Agent-Based Models for Megacities and Social Networks in Disaster

Increasingly agent-based models are being used to study human behavior in response to mass emergencies and disasters. These models have been applied to questions regarding the effects of natural disaster, emergency and humanitarian response, shelter versus evacuation, and crowds and riots. However, developing realistic control populations and the social networks of large-scale urban populations remains a key challenge for research and experimentation. Modelers must balance the need for representative, heterogeneous populations with the computational costs of developing large population sets. These models must also include the social network relationships that influence social interactions and behavioral patterns in emergencies. To address this we use a set of methods and empirical census data to build a synthesized population with social networks embedded in an agent-based modeling environment of the New York megacity. The methodology and modeling code are available and openly shared. The computational framework provides a laboratory for testing the effects of varying disaster impacts and social networks on the survival outcomes and movement of populations in a Civil Defense emergency.