Tagged Generalized Stochastic Petri Nets

This paper introduces an extension of the Generalized Stochastic Petri Net (GSPN) formalism in order to enable the computation of first passage time distributions of tokens. A "tagged token" technique is used which relies on net's structural properties to guide the correct specification of this extension. The extended model is suited for an automatic translation into an ordinary GSPN that can be used for the first passage time analysis. Scheduling policies of tokens in places, that are neglected in ordinary GSPNs, become relevant in the Tagged Generalized Stochastic Petri Net (TGSPN) formalism and specific submodels are proposed which are then used during the translation from TGSPNs to ordinary GSPNs. A running example inspired by a Flexible Manufacturing application is used throughout the paper to introduce the different concepts and to provide evidence of the relevance of the results.