Análise de desempenho de SED via autômatos modulares estocásticos: abordagem via simulação computacional e método analítico

• Ano de defesa: 2019
• Autor(a) principal: Lisley Mara Guimarães de Souza
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Sistemas a eventos discretos; Modelagem de sistemas; Autômatos; Avaliação de desempenho; Robôs; Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/RAOA-BCEH7G

The modeling of systems, in the engineering area, sometimes presents failures because they are large systems with many functionalities. Due to a large number of possible states, the development of a single model that represents the behavior of the system as a whole becomes complex. Therefore, techniques with a modular characteristic become an attractive alternative for modeling systems. In the context of Discrete Event Systems (DES), it is possible to model systems using automata, Petri Nets, max-plus algebra, among others; and to work with a queueing system, client-server, manufacturing, traffic, and many others. This work proposes a methodology that deals with the modeling, implementation, and analysis of the results in order to model and analyze systems in a modular way, by means of stochastic automata. In addition, properties of the Markov chain and queueing theory are used, which makes possible the analysis of systems with large computational complexity. Thus, the same modeling methodology is used for analysis through analytical and simulation methods. One of the main results of this work is the demonstration of the efficiency of stochastic modular automata due to the ease and reduction of errors in the modeling step. The methodology was applied to an automated guided vehicle system as well as a client-server system with impatience.