Multi-agent systems consist of multiple AI agents that interact and collaborate to solve complex tasks. Advantages include parallel task execution, improved problem-solving through collaboration, and robustness in decentralized environments. They are widely used in simulations and distributed systems.