EOS: Developing an General Agent-Based Economic Simulation

Economics has always been an important field of study for the simple reason that, at its idealistic heart, it is the study of how to influence the trade and utilization of limited resources in order to make people as happy as possible. However, economics is also a field in which it is very difficult to prove assertions to be absolutely true because economics concerns itself with the behavior of people, and people all too often don’t know how they themselves will behave in certain situations, never mind predicting how other people will behave. This difficulty motivates a certain skepticism about how successful current economic thought actually is in terms of making people happier. Of course a contemporary example of one of economics’ failures is the current recession, the cause of which is a contentious issue agreed upon by few but has been blamed on everyone from the speculators on Wall Street to mortgage holders to government officials who perhaps failed to regulate the interactions of the preceding parties. The complex human element that predicated this recession is often overlooked by economic theory, which instead tends to focus on aggregate behavior that fits nicely into elegant equations. However, the elegance of these equations is lost in the chaos that ensues when they fail to accurately predict and prevent economic downturns. In this paper we assert that economics lends itself more to analysis by simulation rather than analysis by equations, which tend to abstract away too many relevant details to be valid. More specifically, we assert that a simulation that models the interactions of individual agents in an economy is more useful than equations that aggregate individual behavior away into averages. Furthermore, we discuss the design and implementation of a new framework for such a simulation, EOS, which we argue is an improved version of