Abstract: We present a novel control strategy for automated and individualized mechanical ventilation based on a Reinforcement Learning (RL) controller for a Constrained Markov Decision Process (CMDP) ...