Research On User Access Algorithms And Resource Allocation Mechanism In Ultra-dense Networks

With the rapid development of global information technology,the types of communication services expand rapidly,and the amount of communication data grows thousands of times.At the same time,along with the access of massive mobile Internet devices,the "ubiquitous communication" which surpasses the traditional communication meaning is booming,making the evolution of wireless communication system extremely urgent.The main directions of the evolution of wireless communication systems include the introduction of new wireless transmission technologies,the introduction of new architectures,and the mining of new spectrum resources.Among them,the ultra-dense network,which promotes spatial reuse through small cell encryption deployment,has received widespread attention in the industry as a new type of architecture.The ultra-dense network refers to the dense deployment of low-power nodes in the coverage area of the existing macro base stations to meet the increasing communication demand by improving spatial reuse.Compared with the traditional cellular network,the ultra-dense network effectively shortens the distance between the user terminal and the access point,and the high frequency spectrum reuse makes the limited resources fully utilized,and effectively improves the system capacity and spectrum efficiency.However,the dense deployment of cells and the coexistence of multiple types of nodes result in the complex structure of the network to be accessed,and traditional resource allocation strategies are no longer applicable.At the same time,high-density small nodes will bring considerable energy consumption and violate the development requirements of green communications.In this context,this paper aims to improve network energy efficiency and optimize the quality of user experience,and studies user access and wireless resource allocation strategies based on the characteristics of different service types.Firstly,the ultra-dense heterogeneous cellular network model is constructed according to the principle of separation of control plane and data plane of 5G network architecture,including macro base station,micro base station and femtocell with plug and play characteristics.The base station deployment is characterized by a random geometric process that is closest to the real communication scene.Among them,multiple base stations in the same area share some resources for transmission carried by the control plane,and each base station independently transmits user plane data on its remaining resources to realize the separation of the control plane and the data plane in the 5G ultra-dense networking scenario.Secondly,in view of the coexistence of multiple types of nodes and the complex structure of the network to be accessed,a user access algorithm based on Analytic Hierarchy Process(AHP)is proposed.This algorithm models the user access selection problem under the ultra-dense heterogeneous cellular network model as an analytic hierarchy problem.,fully considering the power consumption,signal strength,coverage,user mobility and other factors related to decision-making.By means of the weight priority coefficient of different levels of elements,it breaks through the limitation of single node type of traditional access strategy and completes the optimal access network selection.The simulation results show that based on the AHP user access algorithm,the traffic offload to the macro base station is effectively realized,and the communication service quality obtained by the user is significantly improved.Finally,to solve the problem that traditional wireless resource allocation strategies are no longer suitable for ultra-dense heterogeneous networks,a hierarchical resource allocation algorithm based on QoE is proposed,which includes two parts: spectrum resource allocation algorithm based on QoE and power allocation algorithm based on deep reinforcement learning.The spectrum resource allocation algorithm based on QoE introduces the user experience quality into the cellular network quality of service evaluation system,and constructs the mapping relationship between rate and QoE.According to the user service type,the ideal transmission rate of each user is determined and the user index is established accordingly.The spectrum resources are allocated according to the index order.On this basis,the power allocation problem was modeled as a constrained multi-objective optimization problem,and the power resource allocation was completed by making full use of the perceptual ability of deep learning and the decision-making ability of reinforcement learning.Firstly,the forward transmission network is constructed under the drive of maximizing the transmission rate,and then the energy efficiency is used as the reward and punishment value,and the error function is constructed to complete the algorithm reverse training.The simulation results show that the hierarchical resource allocation algorithm based on QoE effectively improves the system energy efficiency and user experience quality.