An Agent-based Model for Earthquake Pedestrians’ Evacuation Simulation in Urban Scenarios

The earthquake risk assessment approach actually ignores human behaviors during earthquake. Nevertheless, simulating pedestrians’ motion could be useful to introduce “human” interactions with post-earthquake scenarios. This work proposes an agent-based model for evacuation simulation based on the analysis of videotapes concerning real events. Modifications to the social force model are provided in order to describe typical behaviors. A simulation software is developed for model validation. Tests mainly involve speeds and distances between individuals. The model could be integrated in tools for the analysis of probable pedestrians’ choices in different scenarios, and checking solutions for reducing man-environment interferences during the evacuation process.