Sign In

Communications of the ACM

ACM TechNews

Agent-Based Computer Models Could Anticipate Future Economic Crisis

View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook

The U.S. Department of Energy's Argonne National Laboratory is working to create new economic models that will provide more realistic pictures of different types of markets to help prevent future economic crises. Argonne systems scientist Charles Macal says traditional economic models do not represent the market's true internal dynamics because they ignore the decision-making processes of individual consumers and investors. "The traditional models don't represent individuals in the economy, or else they're all represented the same way--as completely rational agents," Macal says. "Because they ignore many other aspects of behavior that influence how people make decisions in real life, these models can't always accurately predict the dynamics of the market." Macal and other Argonne researchers have created a new group of simulations called agent-based models to better predict how markets will act. The new models partially rely on information collected in surveys that ask individuals about factors that influence their decisions. By having a more thorough understanding of the behavior of individuals, researchers will be better able to predict and prevent economic failures. Agent-based models calculate possible decisions for each individual investor in a model, using the results of these decisions to see what impact they could have on other agents. Creating such detailed simulations relies on the availability of high-performance computers capable of managing the challenge of mathematically representing an enormous number of individual actors. Macal says five years ago it was impossible to model more than a couple of agents, but now millions of agents can be modeled.

From Argonne National Laboratory
View Full Article


No entries found