Research On Integrated Resource Allocation Method Of Access And Backhaul Based On User Characteristics In Heterogeneous Wireless Networks

Heterogeneous Networks(Het Nets),which is a combination of macro stations,micro stations and pico stations,is one of the key technologies to meet the growing data service needs of users and improve network capacity.At the same time,the introduction of wireless backhaul technology between the macro station and the micro station and the pico station can greatly improve the flexibility of network deployment.However,due to the diversity of user distribution and user service requirements,and the conflict of access and backhaul resource utilization,it brings great challenges to network integrated resource allocation technology.In this paper,multi-cell association and cell merging technology are applied to heterogeneous wireless access and backhaul integrated networks,and combined with network traffic prediction.Aiming at the scenarios of using large-scale antenna arrays in macro-stations and intensively deploying in-band wireless backhaul micro-stations in heterogeneous networks,this paper studies the bandwidth resource allocation method under downlink multi-cell Association and cell consolidation in heterogeneous wireless access and backhaul integrated networks.When the network changes dynamically,a machine learning fusion model based on user characteristics is established to predict the resource allocation pre-planning method of network fluctuation.This paper proposes a joint optimization scheme for multi-cell association and cell merging.It can effectively improve the average user rate by providing services to users through multiple cells(including macro stations and micro stations)under uniform bandwidth allocation.When the macro station is overloaded,the user is associated with the micro-station to reduce the load.Fully coordinate micro-station resources to increase network capacity and user average rate.However,due to the intensive deployment of base station antennas in the actual network,serious inter-channel interference occurs between different cells,resulting in a low rate of cell edge users.In this paper,the user structurebased cell merging algorithm is used to adjust the network structure to eliminate adjacent channel interference,improve network fairness and cell edge user service quality.Therefore,this paper develops an iterative algorithm based on multi-cell association and cell merging techniques to maximize the logarithmic total capacity as the objective function in heterogeneous networks.The mathematical model of logarithmic total capacity in heterogeneous networks is expressed as a non-convex Mixed Integer Nonlinear Programming(MINLP)problem,and transformed into two convexes by the Lagrangian dual method of relaxation optimization layered decomposition.Sub problem solving.The simulation results show that the proposed joint optimization algorithm obtains the approximate optimal solution of the total logarithm of the user in the heterogeneous network,and the average user rate is increased by at least 12%,and the network capacity is significantly improved.In addition,because of the different characteristics of user distribution,density and service type,the throughput and the number of online users in heterogeneous networks change dynamically,and the heterogeneous networks can not adjust the network,change the cell association of users and adjust the allocation of bandwidth resources in real time due to cost constraints.Therefore,this paper designs a two-tier fusion model to predict the network throughput and the number of online users in the next period.According to the predicted value,the network bandwidth resources can be allocated reasonably and the utilization rate of network bandwidth can be improved to solve the network traffic congestion and improve the quality of service of users.This paper uses real operator data on Kaggle platform to clean up data and analyze feature engineering.By comparing the simulation results of two-level fusion model and traditional machine learning model,it proves that the fusion model effectively improves the accuracy of prediction value.Based on the prediction results of traffic and number of online users,combined with the bandwidth allocation algorithm of multi-cell Association and cell merging proposed in Chapter 3,resource allocation and network planning are carried out in advance.The experimental results show that in heterogeneous wireless access and backhaul integrated networks,the bandwidth resource allocation algorithm based on multi-cell Association and cell merging predicted by user characteristics significantly improves the overall network throughput and user quality of service.