The emergence of Zipf's Law in a system of cities: An agent-based simulation approach

This paper develops a spatial agent-based model to generate a system of cities that exhibits the statistical properties of the Zipf's Law. The numerical results suggest that the combination of bounded rationality and maximum heterogeneity of agents can produce a generic power-law relationship in the size distribution of cities, but does not always generate the Zipf's Law. We found sufficient conditions on the probability distribution of spatial reach to generate the Zipf's Law associated with unit power coefficient. Our model also indicates that the Zipf's Law breaks down unless the extent of agglomeration economies overwhelms the negative disagglomerating forces.