Ana Paula Couto da Silva,  Rosa Maria Meri Leão,  Edmundo de Souza e Silva. 

Multimedia networks have been extensively studied in the literature. Markovian reward models have been used to model and to obtain performance measures from multimedia networks. In these models, reward rates are associated with states and/or reward impulses are attributed to transitions between states. We have developed an algorithm to obtain the queue size distribution of a finite buffer when it is fed by a fluid. We base our analysis on impulse reward models. The probability transition matrix generated from the model we obtained has a special structure that is the key to our technique. The algorithm presents computational advantages when compared with others approaches in the literature.

http://www.lbd.dcc.ufmg.br:8080/colecoes/sbrc/2002/021.pdf

Processando consulta.