Safe Multi-Agent Reinforcement Learning via Shielding
This work addresses safety guarantees in MARL for safety-critical applications, representing an incremental improvement over existing methods.
The paper tackles the problem of ensuring safety in multi-agent reinforcement learning (MARL) for critical applications by proposing two shielding approaches: centralized shielding monitors all agents' joint actions, and factored shielding uses multiple shields based on state space factorization. Experimental results show that both approaches guarantee safety during learning without compromising policy quality, with factored shielding being more scalable in agent numbers.
Multi-agent reinforcement learning (MARL) has been increasingly used in a wide range of safety-critical applications, which require guaranteed safety (e.g., no unsafe states are ever visited) during the learning process.Unfortunately, current MARL methods do not have safety guarantees. Therefore, we present two shielding approaches for safe MARL. In centralized shielding, we synthesize a single shield to monitor all agents' joint actions and correct any unsafe action if necessary. In factored shielding, we synthesize multiple shields based on a factorization of the joint state space observed by all agents; the set of shields monitors agents concurrently and each shield is only responsible for a subset of agents at each step.Experimental results show that both approaches can guarantee the safety of agents during learning without compromising the quality of learned policies; moreover, factored shielding is more scalable in the number of agents than centralized shielding.