Um estudo da eficiência da autocentralidade no problema de isomorfismo de grafos

This work treats the application of the eigenvector centrality in solving the Graph Isomorphism Problem. This property, taken from spectral graph theory, was used by Philippe Santos in [SANTOS 2010] to propose a spectral algorithm for solving this problem. An adaptation of the power method is propos...

ver descrição completa

Detalhes bibliográficos
Autor: Baroni, Marcos Daniel Valadão
Tipo de documento: dissertação
Estado:Versão publicada
Data de publicação:2012
País:Brasil
Recursos:Universidade Federal do Espírito Santo (UFES)
Repositório:Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
Idioma:português
OAI Identifier:oai:repositorio.ufes.br:10/4257
Acesso em linha:http://repositorio.ufes.br/handle/10/4257
Access Level:Acceso aberto
Palavra-chave:Graph theory
Isomorphims
Algorithms
Eigenvectors
Teoria dos grafos
Algoritmos
Autovetores
Isomorfismos (Matemática)
Ciência da Computação
004
Descrição
Resumo:This work treats the application of the eigenvector centrality in solving the Graph Isomorphism Problem. This property, taken from spectral graph theory, was used by Philippe Santos in [SANTOS 2010] to propose a spectral algorithm for solving this problem. An adaptation of the power method is proposed to compute the eigenvector centrality, producing a competitive version to the spectral algorithm of [SANTOS 2010]. Based on this adaptation, the efficiency of the eigenvector centrality in solving the problem is studied. In addition, it is proposed an iterative labeling algorithm, called Centrality Based Iterative Algorithm, which can be applied to any type of graph, including regular ones. Several tests are performed to compare the two proposed algorithms with some others well-known algorithms from literature, such as Nauty.