradeoff of 61(1/V),6(V)?between power consumption and delay.The proposed algorithm can reduce the power con-?sumption over the existing scheme by up to 73.3%under the same traffic delay.(4)In traditional cellular networks,the random access process and data transmission process are relatively independent,which makes it difficult to support the differentiated QoS requirements of massive devices.To deal with this problem,we develop a priority-queueing-based access class barring(ACB)model,and propose a joint random access and data transmission protocol,to guarantee end-to-end QoS requirements of different traffic types.First,tak-ing into account the random access and short packet transmissions,we derive a novel effective capacity of the massive machine type communication(mMTC)network.Based on the derived effective capacity,the barring policy is modeled as a non-cooperative game.A distributed iterative algorithm and a price-update algorithm are proposed to maximize the network utility.The simulation also shows that the total effective capacity of the proposed distributed algorithm is increased to about 2 times compared with the full device activation scheme,and the total effective capacity of the proposed dynamic value algorithm can be improved by 35.5%compared with the proposed distributed algorithm.