A Concept for the Modular Description of Stochastic Petri Nets

Petri nets represent a modeling paradigm which allows to build general models by using a small number of graphical primitives. The strong point of Petri nets lies in their ability to model system aspects such as concurrency and synchronization and to represent these aspects graphically. Due to this reason, stochastic Petri nets (SPNs) are frequently used for model-based performance and dependability evaluations. However, the main drawback is caused by the problem of complexity, both of model specification and of model evaluation. A number of approaches have been suggested in order to manage the complexity problem, e.g., g e n e r a l i z e d stochastic Petri nets [1], stochastic reward nets [4], stochastic activity networks [10], coloured Petri nets [7], and stochastic well-formed nets [3]. In PROTOB [2], Pascal-like language constructs have been added to Petri nets in order to describe models in a modular fashion. In the recent years there has been considerable interest in applying objectoriented technology to Petri nets [8, 9]. Another direction of research is to use operators from process algebras for the composition of SPNs, e.g., [5].