Churning, power laws, and inequality in a spatial agent-based model of social networks

Abstract

Amidst many previous network models lacking a spatial dimension, this paper proposes a dynamic agent-based model of social network formation that explicitly considers space. We find that varying the dynamics of agent interaction causes the emergence of differential degree distributions as well as nonlinear dynamics in social and spatial inequalities. The scale-free property of degree connectivity vanishes when tie formation dominates tie dissolution, with power laws re-emerging when tie dissolution is of equal strength or stronger than tie formation. Furthermore, we find a nonlinear relationship between network density and agent inequality in social resources. In particular, multiple phase transitions occur where the relationship is positive in one phase but negative in another. This suggests that, contrary to intuition, higher connectivity can have an adverse distributional impact by benefiting the already privileged. Critically, we find a tradeoff between agent inequality and spatial inequality where the geographic concentration of social resources accompanies a more equal distribution of connectivity. Finally, the disadvantage of agents with limited spatial reach is exacerbated as network density increases. Our results thus highlight the importance of distinguishing between social and spatial inequality in policymaking.

Bridging of different sites by bohemians and tourists: analysis by agent-based simulation

“…As independent parameters, we changed the number of duals, or d, and the other parameters (f, p, q). The number of duals d is prepared from 0 to 100, with 10 digits, because the number of agents is 10 for each value of resources w (6)(7)(8)(9)(10)(11)(12)(13)(14)(15). For each w, we prepare the same number of duals, from 0 to 10; singulars and duals are the same with respect to w. During the simulation, duals may select the same site for residence and visits.…”

“…Accordingly, if individuals want to form networks with other individuals through which they can profit, the former should escalate into areas of aggregation, which further increases population density and proximate interaction in the areas. Based on these findings, an assumption of studies has been that proximity is the necessary condition for network formation [1,15,41].…”

Generating random bigraphs with preferential attachment

“…In this respect, agent-based models have become of increasing importance in different fields of application regarding the modeling and simulation of complex systems (see, for example, [2,3,4,5]). Within this computational model, individual entities of a complex system are modeled as agents, whereas the behavior and interaction between them are determined by a set of predefined rules.…”