I came across an excellent ABM exploring the emergence of cities, and thought I’d pass along the link. The model is by André Ourednik and Pierre Dessemontet. Their abstract:

The very existence of urban formations on all inhabited continents and throughout the history of mankind since the 3rd millennium B.C. leads to suppose a tendency of some structured societies to maximize interaction by minimizing physical distance. Were this tendency unconstrained, it should eventually lead to the concentration of all of the society’s population into one single point: a situation only partially realized by the distribution of urban populations at the global scale. Models of constraints preventing its realization have thus to be proposed. We have set up one such model, using agent based simulation of food production and accessibility, in order to account for the structural constraints particular to the physical space. The simulations have notably shown that, while necessarily emerging from a society investing agricultural surplus into the upholding of specialists, an upper limit to city-growth is imposed by the phenomena of spatial friction.

So I began to experiment with their model. One of the pre-sets that they thoughtfully include is for ‘parasitic cities’. Rome being the archetypical parasitic, or ‘consumer’ city (thanks, Moses!) I thought I would try it out. Lo! and behold, an enormous city emerges, with one other major centre….

The model is conceived in the tension between agricultural producers and urban ‘specialists’.  I wonder about the conclusion regarding an upper limit to city growth. The world of the simulation is a torus, so I wonder if there is an effect there. Hmm. On second thought, probably not. It would be interesting to run this simulation on top of a real-world map, rather than the idealized donut of the torus.

An extremely cool model! Check out the pdf with more detail here.