多能源管理系统的自我完善硬约束条件下安全的强化学习（英文版）

英文标题：Safe reinforcement learning with self-improving hard constraints for multi-energy management systems中文摘要：本文介绍了两项新的安全强化学习方法，OptLayerPolicy 和 self-improving hard constraints，将约束函数与 RL 形式解耦，以提高初始效用和准确性，提供了在模拟的多能源系统案例研究中实现 92.4%（OptLayerPolicy）的初始效用和 104.9%（GreyOptLayerPolicy）的策略的结果。英文摘要：Safe reinforcement learning (RL) with hard constraint guarantees is apromising optimal control direction for multi-energy management systems. Itonly requires the environment-specific constraint functions itsel