基于压缩感知的异构网络数据动态重构算法

Dynamic Reconstruction Algorithm of Heterogeneous Network Data Based on Compressed Awareness

针对异构网络数据资源存在节点及链路均衡度过低且数据重构资源分配路径误差较大的问题,设计一种基于节点拓扑感知的异构网络数据动态重构算法。利用模糊核聚类算法将数据样本集映射至高维空间内,聚类目标数据,根据节点在网络中对应的坐标构建异构网络拓扑模型及数据权矩阵模型;设置路由器物理链路通信图,根据链路带宽约束优化数据重构资源的分配路径,将节点核心区数据重构,并分配至聚类节点来感知数据的动态变化,为每一个核心聚类节点拓扑感知一个路由器数据的动态变化,来实现全局的数据动态重构,至此完成基于节点拓扑感知的异构网络数据动态重构算法的设计。设计仿真实验,测试算法的节点均衡度与链路均衡度。实验结果表明,设计的算法的节点均衡度为0.93,链路均衡度为0.90,均高于对比方法,因此可以得出,该算法的资源均衡度更好。

In order to solve the problem of low node and link equalization in heterogeneous network data resources and large error in the distribution path of data reconstruction resources,a dynamic data reconstruction algorithm based on node topology perception is designed.The fuzzy kernel clustering algorithm is used to map the data sample set to the high-dimensional space,cluster the target data and construct the heterogeneous network topology model and data weight matrix model according to the corresponding coordinates of nodes in the network.The router physical link communication diagram is set.According to the link bandwidth constrained optimization path data,the resources distribution path and the joint core data are reconstructed and assigned to the cluster nodes to perceive the dynamic change of the data,and the dynamic change of the router data is perceived for each core cluster node topology,so as to realize the global dynamic data reconstruction.Thus,the design of the heterogeneous network data dynamic reconstruction algorithm based on the node topology awareness is completed.The simulation experiment is designed to test the node equilibrium degree and link equilibrium degree of the algorithm.The experiment shows that the node equilibrium degree of the designed algorithm is 0.93 and the link equilibrium degree is 0.90,both of which are higher than the comparison method.Therefore,it can be concluded that the resource equilibrium degree of the designed algorithm is better.