Andries Smit
Papers
1
Total Citations
12
H-Index
1
About
Andries Smit is a leading researcher in multi-agent reinforcement learning (MARL), with a particular focus on scaling complex coordination tasks to real-world-inspired domains. His most-cited work, "Scaling multi-agent reinforcement learning to full 11 versus 11 simulated robotic football" (2023, 12 citations), tackles one of AI’s grand challenges: enabling a full team of autonomous agents to learn cooperative strategies in a high-dimensional, adversarial environment. Smit’s key contribution lies in demonstrating that MARL can outperform traditional heuristics and handcrafted rules in the RoboCup simulated football league, where agents must coordinate in real time under partial observability. By successfully scaling learning to 11 versus 11 settings, his work bridges the gap between toy problems and practical multi-agent systems. This achievement not only advances the state of the art in robotic football but also provides a scalable framework for other multi-agent coordination tasks, such as autonomous driving and swarm robotics. Smit’s research is foundational for students and practitioners aiming to deploy MARL in complex, real-world scenarios.
Research Focus
Key Achievements
Top Papers
- 1