Agent-Based Computational Economics (ACE)

What is it?

ACE is a computational approach to studying dynamic economic systems in which heterogenous software agents autonomously interact within virtual worlds.  The approach allows many standard assumptions of economics, e.g. the representative agent, to be dropped in the building of economic models.

Two excellent resources for starting to learn about ACE modeling are:

Tesfatsion's Site:  The definitive collection of materials related to ACE.  A must visit for anyone interested in what is being down and what is available.  The site includes self-study courses in ACE and in the computational tools.  Visit her site.

Handbook of Computational Economics, Vol. 2: Agent-Based Computational Economics.  Leigh S. Tesfatsion and Kenneth L. Judd. (North-Holland, 2006):  A thorough overview of the state of the literature as of 2006 and a must read and excellent reference for anyone starting to undertake research in the area.

Software Platforms

Many ACE researchers "role their own" models in a programming environment of their choice.  Being able to write code can be a barrier for some economists not familiar with or comfortable programming.  The following environments reduce the overhead of getting into developing your own models and are some of my favorites.  Tesfatsion has a more complete list here.

Netlogo:

A very good tool for getting into ACE and agent-based modeling; particularly if you have little or no programming experience.  Netlogo is my environment of choice for working with computational labs for use in the classroom. Available here.

Mason:

My toolkit of choice for research.  Mason provides a very flexible structure for separating the structure of the model from the visualization of the model.  Knowledge of Java is necessary to program with the Mason libraries. Available here.

Last Modified: 2016-01-06