Medical Drone Cloud with Binary Weather-Dependent Failures: An RL Approach to Dynamic Decision Making

This paper presents a Medical Drone Cloud (MDC) model for emergency healthcare with dynamic binary failures influenced by weather conditions. Unlike traditional models, we formulate the problem as a Markov Decision Process (MDP) with explicit weather-dependent failure states and solve it using Reinforcement Learning (RL). The model maintains closed-form transition probabilities while enabling adaptive decision-making under varying weather conditions. Numerical results demonstrate the significant impact of request arrival rate λ on system performance and allocation strategy, providing insights for real-time medical applications.