Smac

The StarCraft Multi-Agent Challenge (SMAC) is a multi-agent reinforcement learning (MARL) benchmark platform built on the StarCraft II game engine. It is designed to provide an ideal environment for researching the performance of cooperative multi-agent systems by incorporating elements such as partial observability, dynamic challenges, and high-dimensional observation spaces. Each game unit acts as an independent reinforcement learning agent, working together to accomplish complex tasks, thereby advancing the development and application of MARL algorithms.