Researchers Develop AI-Powered Adaptive Evacuation Routing for Shooting Events

GPEvac is a research framework using graph neural networks (GNNs) and PPO to compute adaptive evacuation routes during shooting events, accounting for crowd dynamics and adversarial uncertainty.

GPEvac outperforms intelligent baselines in reducing total threat exposure across different architectural layouts, and it is significantly faster than traditional computationally intensive methods.

Yes, the system can integrate with live surveillance systems due to its low latency (14.73 ms), making it suitable for smart cities and critical facilities in the region.

The methodology is transferable to other domains such as critical infrastructure, intelligent transportation systems, and adaptive sensor networks.