Função estocástica de tempo de percurso

• Ano de defesa: 2010
• Autor(a) principal: Paula de Campos Oliveira
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Redes de filas; Performace; Simulação; Dependente do estado; Estatística; Análise estocástica; Métodos de simulação; Teoria das filas
• Link de acesso: http://hdl.handle.net/1843/BUBD-8AVM68

We apply finite state-dependent queueing networks to model travel time in mobile communication systems. Although they have been successfully used in the past to model vehicular traffic, state-dependent models have not been applied to mobile communication systems, to the best of our knowledge. The novelty of state-dependent stochastic mobilitymodels is that the congestion phenomenon is explicitly considered, that is, the user speeds fall when the number of users in the system increases. We present a detailed description of the simulation model used to estimate the performance measures of the queueing networksand show computational results for a comprehensive set of instances. As we show, finite state-dependent stochastic models bring interesting new insights, for instance, that in some cases mixed bimodal distributions will better describe the cell residence time of a call than the classical probability distributions used in the past.