The Application of EB-Equivalence Rules to the Structural Reduciton of GSPN Models

Abstract The paper presents a set of structural reduction rules for GSPN that allow in many cases the elimination of immediate transitions, and that can recognize and reduce some symmetries present in the timed behavior of the model. These rules formalize some modeling tricks that experts have been using for a long time, thus paving the way for their automated implementation, hence making them available also to novice GSPN users. The reductions were inspired by a set of rules originally defined in the framework of exhibited behaviour equivalence (EB-equivalence), a notion developed for untimed 1-safe Petri nets whose transitions can be labeled by observable or unobservable actions. This set of reduction (or rewriting) rules are reformulated to deal with the class of Petri nets with arc multiplicities, no place capacities, and with transition priorities by means of a suitable interpretation of the transition labelling function. Furthermore, a calculus for the computation of transition weights is introduced, so that the new reduction rules can be applied to GSPN models.