Research On Resource Allocation Technology For Wireless Network Virtualization In LTE-A Uplink Networks

Wireless network virtualization technology can abstract,isolate and share the physical infrastructure network equipment,which decouple the physical resource supply process and service provisioning process,physical mobile network operator(PMNO)and virtual mobile network operator(VMNO)manage the two processes above respectively,Through wireless network virtualization,network infrastructure separates itself from the service,multiple types of service can be served in a basic network,the overall cost of mobile cellular deployment and operation can be significantly reduced,the problem that mobile network architecture is becoming rigid and can't meet the need of user is solved.Existing research on wireless network virtualization mostly focused on the network layer above the data link layer,which is mainly due to the virtual network deployment is based on application service flow level,therefore,physical layer characteristics,such as channel state information,are neglected in the research of wireless network virtualization,which leads to channel resources can't be used completely when resource scheduling is performed and the physical resources are wasted.In addition,only the bandwidth service is guaranteed when traffic quality of service requirement is considered in a few literatures on physical layer virtualization technology,quality of service requirement such as traffic flow delay is not studied deeply.This thesis aims to study the resource allocation algorithm for LTE-A network uplink wireless network virtualization,firstly,the usage of the finite state Markov channel model is proposed to predict the channel state information in a scheduling slot in this thesis,and the average capacity is achieved in a scheduling slot,then then fading and varying can be ignored and the capacity mean of the wireless channel can be operated directly to study resource scheduling.Secondly,traffic admission control is performed under the condition of fixed time delay quality of service constraints in this thesis,time delay quality of service constraints is combined with arrival rate and capacity,this thesis proposes that customized service is provided to subscribers by the effective bandwidth and effective capacity model at different layers respectively,and then the quality of service characteristic of the service is realized.Finally,this thesis design a dynamic interaction mechanism considering resource allocation and service access jointly,according to the different functions of PMNO and VMNO,in the physical layer,the resource allocation model is established according to the principle of on-demand allocation,in the MAC layer,the service access model is established by using the non-cooperative game according to the competition among the virtual networks.The optimized results of two models interact with each other,MAC layer presents the effective bandwidth requirement with quality of service characteristic to physical layer,the resource allocation variables of the physical layer resource allocation model can be solved by branch-and-bound algorithm,effective capacity can be got by resource allocation variables and channel state information,the delay quality of service requirement is satisfied according to the relationship between the effective capacity and arrival rate,traffic admission control variable is obtained by the gradient iterative algorithm.PMNO and VMNO transfer the resource allocation variables and traffic requirement repeatedly,when the traffic admission control variable tends to be stable,it means the whole system has reached balance.The simulation results demonstrate the superiority and stability of quality of service guaranteed wireless network virtualization proposed in this thesis in terms of the access rate,the utility function and the delay-outage probability.