The overuse of hospitals in Burkina Faso, especially during emergencies, is largely due to chronic underfunding and weak coordination among health centers. These challenges create critical bottlenecks in emergency care, where resources are limited and demand is often unpredictable. Motivated by the need for improved management of emergency patients in resource-constrained urban settings, this study focuses on optimizing emergency resource allocation at the city level through a hybrid methodological approach. Our framework is organized into three complementary steps. First, Linear Programming (LP) is applied to establish a theoretical basis for optimal resource allocation, providing an initial structure for decision-making. Second, Particle Swarm Optimization (PSO) is introduced to extend the LP model by incorporating nonlinear constraints and enabling flexible redistribution of resources when they are unevenly available across hospitals. Finally, Artificial Neural Networks (ANN) are integrated to support real-time decision-making under uncertainty, allowing the system to adapt dynamically to evolving emergency conditions. The results from the deployed platform highlight the effectiveness of this mixed-method design. The ANN model achieves an F1-score of up to 87% and correctly addresses over 76% of emergency cases, while the PSO component significantly enhances allocation flexibility across multiple hospitals. By combining mathematical modeling, metaheuristic optimization, and machine learning, our approach improves both the efficiency and fairness of emergency care delivery in Burkina Faso, offering a scalable model for similar low-resource health systems.

Adaptation Models, Uncertainty, Hospitals, Decision Making, Urban Areas, Artificial Neural Networks, Linear Programming, Mathematical Models, Resource Management, Particle Swarm Optimization, Linear Programming, Particle Swarm Optimization, Artificial Neural Networks, Emergency Resource Allocation, Healthcare Optimization, Burkina Faso