基于模型的动态屏蔽技术，用于安全高效的多智能体强化学习（英文版）

英文标题：Model-based Dynamic Shielding for Safe and Efficient Multi-Agent Reinforcement Learning中文摘要：该论文提出了一种基于模型的动态屏蔽（MBDS）方法来支持多智能体强化学习算法设计，同时在强化学习和部署阶段实现形式化安全性保证。该算法合成分布式屏蔽，可以在与每个 MARL 代理并行运行的情况下监视和纠正不安全行为，从而实现对多智能体复杂环境的有效监控，并具有强有力的安全性保证。英文摘要：Multi-Agent Reinforcement Learning (MARL) discovers policies that maximizereward but do not have safety guarantees during the learning and deploymentphases. Although shielding with Linear Temporal Logic (LTL) is a promisingformal method to ensure safety in