Algoritmo para conversão automática de modelos SAN GTA para modelos SAN CTA

Abstract

This work presents a formalism for modeling systems called Stochastic Automata Networks (SAN), SAN formalism aims to increase the abstraction s level and provides a sophisticated alternative model to the tadicional formalism of Markov Chains (MC). SAN uses both Classical (CTA) and Generalized Tensor Algebra (GTA) to simplify the matrix of transitions between states of the model. Despite all models described with GTA having at least one equivalent model described using CTA, and that the solution of certain models based on CTA could be faster than the equivalent GTA based model, this dissertation proposes an algorithm for translating a model described in GTA into the equivalent model described in CTA. It is expected that some models described using functions (using GTA) could be solved more quickly or taking less memory through the solution of its CTA-converted model