A proximal space approach for embedding urban geography into CA models

In the great majority of urban models based on Cellular Automata (CA), the concept of proximity is assumed to reflect two fundamental sources of spatial interaction: (1) the accessibility of places and (2) the distance "as the crow flies". While the geographical space defined by the latter clearly has an Euclidean representation, the former, based on the accessibility, does not admit such a regular representation. Very little operational efforts have been undertaken in CA-based urban modelling to investigate and provide a more coherent and cogent treatment of such irregular geometries, which indeed are essential and crucial feature of urban geography. In this paper, we suggest an operational approach - entirely based on cellular automata techniques - to model the complex topology of proximities arising from urban geography, and to entangle such proximity topology with a CA model of spatial interactions. © 2010 Springer-Verlag.