Real-world industrial supply networks are highly complex structures made up of a multitude of competing individual companies. Today’s structures span the whole planet and link processes over a timeline that measures in months. In this article we focus on one of the most complex networks, the supply in the automotive industry.
The observed dynamics emerge from the physical and virtual interactions of the individual components of the supply network. The complexity of the net makes an analytic description of the system-level behavior infeasible. Instead, we have to resort to models of the individual dynamics that are then explored in simulation experiments. In this article we compare two modeling approaches—equation-based and agent-based modeling—and we report on two research projects at ERIM’s Center for Electronic Commerce that applied agent-based modeling in the analysis of simple supply structures.
Simulation of system dynamics is a central element in supply network management research. Agent-based models of real-world supply chains can be built by domain experts that do not have to be versed in information technology (IT). Using these models, a quantitative evaluation of the impact of parameters and strategies in the supply network design can show the financial advantage of the introduction of supply network management. The bulk of this article reports on a simulation exercise at the DaimlerChrysler Corporation that identified a potential win-win situation for all partners along the supply chain if a new forecast policy is adopted.