基于关联矩阵的网络拓扑辨识方法研究

提出一种基于关联矩阵的网络拓扑辨识方法．通过对网络的链状和树状基本结构关联矩阵的分析，认为关联矩阵可以完全反映网络基本结构的拓扑信息．网络关联矩阵经过矩阵变换形成三角矩阵，可以反映网络拓扑结构；对三角矩阵进行分级和分解运算，可以提取网络的拓扑信息．这种方法适用于实际应用中常见的树状结构拓扑的辨识，甚至适用于带并联支路（环路）的网络拓扑辨识．该方法简洁实用，结合系统设计仍有更进一步扩展的潜力。

Method for Network Topology Identification Based on Incidence Matrix

Based on incidence matrix,a new method was developed for network topology identification.By analyzing the incidence matrix of chain structure model and tree structure model,incidence matrix was obtained,then topology information of basic network structures could be reflected.Topology information of network structure was also obtained through converting an incidence matrix into standard triangular matrix.Hierarchizing and decomposing the triangular incidence matrix,topology information of network is extracted,which is suitable for network topology identification of common tree structure,even suitable for network topology identification of the structure with parallel branch(circle branch).This method is simple for use and also has expandable ability if combining with the system design.