Gerald F. Frasco, Jie Sun, Hernan D. Rozenfeld, Daniel ben-Avraham
We propose a bare bones stochastic model for the geographical distribution of people within a country, including also their complex network of connections. The model, which is designed to give rise to a scale-free network of connections and to visually resemble the geographical spread seen in satellite pictures of earth at night, gives rise to a Zipf distribution for the ranking of cities by population size, and re ects the notion that highly connected individuals tend to live in highly populated areas. It also yields some interesting insights regarding Gibrat's law for the rates of city growth (by population size), in partial support of the ndings in a recent analysis of real data [Rozenfeld et al., PNAS 105, 18702 (2008)], and exhibits a non-trivial relation between city population and city population density which seems quite in line with real data.
View original:
http://arxiv.org/abs/1306.0257
No comments:
Post a Comment