Epstein, J.M. & Axtell, R. (1996)
Growing Artificial Societies: Social Science From the Bottom Up
MIT Press / Brookings Institution Press (review)

Its funny how some important books just slip under one's radar. This one is clearly not new and is highly relevant to the kind of work I do --- but I'd not read it until just now! Anyway, its a great little book describing a number of experiments with the authors' virtual ecosystem Sugarscape. In this virtual world agents consume sugar which appears in their gridded landscape. This resource may be gleaned, traded, fought for etc. by virtual evolutionary agents inhabiting a planar world. In addition the authors discuss a model of disease transmission (including a basic model of immunity and epidemiology) on Sugarscape not disimilar to my own independently developed (but considerably less old) model.

The reasons for the authors' experiments are several fold, but in particular, as the book's title indicates, the authors propose that simulation may be an effective means for doing Social Science. They demonstrate various properties of the environment and the behaviour of their virtual agents that give rise to phenomena that can also be observed in human interactions. Observed phenomena include splits in wealth distribution, the emergence of social networks, changes in population numbers etc.

I don't really expect that a seasoned a-lifer would find anything surprising in these results. However the book is valuable for a couple of reasons : (i) it is an early record of the value of Artificial Life experiments within simulated environments to a field other than Artificial Life itself; (ii) it is a clearly written text on a simply but usefully designed agent-based simulation.

Thumbs up!

Alan Dorin, 10 May 06