A simulation framework for emergency evacuation, considering navigation errors

Abstract

The rapidly growing tourism industry has brought forth substantial safety concerns, particularly in the context of emergency evacuations necessitated by natural and human-induced disasters. Tourists often lack the necessary orientation, information, and preparedness, rendering them vulnerable during such crises. While research has extensively explored tourist behavior and evacuation procedures independently, the intersection of these two fields remains underexamined.  This study introduces a simulation model utilizing MATSim to characterize the behaviors of tourists during emergencies, highlighting navigation errors and decision-making processes. Two novel routers - “Random Walk” and “Landmark Assisted” - have been developed to better reflect tourist navigation challenges. A case study in Conegliano, Italy, demonstrates the effectiveness of these routers under two evacuation policies: predefined and undefined destinations. Results indicate significant disparities in evacuation times: optimal routes average 50 minutes, while random navigation extends this to 544 minutes. The Landmark Assisted router improves evacuation to 73 minutes, underscoring the importance of identifiable landmarks. Additionally, managing intersections further reduces evacuation times. This simulation framework serves as a decision-making tool for evaluating evacuation policies, providing insights into optimizing resource allocation and enhancing overall efficacy in emergency scenarios. Future research should focus on developing optimization algorithms for intersection management selection, reinforcing the practical applicability of this model in real-world contexts.