Multi-Agent Training

Reinforcement learning (RL) for real-time strategy (RTS) games has long been bottlenecked by slow simulation, rigid environment interfaces, and high…