Real-time heuristic evacuation pathfinding algorithm for flood avoidance

Floods rank among the world’s most devastating annual disasters in terms of casualties and economic losses, making evacuation route planning crucial for mitigating flood risks. This study proposes a Real-time Heuristic Flood Evacuation Path Planning Algorithm (RHEPFA) based on dynamic flood simulations using the ANUGA 2D model. Its core innovation lies in a novel dual-range (dangerous and safe) heuristic weighting mechanism that efficiently anticipates flood progression. Uniquely coupled with hydrodynamic forecasts, RHEPFA conducts real-time pathfinding based on moving speeds, terrain constraints, and inundation dynamics. This innovative approach maintains over 90% accessibility, which outperforms static Dijkstra and ant colony optimization algorithms. Compared to the dynamic Dijkstra algorithm, RHEPFA still demonstrates outstanding performance with 12.5% higher accessibility, 22.8% shorter path length, 81.3% lower computational cost, and more importantly, 10.43% longer evacuation time window. Entrapment rate is further found to have a negative correlation with evacuation speed, as taking a faster transit means can significantly reduce entrapment rates.