Deep Reinforcement Learning-Based Approaches for Optimizing Healthcare Transportation
Résumé
Recent advancements in Artificial Intelligence (AI), particularly in Deep Reinforcement Learning (DRL) [2], offer promising alternatives for solving Ambulance Routing Problem (ARP) [3]. DRL enables agents to learn optimal policies directly from interactions with the environment, bypassing the need to re-run algorithms for every instance. In contrast to classical solutionbased methods (heuristics and metaheuristics) [5], DRL is a policy-based approach where, once trained, the agent can generalize to solve new instances without requiring retraining. This characteristic is particularly advantageous for large-scale ARPs, where diverse instances are common and solution speed is critical. Additionally, DRL's potential to handle dynamic problems, such as stochastic demand and travel times, makes it a compelling approach for real-world ARPs.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |