Network Creation with Homophilic Agents

TL;DR

This paper studies how heterogeneity and homophily influence network formation, revealing that stable networks tend to be similar across different models and providing insights into social segregation phenomena.

Contribution

It introduces and compares two models of network creation with homophilic agents and non-uniform costs, highlighting the robustness of resulting network structures.

Findings

Stable networks are similar across models despite different initial conditions.

Homophily leads to emergent social segregation.

Theoretical and experimental analysis confirms robustness of network structures.

Abstract

Network Creation Games are an important framework for understanding the formation of real-world networks. These games usually assume a set of indistinguishable agents strategically buying edges at a uniform price leading to a network among them. However, in real life, agents are heterogeneous and their relationships often display a bias towards similar agents, say of the same ethnic group. This homophilic behavior on the agent level can then lead to the emergent global phenomenon of social segregation. We initiate the study of Network Creation Games with multiple types of homophilic agents and non-uniform edge cost. Specifically, we introduce and compare two models, focusing on the perception of same-type and different-type neighboring agents, respectively. Despite their different initial conditions, both our theoretical and experimental analysis show that the resulting stable networks are almost identical in the two models, indicating a robust structure of social networks under homophily. Moreover, we investigate the segregation strength of the formed networks and thereby offer new insights on understanding segregation.