Un modelo flexible para la simulación de distribución de ciudades

Abstract

The fulfillment of the Zipf's law is widely recognized in urban economics literature. Indeed, its coefficient is used as a measure of spatial concentration of a population. Whereas the concentration excess has problems associated with social welfare due to the existence of regional disparities and the congestion in some cities, the dispersion does not take advantage of, for example, the benefits of reducing costs of production generated by economies of scale. Several studies have been performed in order to give an explanation to the fulfillment of Zipf's law, where the top-down models, that force the fulfillment of Zipf's law, have dominated the literature. Some proposals have attempted to explain the fulfillment of Zipf using cellular automata. These bottom-up proposals generate the distribution and sizes of cities in an emergent way. However, these models are not flexible and extensible, therefore, these proposals cannot be adapted to different geographical areas. This paper proposes an initial implementation of a flexible and extensible model based on cellular automata, called CityCA, which tries to explain the Zipf coefficient. CityCA was evaluated and adapted to a real scenario, Chile. The simulation results were as expected, since they show that the ratio is closer to Zipf over-concentration than the dispersion, as in reality. With this model and implementation, urban economists do not require advanced programming knowledge, since CityCA already integrates specific knowledge.