Um algoritmo Eficiente de Soluc ¸˜ ao de Modelos com Aplicac ¸˜ ao a Redes Multim´ idia

This work presents a new approximate solution technique to solve models with a special structure that is common in many computer network models. In particular, the class of models we address is that in which the resulting state transition matrix has entries with values that differ in orders of magnitude. This method is based on the GTH algorithm and maintains its main advantages. Furthermore, it is capable of solving complex system models with state space cardinalities and transition matrix structures that would be too costly to solve using GTH. Error bounds are also obtained for the new algorithm. The applicability of the technique is illustrated by solving multimedia system models.