online read us now
Paper details
Number 4 - December 2014
Volume 24 - 2014
An analytical method for well-formed workflow/Petri net verification of classical soundness
Julio Clempner
Abstract
In this paper we consider workflow nets as dynamical systems governed by ordinary difference equations described by a
particular class of Petri nets. Workflow nets are a formal model of business processes. Well-formed business processes
correspond to sound workflow nets. Even if it seems necessary to require the soundness of workflow nets, there exist
business processes with conditional behavior that will not necessarily satisfy the soundness property. In this sense, we
propose an analytical method for showing that a workflow net satisfies the classical soundness property using a Petri net.
To present our statement, we use Lyapunov stability theory to tackle the classical soundness verification problem for a class
of dynamical systems described by Petri nets. This class of Petri nets allows a dynamical model representation that can
be expressed in terms of difference equations. As a result, by applying Lyapunov theory, the classical soundness property
for workflow nets is solved proving that the Petri net representation is stable. We show that a finite and non-blocking
workflow net satisfies the sound property if and only if its corresponding PN is stable, i.e., given the incidence matrix A of the corresponding PN, there exists a Φ strictly positive m vector such that AΦ ≤ 0. The key contribution of the paper is the analytical method itself that satisfies part of the definition of the classical soundness requirements. The method is designed for practical applications, guarantees that anomalies can be detected without domain knowledge, and can be easily implemented into existing commercial systems that do not support the verification of workflows. The validity of the proposed method is successfully demonstrated by application examples.
Keywords
Petri nets, decidability, workflow nets, Lyapunov stability, soundness, verification