The Early Days of GSPNs

The paper reports on the modelling experience that lead to the proposal of Generalized Stochastic Petri Nets (GSPN). A research program, named “Progetto Finalizzato Informatica” (PFI for short), comprised a number of research lines, including one devoted to multiprocessor systems, aiming at the achievement of both experimental and theoretical results. Markov chains were the models found to be more effective for the objectives of the research, but the complexity of the phenomena that had to be analyzed quickly made them inpractical. Stochastic Petri Nets where then used first as a graphical method for constructing large Markov chains, but extension where soon proposed to exploit the theoretical results already available in the literature and to easily deal with delays with quite different time scales thus yielding GSPNs.