Algoritmos genéticos multiobjetivos aplicados ao roteamento multicast com qualidade de serviço

• Ano de defesa: 2009
• Autor(a) principal: Vita, Stéfano Schwenck Borges Vale
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; BR; Programa de Pós-graduação em Ciência da Computação; Ciências Exatas e da Terra; UFU
• Palavras-chave em Português: Roteamento multicast; Algoritmos genéticos multiobjetivos; NSGA; NSGA II; Dominância-e; Qualidade de serviço; Redes de computadores; Algoritmos genéticos; Multicast routing; Genetic algorithms multi-objective; E- dominance; Quality of service; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufu.br/handle/123456789/12480

Multicast Routing is an effective way to communicate between multiple routers into computer networks. In general, the quality of service (QoS) is required in most of multicast applications. Several researchers have investigated the application of genetic algorithms in multicast Routing with QoS restrictions. The evolutionary environments proposed in this dissertation employ a multi-objective approach embracing the concept of Pareto Optimum to solve the Routing calculus and to deal with several QoS metrics. Basically, four multiobjective environments were built to solve the problem of multicast Routing with QoS. The first was based on NSGA and the second was based on NSGA-II; they adopted the original concept of Pareto dominance. The third multi-objective environment built is an adaptation of NSGA-II which incorporates the e-dominance. The fourth environment is also an adaptation of NSGA-II, but it employs a variation of e-dominance, the e -dominance. Five different pairs of objective functions were evaluated: the first objective in each pair is related to the total cost of a multicast route. The second objective accounted for: (i) the total delay of the multicast tree, (ii) the average of accumulated delay from the source to each destination node, (iii) the maximum accumulated delay from the source to each destination node and (iv) the total number of routers in the multicast tree. Our results indicate an assessment of the four multi-objectives environments. These algorithms were applied find routes in two network topologies named REDE0 and REDE1.