By Jillian Cordes*
An Agent Based Model (ABM) is a model that simulates interactions between individuals and groups in social and environmental settings. These theoretical models are developed as a way to analyze behavior based on methodological situations. Using computation-intensive procedures, the concept of ABM was developed as early as the 1940s, but did not gain momentum until the 1990s. These models can be used to solve and better understand a variety of business, technological and social problems. For a no-nonsense explanation of how ABM works, see What is Agent Based Modeling?
Typically researchers use real data to “model” the behavior of their “agents” in order to see what kind of changes can be expected to happen when small variables are adjusted. Today, Development Roast brings you five different games that use ABM to simulate and teach real-life change:
1. The appropriately named Game of Life was developed by John Conway, a mathematician who is currently a professor at Princeton University. This game aims to help scientists understand the behavior of cells, animals, and even diseases. There are no winners or losers in this game. Just by playing around on the grid “live cells” are created which are born and die after interacting with neighbor squares on the grid. Every dark square represents a living cell. After the player creates an initial pattern, the cells move about until they reach a stable state. The game embodies the idea of “self-organizing systems” and, based on the rules provided, one can understand how complex patterns and behaviors emerge. The main website also provides a few explanations of common patterns that occur after the simulation has steadied. The stabilization of the screen symbolizes a stop in population growth. The game is written in java and requires a java plug-in to run.
2. An example of an economic simulation is an ABM known as JAMEL (Java Agent-based MacroEconomic Laboratory). Developed by Pascal Seppecher, an economics professor at the Université de Nice-Sophia Antipolis, JAMEL is a cross-platform macroeconomic simulator in which the player can define the conditions of the economy and watch the effects of certain shocks such as a productivity shock or the simulation of a credit bubble. The model covers forty three different macroeconomic indicators such as unemployment, inflation, and wages. Very thorough in nature, JAMEL shows long-term line graphs for each of these indicators in a fictitious economy. Though it is only a prototype, JAMEL can be useful for anyone studying economics or simply interested in policy-making. Professor Seppecher believes that with some upgrades JAMEL could develop into a way to test different economic configurations or policies. He believes it would be a good teaching utensil and students would find it more active (and possibly more entertaining) than traditional learning methods in economics. The model used is closely associated with Keynesian thinking—John Maynard Keynes was a British economist who in the 1930s theorised that state intervention is necessary to moderate the natural ups and downs of the free market—and uses an agent-based approach representing a dynamic out-of-equilibrium economy composed of two primary sets of agents (firms and households) associated with two main functions (production and consumption). A more in depth description of JAMEL can be found in Professor Seppecher’s paper A Java Agent-based MacroEconomic Laboratory and for an account of using JAMEL in an actual experiment see Flexibility of Wages and Macroeconomic Instability in an Agent-Based Computational Model with Endogenous Money.
3. Another economics-centered game run with java is the Small-World Trade Network Demo developed by economics Professor Mark McBride of Miami University in Ohio. The simulation examines the market efficiency of trades made between buyers and sellers of goods numbering from 200 to 10,000 in quantity. This simulation allows for more player participation than JAMEL and has fewer variables. This perhaps makes it less realistic but easier to understand for someone with a growing knowledge of economics. Professor McBride uses it in his undergraduate class on computational economics and has found it to be a fairly successful learning instrument. In the game, the player is able to change factors such as the number of buyers and sellers, and the maximum buyer value and seller cost. Based on these factors, the player is able to see a life-like supply and demand curve, price dispersion, market efficiency, and a few other outcomes. The overall message of the simulation is that buyers and sellers do not need to make rational decisions in order to produce market efficiency and that efficiency is derived mainly from a market institution.
4. Another ABM simulation, known as Sugarscape, creates an artificial society in order to study human social phenomena. Citizens are created that “move around” on the grid, gathering food, giving birth, and acting the way people are thought to have done in a pre-agricultural society. Through this researchers are able to better predict the effects of population growth on soil fertility and economic outcomes. This model, which also uses java to run, requires minimal player interaction, but is still informative and was developed by Professor Joshua Epstein of John Hopkins University and Professor Rob Axtell of George Mason University. The project was developed as an international effort to identify conditions for a sustainable global system and to design policies to assist in creating such a system.
5. Lastly, SimPachamama is a policy game released this month by the Institute for Advanced Development Studies (INESAD) and Conservation International in Bolivia, and the London School of Economics (LSE), Ecosystem Services for Poverty Alleviation (ESPA), and two other U.K.-based partners. In the game, the player takes on the role of the mayor of a small Bolivian community, who has to put into place different policies in order to achieve maximum wellbeing of citizens, while reducing deforestation of the surrounding Amazon. Similar to the format of the Small Trade Network Demo, SimPachamama players need to account for a limited budget, as well as the long-term effects of policies and potential unintended side effects.
SimPachamama differs from other ABM simulations in that it is being used along with two other educational simulation tools in order to influence policy and make a real impact. The simulation, which is based on extensive data from a real life Bolivian forest community, shows that little reduction in deforestation would occur without two important policies: tax on deforestation in Bolivia and international payments for reduction of deforestation from rich countries. The project team has designed an infographic that explains the importance of these two policies in raising US$1 billion every two years, which is predicted to tip the balance in favor of keeping more of Bolivian forests standing. To learn more visit www.inesad.edu.bo/simpachamama.[contact-form to=’firstname.lastname@example.org‘ subject=’SIM 5 ABMs’][contact-field label=’Like this article? Enter your email for weekly updates from Development Roast’ type=’email’ required=’1’/][contact-field label=’Your name’ type=’name’ required=’1’/][/contact-form]