The Potentials and Limitations of Agent-Based Models for Urban Digital Twins: Insights From a Surveillance and Behavioral Nudging Simulation

Abstract

Although urban digital twins are still at an embryonic stage of development, their use cases are multiple, ranging from big data aggregation to simulations. Additionally, predictions can be rendered and quickly implemented using actuators to transform physical environments and influence urban life. In this article, we investigate the potential of an agent-based model in a smart city setting to predict emergent behavior in relation to the suppression of civil violence by implementing crowd management practices. To this end, we designed a simulation environment that includes cameras in public spaces and wearable sensors, and considers nudging and self-nudging processes supported by a surveillance apparatus. Building on Epstein’s threshold-based model of civil violence, the proposed simulation is informed by surveillance theories and contemplates methods for crowd monitoring and social control. The experiments’ results provide insights into how specific measures and combined actions may influence the suppression of civil violence in public spaces and can be useful to inform crowd management activities and policymaking. Moreover, we use the simulation to reflect upon the potentials and limitations of integrating agent-based models into urban digital twins and emphasize the imminent risks for individuals and democratic societies of employing a ubiquitous surveillance apparatus endowed with the autonomy to trigger actuators.