Simulating optimal flood evacuation using heuristic algorithms and path-choice behaviors

Effective path planning in flooding emergency rescue scenarios is essential for ensuring timely evacuation while minimizing safety risks. Conventional path-planning algorithms often prioritize the shortest or most cost-efficient routes, potentially neglecting safety considerations. To address this limitation, this study introduces an improved path-planning method using a behavior-based A-star (A*) algorithm designed for evacuation scenarios. A cellular automata (CA) environment is applied to address common challenges associated with traditional A* algorithms, including path inefficiencies, longer distances, and difficulties in handling dynamic flood environments. The key innovation of this study is the optimization of a heuristic function by integrating depth sensitivity perception (DSP), which directly influences evacuation behavior by prioritizing safer paths based on real-time water depth assessments during path selection. Experimental results across diverse flood scenarios demonstrate that the optimized A* algorithm significantly outperforms traditional A-star and Dijkstra’s algorithms, achieving reductions in explored nodes by 90.06 % and 93.13 %, lowering safety risks, and shortening computational times by 87.65 % and 88.06 %, respectively. These findings validate the efficacy of the depth-sensitive heuristic in enhancing evacuation pathfinding within complex flood environments.