AGENT-BASED MODELING FRAMEWORK FOR WILDFIRE EVACUATION IN DAMAGED TRANSPORTATION SETTINGS

Abstract

The main goal of this project was to support effective evacuation planning by developing an agent-based modeling (ABM) framework for wildfire evacuation in damaged transportation settings. More specifically, the framework integrates wildfire simulation and vulnerability assessment with ABM to adequately represent both human behaviors during an evacuation and time-dependent network functionality in microscopic traffic simulation. The framework predicts traffic conditions during an evacuation and identifies the critical parts of the transportation network for pre-fire risk mitigation actions aimed at improving mobility during a wildfire evacuation. The proposed framework is illustrated with the City of Santa Clarita, affected by the Rye Fire, to demonstrate its applicability to a real community. The contribution of this project is twofold: (a) The framework incorporates an advanced wildfire hazard modeling and vulnerability assessment to improve the accuracy of wildfire evacuation in damaged transportation settings; and (b) This project constructs an evacuee response model based on a stated preference survey to predict individual evacuees’ behaviors as a firefront approaches.