Safe and Generalizable Hierarchical Multi-Agent RL via Constraint Manifold Control
Addresses the safety-performance trade-off in multi-agent systems for safety-critical applications.
Proposed a hierarchical multi-agent RL framework that enforces hard safety constraints via a constraint manifold, achieving nearly perfect safety rates and competitive performance while generalizing to varying numbers of agents and obstacles.
Multi-agent systems are widely used in safety-critical applications that require coordinated behavior under strict safety constraints. Existing approaches face a fundamental trade-off: learning-based methods achieve strong empirical performance but lack theoretical safety guarantees, while control-theoretic methods enforce safety but often lead to overly conservative and inefficient behaviors. We propose a hierarchical multi-agent reinforcement learning framework that enforces hard safety constraints under mild assumptions at low level via a constraint manifold, while enabling effective coordination through high-level policy learning. Our approach provides theoretical safety guarantees in the multi-agent setting and yields stationary learning dynamics, thereby enabling stable and efficient training. Empirically, our method achieves competitive performance while maintaining nearly perfect safety rates, and generalizes effectively to varying numbers of agents and obstacles.