A Cluster-Based Wireless Networks And Resource Allocation Method

Due to the rapid development of communication and computer techniques,Mobile Internet has been widely used in real life.As the scale of the network becomes larger and larger,more and more entities and inter-entity relations need to be dealt with in the different networks.These changes make resource allocation problems with various types of resources in existing networks very complicated.In this thesis,based on the graph representation of the resource allocation problem in the communication network,the performance of the current resource allocation algorithm is improved by using the clustering technique which is widely used in the computer field.This thesis has carried out the following research contents:We use clustering algorithms to improve the performance of existing resource allocation algorithms.On the basis of using the graph method to represent the resource optimization problem of downlink transmission in the wireless communication network,we propose a clustering structure based on the Kmeans algorithm and spectral clustering algorithm respectively to improve the existing resource allocation algorithm performance.We use an existing dynamic graph-based scheduling algorithm to calculate the total number of clusters in the network,and use the two proposed clustering strategies to deal with the communication network optimization problem respectively.By comparing the spectrum efficiency of the whole network,it can be seen that the two clustering algorithms proposed by us have better performance than the dynamic graph-based scheduling algorithm.Meanwhile,the performance of the Kmeans clustering strategy is slightly better than that of the spectral clustering algorithm.But the spectral clustering algorithm has better adaptability than the Kmeans algorithm in different graph representation methods.In addition,we also discuss the influence of a series of key parameters in simulation on the network spectral efficiency.We use the density peak algorithm to improve the selection of the initial cluster center and the total number of clusters in the clustering problem.We propose a new resource allocation algorithm based on the density peak strategy and discuss the parameter setting and selection in the algorithm.The performance of the Kmeans algorithm and spectral clustering algorithm is further improved by using the selection of the initial cluster center and the total number of clusters based on the density peak algorithm.The performance of the three algorithms proposed in this thesis is better than the dynamic graph-based scheduling algorithm.In the case of the same cluster number,the performance of the three proposed algorithms is very close.