PrioritEvac, an adaptive model for evacuation: agent based simulation of the Station nightclub fire

Abstract

Fire evacuation modeling benefits from application of social science both in terms of accuracy and greater possibility of external validation. This thesis describes a novel simulation framework, PrioritEvac, which incorporates social dimensions of fire evacuation in an agent-based modeling framework. This model uses individual priorities, making for a dynamic approach that allows greater agency and nuance. The agent-based model was programmed in NetLogo and then validated using data from the Station nightclub fire, finding that it is in good agreement on multiple metrics.