Evaluation of Robust Evacuation Strategies for Resilient Urban Infrastructure Through Microscopic Traffic Simulation

Abstract

Natural disasters are a global threat, highlighting the urgent need for effective disaster management systems worldwide. Many countries, both developed and developing, are not adequately prepared, emphasizing the importance of governmental action. Key to disaster management is the creation of specialized disaster management units that develop and implement rapid response plans for potential risks. A crucial aspect of disaster management is evacuation-the process of moving vulnerable populations to safer areas. However, evacuations face challenges such as timely alert issuance, traffic congestion, resident reluctance to evacuate, and potential damage to transportation infrastructure. These challenges can be mitigated through comprehensive evacuation plans that ensure smooth relocation to shelters. This paper addresses these issues by developing and evaluating traffic routing conditions in an evacuation study area using the microscopic simulator SUMO. It examines two algorithms, Dijkstra and A-star (A*), which optimize vehicle routes under different network conditions. By focusing on criteria such as Minimum Travel Time and Maximum Number of Evacuations (clearance time), the research aims to improve disaster response and resilience. The objective is to enhance evacuation procedures, thereby strengthening disaster management and ensuring the safety of affected populations. Results show that the A* algorithm outperforms Dijkstra, reducing travel times by up to 18% and network clearance times by up to 6.8% under optimal conditions. The Manhattan-based network design further enhances evacuation efficiency, reducing average waiting time by up to 35% compared to the actual map.