Generalized Stochastic Petri Nets vs. Generalized Semi-Markov Process models

 

 

We want to compare two formalisms based on their modeling power.  The term modeling power needs a separate discussion since it may be not clear without a context what is meant by that. For the untimed formalisms for example, it may mean different thing than for the timed ones.

 

For now we will just informally say that two formalisms have equivalent modeling power if they can model the same class of (discrete-event) systems.

 

We’d like to claim that Gen. Stochastic Petri Nets and GSMP as defined here have equivalent modeling power.

 

P. Haas in [1] shows that for any GSPN there exists a GSMP that strongly mimics the behavior of the GSPN and vice versa., i.e. both formalisms have the same modeling power (in the sense of mimicry). Strong mimicry is defined in [1] as identical stochastic behavior. Even though there are slight differences between our definitions of GSPN and GSMP models and the corresponding definitions in [1] it seems that this result should hold still.

 

 

References:

1.      P. J. Haas, Stochastic Petri Nets, Springer, 2002.