A novel emergency evacuation route optimization model in flood disasters using hydrodynamic model and intelligent algorithm

In recent years, flood disasters have intensified due to frequent extreme rainfall events worldwide. Emergency evacuation is regard as the most important response action for reducing the loss of human life in floods. However, there are still few studies on the optimal selection of emergency evacuation routes for flood disasters. This study aims to propose a novel emergency evacuation route optimization model for flood disasters using hydrodynamic model and intelligent algorithms. The hydrodynamic model is applied to simulate the flood evolution and then determine the transfer village and resettlement village, while intelligent algorithms are used to solve emergency evacuation routes between them. First, the TELEMAC-2D couples a novel fine-scale roughness assignment method are selected as the hydrodynamic model for flood evolution simulation, determining more accurate transfer and resettlement villages. Second, the Mask-RCNN image segmentation technique and Thiessen polygons are introduced to solve the missing population data in those villages, which is an important link to determine the emergency evacuation routes. Thirdly, a novel emergency evacuation route optimization model is developed combining intelligent algorithms like ant colony algorithm considering search radius, genetic algorithm coupling indexed matrix and road congestion, greatly improving the selection efficiency of flood emergency evacuation routes between transfer and resettlement villages. The research results are applicable to the emergency evacuation before and during floods, providing advance warning and process evacuation guidance.