This article discusses General Motors’ North American Enterprise Model, a system dynamics model of the entire North American automobile market. The Enterprise Model takes a broad look across the corporation and its marketplace, combining internal activities such as engineering, manufacturing and marketing with external factors such as competition for consumer purchases in the new and used vehicle marketplaces. Eight groups of manufacturers compete monthly for a decade across eighteen vehicle segments, making segment-by-segment decisions about price, volume and investment. The model enables Monte-Carlo analysis of alternative strategies. The goal is to find and assess the likely impact of improved strategies for managing the business that are robust across uncertainty about consumers, competitors, and the macro-economy.
This article presents a high level overview of the model. We discuss why and how the model was built and what sorts of results came from it. We discuss software tools we wrote to supplement Vensim: a profiling tool for finding inefficient equation formulations, and a syntax coloring tool for automatically color coding Vensim sketch diagrams according to selected criteria. Finally, we discuss the limitations of the System Dynamics paradigm for large models, and how Agent Based Models might complement traditional system dynamics.
A small piece of a RANDOMIZED second choice matrix