Well-Defined Stochastic Petri Nets

Formalisms based on stochastic Petri Nets (SPNs) can employ structural analysis to ensure that the underlying stochastic process is fully determined. The focus is on the detection of conflicts and confusions at the net level, but this might require to overspecify a given SPN model. The problem becomes even more critical when reward processes of interest derived from the basic underlying process are considered. Typical examples are state-dependent impulse reward measures. We propose a definition of well-defined SPNs, which takes into account whether the basic underlying stochastic process or the derived reward processes are determined. A state-space-based algorithm to determine whether a given SPN is well-defined is provided.