A tool for tailored code generation from Petri net models

The use of Petri nets for the modelling of discrete-event systems is well-studied. Yet, the tools allowing the implementation of these models, and supporting code generation, are still very few, almost non-existent. This paper starts by presenting a Petri net class, based on place/transition nets and well-known concepts from synchronised and interpreted Petri nets. This Petri net class allows the association of external input signals to transitions and the association of external output signals to transitions and place markings. Additionally, the class provides support for the specification of input and output events. Next, the paper presents a code generator able to output optimised executable code from these nets. The generated code can be optimised by several distinct strategies, which facilitate the creation of code tailored to specific platforms, as well as for specific classes of Petri nets.

1. Introduction

A common way to model embedded systems uses a decomposition based on a controller that interacts with its environment. By using a synchronising signal (a global external clock), this interaction can be forced to happen at periodic points in time. These discrete event systems can be conveniently modelled by Petri nets and this fact has been known for quite a long time (see, for example, the article by Holloway et al. [6]). Notwithstanding this known adequacy, the number of code generators from Petri net models is extremely scarce as attested by the Petri Nets Tool Database [7].