In: System Informatics 9, pages 135-194. 2004. http://www.iis.nsk.su/persons/itar/spnsi.pdf.
Abstract: Petri nets (PNs) are a well-known formal model for qualitative analysis of discrete dynamic systems. Stochastic Petri nets (SPNs) are an extension of PNs with an ability to calculate performance of the systems i.e., with a feature of quantitative analysis. The behaviour analysis of a modeled system consists in building a stochastic process based on the corresponding SPN, and in further investigation of this process. At present, there are a lot of approaches to definition of SPNs. For example, discrete and continuous time SPNs were proposed as well as those with different types of transition delays, priorities and inhibitor arcs. In addition, other structural and behavioural extensions of SPNs were introduced for more adequate treatment of particular problems of quantitative simulation. In this paper we describe the most well-known classes of SPNs and the related analysis methods. We compare the presented SPN classes, give illustrative examples, describe application areas. In addition, we discuss how to propose transition labeling for the SPN classes with a purpose to compare SPNs by behavioural equivalences.
Keywords: Standard; inhibitor-priority and stochastic Petri nets; probability distrubutions; Markov processes and chains; transient and steady states; labelling; equivalence.