Research On Efficient Resource Allocation Technology Based On Clustering In Ultra-dense Networks

Ultra-Dense Network(UDN)forms dense small cells by densely deploying a large number of small-power micro base stations,which can achieve seamless coverage of wireless cellular networks,improve network throughput,improve user service quality,and reduce The transmission delay of the network.UDN has become a key technology of the fifth generation mobile communication system.However,UDN also has some technical challenges.The number of small cells has increased sharply and is spatially overlapped with macro cells,which makes the network topology very complicated.The resource management of base stations is becoming more and more difficult,and the network energy consumption is becoming higher and higher.How to achieve efficient and low-complexity wireless resource allocation in UDN to improve network throughput and energy efficiency has become a critical issue to be solved.In view of the fact that clustering can simplify the topology of ultra-dense networks and reduce the complexity of resource allocation in UDN,this paper studies the efficient allocation of UDN resources based on clustering.The main research contents and results include:(1)The UDN small cell base station dynamic clustering technology is studied,and a K-means clustering algorithm based on weighted density is proposed to achieve efficient dynamic clustering of small cell base stations.In fact,when the traditional K-means clustering algorithm is used to cluster UDN small and medium-sized cells,the clustering result is likely to fall into the local optimum due to the random selection of the initial cluster center.To solve this problem,this paper designs a new cluster center selection strategy combining the distribution density of base stations and the minimum distance between base stations,and proposes an improved K-means clustering algorithm based on weighted density.The research results show that,compared with the traditional clustering algorithm,the base station clusters obtained by the algorithm in this paper are more uniformly distributed,without overlapping between clusters,and the algorithm has better adaptability to the dynamic changes of the network topology.(2)The resource allocation scheme based on clustering in UDN is studied,and a two-stage resource compensation allocation algorithm for distinguishing user priority is proposed to improve network throughput.Considering the allocation of OFDM(Orthogonal Frequency Division Multiplexing,OFDM)time-frequency resource blocks in UDN small cell base stations,the optimization problem of time-frequency resource allocation is first established with the goal of maximizing network throughput;Then introduce base station clustering to reduce the complexity of solving the optimization problem,and transform the optimized global resource allocation of UDN into the optimized resource allocation within the cluster;Finally,considering the user priority,a two-stage resource compensation allocation algorithm is proposed to optimize the allocation of resources within the cluster.The research results show that the proposed resource compensation allocation algorithm improves UDN throughput while taking into account user fairness and service quality.(3)The problem of energy efficiency improvement based on clustering in UDN is studied,and an energy efficient non-cooperative game power allocation algorithm is proposed.Aiming at the problem of high system energy consumption caused by a large number of densely deployed small cell base stations in UDN,an energy efficiency model of UDN is established based on base station clustering,and the energy efficiency optimization problem of joint time-frequency resources and power allocation is elaborated and solved Coupled to two sub-problems of time-frequency resource block allocation and power allocation.Aiming at the power allocation sub-problem,a distributed non-cooperative game iterative power allocation algorithm is proposed,based on which an optimal power allocation scheme that maximizes the energy efficiency of UDN is obtained.Simulation results show that the proposed power distribution algorithm can effectively improve the energy efficiency of UDN.