Artificial worlds and economics,part II

This paper continues the discussion of artificial worlds (AWs) begun in Lane (1993b). Here, the focus is on two kinds of AWs. The first, classifier systems, can be used to represent agents that are capable of generating complex behaviors in response to intermittent rewards from an environment of which they are a part. A collection of such agents, engaging in economic interactions with one another, produces another kind of AW, in which such interesting aggregate behaviors as the formation of bubbles and crashes and technical trading in an artificial stock market, may arise. The second kind of AW considered in this paper is artificial economies. These AWs can provide a dynamic, nonequilibrium, microfounded account of such aggregate-level or macroeconomic phenomena as stable growth paths, business cycles, and Pareto firm-size distributions.