Complex Systems

Strong and Weak Spatial Segregation with Multilevel Discrimination Criteria Download PDF

Philippe Collard
Université Côte d'Azur, CNRS, I3S, France
UNS, Parc Valrose, 06108 Nice, France


This paper is influenced by the research of Thomas Schelling on spatial segregation; in his seminal work on the subject he used simple simulations to show that even highly tolerant individuals end up being spatially aggregated far beyond the local requirement of their tolerance level. In this paper we are not seeking to find the conditions, in terms of density of population and tolerance level, that lead to a global stable state where all the individuals are satisfied in view of their own neighborhoods. Here the context is: (i) a space full of agents where each individual is in continual contact with a maximum number of neighbors; and (ii) where both a principal and a secondary discrimination criterion compel people to leave their places. As, in general, the first hypothesis does not allow the population to converge within the meaning of Schelling, only incomplete segregation phenomena are observable. So the problematic will be to determine, according to the respective strength of the two discrimination criteria, the spatial repartition of the agents resulting from their moves; in such a general context, we will refer to segregation as being strong or weak or even mixed.

Keywords: spatial segregation; computational sociology; agent-based simulation; multilevel discrimination criteria