Veröffentlichung

Titel:

Synthesizing Petri Nets from Labelled Petri Nets.

AutorInnen:

Robin Bergenthum
Jakub Kovář

Kategorie:

Konferenzbandbeiträge

erschienen in:

Proceedings of PETRI NETS 2025, Lecture Notes in Computer Science 15714, 63-85, Springer 2025.

Download:

electronic edition via DOI (open access)

Synthesis automatically generates a process model from a behavioural specification. If the desired process model is a Petri net, synthesis is addressed through so-called region theory. Region-based synthesis has been extensively studied for cases where the specification is a transition system, a step-transition system, a language, or even a partially ordered language. Although the ideas of region-based synthesis are consistent across different types of specifications, each specification type has its own definition of regions and uses different representations of the set of all regions to synthesize a result. Up to this point, state-based and language-based regions are simply two distinct concepts. In this paper, we introduce Petri net regions, to synthesize a Petri net from a set of labelled Petri nets. We synthesize a result that can simulate the specified behaviour using a set of minimal regions. Thus, we advance region theory to the next level. Furthermore, we show that both language-based and state-based regions are Petri net regions as well. There is no longer any need to distinguish these concepts any more. Using Petri net regions, we present an implementation of a synthesis algorithm that handles state-based, language-based, and Petri net-based input.