A trace semantics for petri nets

Abstract

A generalization of the notion of trace is proposed. This enables us to associate with each Petri net a single behavioural object, namely a poset of (generalized) traces. A characterization is given of the trace languages defined by Petri nets. We show that the general event structures of Winskel and the stable event structures can also be characterized in terms of our trace languages. One consequence is that in this framework, stable event structures, general event structures, and Petri nets constitute a strictly ascending chain in terms of expressive power.