Finding evacuation routes using traffic and network structure information

Given the unpredictable nature of transportation networks, a strategic disaster evacuation plan involving taking detours, which avoids congested and less connected links, is of practical importance. It guarantees the seamless evacuation of evacuees to safe exits while reducing the chance that their safety is jeopardized. This study develops a vehicle evacuation algorithm that employs a link-based centrality metric to dynamically identify agile evacuation paths based on the topological features and the changing spatio-temporal traffic concentration on links within the network. By employing real traffic volume data corresponding to the Yeouido Island's transport network in Seoul, Korea, we demonstrate that, relative to the conventional shortest path approach, considering the structural and operational features of transport networks using a link-based betweenness centrality measure has a better evacuation performance during high evacuation demand periods by lowering network clearance time. Extensive experimental results, pertinent findings, and recommendations for developing effective evacuation plans are discussed.