Abstract: Model-free deep reinforcement learning has emerged as a promising method for addressing the scheduling challenges in integrated energy systems. However, uncertainty in system states ...