An iterative technique for the computation of approximate performance
indices of a class of stochastic Petri nets models is presented. The p
roposed technique is derived from the mean value analysis algorithm fo
r product form solution stochastic Petri nets. in this paper, we apply
the approximation technique to stochastic marked graphs; In principle
, the proposed technique can be used for other stochastic Petri net su
bclasses in the paper, some of these possible applications are present
ed. Several examples are presented in order to validate the approximat
e results.