Análise de medidas de centralidade utilizando distância de Hellinger

Abstract

The theory of complex networks has among its characteristics the interdisciplinarity, besides being concise and clear can simultaneously involve several areas with broad applicability in real modeling. Some networks, for example, the Internet, airports routes, scientific collaborators, electricity and sewage have a physical structure that characterize them as complex networks. Because they are dynamics these networksmight be affected by topology changes. In this case, it is necessary to know physical and analytical characteristics to better understand such events. In particular we analyzed a set of Autonomous Systems (ASs) composed of customer networks, university networks, beyond points exchanges, and others. It is the vertices and edges each AS the connections between the ASs. This study has the objective to verify a real Internet network of ASs, through their respective graph to understand their behavior and their characteristics by way of centralizedmeasures based on vertices. To achieve this goal we use statistical studies associated of the Monte Carlo Method and modified randomly your graphwith perturbations for newgraphsmodified topologically. Through the relative frequencies of the centrality measures based on these graphs vertices, original and modified. We obtained through Hellinger distance quantifier many relevant results. About doing perturbations, removing edgeswas themost kind of perturbations that had altered the topology compared the others networking. According to the centralitymeasures, the degree of Closenessmeasure was themost vulnerable in general compared another perturbations.