ALLOS: uma ferramenta para solucionar modelos de redes de filas usando cadeias de Markov.

Abstract

The relevance of Markovian models in systems performance evaluation and the The need for efficient model specification and solution tools attention of the new generation of competing systems. These tools should be sophisticated to the point of making its users transparent, the mathematical definition of modeled system and the details inherent in model solution techniques. And in this sense that this dissertation presents ALLOS, a friendly tool for troubleshoot queue network models using Markov chains. The process of generating the state space for a specific model and "made automatically from a state and rules that describe system behavior. ALLOS is part of an intelligent, integrated environment for modeling and solving queuing network systems called SAVAD, a multidisciplinary project involving Computer Networking and Artificial Intelligence Groups of the Federal University of Parafba - CCT / UFPB. The use of ALLOS is intended primarily for systems that contain resources, namely computer systems and computer networks, and others that can be modeled using the queuing network paradigm.