Predictive ACO-ML Framework for Proactive Evacuation Route Adjustment in Dynamic Disaster Environments

Abstract

Disaster logistics and evacuation planning are increasingly challenged by dynamic environmental conditions, infrastructure failures, and uncertain resource demands. Conventional optimization methods remain largely reactive and are unable to adapt proactively to rapidly evolving disaster scenarios. This study developed a Predictive Ant Colony Optimization–Machine Learning (ACO–ML) framework for proactive evacuation route adjustment and disaster logistics optimization in dynamic disaster environments. The study adopted a Design Science Research (DSR) methodology integrated with simulation-based evaluation. The framework was developed by combining Machine Learning prediction models with an enhanced Ant Colony Optimization algorithm. Predictive models estimated road accessibility, traffic congestion, shelter demand, and infrastructure conditions using heterogeneous data sources, while the ACO algorithm generated adaptive evacuation routes. The proposed framework was evaluated through simulation and benchmarked against conventional optimization approaches. The findings showed that integrating predictive Machine Learning with Ant Colony Optimization significantly improved optimization performance by enhancing convergence speed, computational efficiency, route quality, predictive accuracy, scalability, robustness, and adaptability. The framework also improved situational awareness through the integration of multi-source disaster data, enabling proactive route adjustment and more effective disaster logistics and evacuation planning. The study concludes that the proposed Hybrid ACO–ML framework provides a predictive, adaptive, and scalable decision-support solution for disaster logistics and evacuation planning. The framework advances computational disaster management by combining predictive intelligence with swarm optimization and offers a practical foundation for next-generation intelligent disaster response systems, Smart Cities, and AI-enabled emergency management.