Continuous utility factor in segregation models

TL;DR

This paper introduces a continuous utility factor in a Schelling segregation model, revealing significant differences in segregation patterns and phase transitions between two proposed models, with implications for understanding social clustering.

Contribution

It proposes two models with continuous utility factors, demonstrating how they affect segregation behavior and cluster formation differently from traditional discrete models.

Findings

Model B produces larger segregated clusters.

Model A results in smaller clusters and frozen states.

Model B shows no frozen states even at low tolerance.

Abstract

We consider the constrained Schelling model of social segregation in which the utility factor of agents strictly increases and non-local jumps of the agents are allowed. In the present study, the utility factor u is defined in a way such that it can take continuous values and depends on the tolerance threshold as well as the fraction of unlike neighbours. Two models are proposed: in model A the jump probability is determined by the sign of u only which makes it equivalent to the discrete model. In model B the actual values of u are considered. Model A and model B are shown to differ drastically as far as segregation behaviour and phase transitions are concerned. In model A, although segregation can be achieved, the cluster sizes are rather small. Also, a frozen state is obtained in which steady states comprise of many unsatisfied agents. In model B, segregated states with much larger cluster sizes are obtained. The correlation function is calculated to show quantitatively that larger clusters occur in model B. Moreover for model B, no frozen states exist even for very low dilution and small tolerance parameter. This is in contrast to the unconstrained discrete model considered earlier where agents can move even when utility remains same. In addition, we also consider a few other dynamical aspects which have not been studied in segregation models earlier.