A Comparison of Petri Net and System Dynamics Approaches for Modelling Dynamic Feedback Systems

Abstract

Petri nets are a valuable tool that can be used to simulate workflow systems. They are based on a state-transition approach, and in common with discrete event simulation, events can be scheduled to fire at different time intervals. This is in contrast to the stock and flow approach of System Dynamics, where workflows are aggregated and state transitions modelled continuously through sets of integral equations. This paper continues a theme explored at the Boston 2005 conference, where a paper was presented that identified a common simulation problem, and presented solutions using both discrete event simulation and system dynamics. The benefit of this approach is that it can provide modellers from different methodological worldviews insights into how common problems may be addressed. The problem approached here is a one-actor model of the Beer Game. Two models are developed, detailed experimentation is performed, and overall results and conclusions presented.