MARLadona -- Towards Cooperative Team Play Using Multi-Agent Reinforcement Learning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913744156622848 |
|---|---|
| author | Li, Zichong Bjelonic, Filip Klemm, Victor Hutter, Marco |
| author_facet | Li, Zichong Bjelonic, Filip Klemm, Victor Hutter, Marco |
| contents | Robot soccer, in its full complexity, poses an unsolved research challenge. Current solutions heavily rely on engineered heuristic strategies, which lack robustness and adaptability. Deep reinforcement learning has gained significant traction in various complex robotics tasks such as locomotion, manipulation, and competitive games (e.g., AlphaZero, OpenAI Five), making it a promising solution to the robot soccer problem. This paper introduces MARLadona. A decentralized multi-agent reinforcement learning (MARL) training pipeline capable of producing agents with sophisticated team play behavior, bridging the shortcomings of heuristic methods. Furthermore, we created an open-source multi-agent soccer environment. Utilizing our MARL framework and a modified global entity encoder (GEE) as our core architecture, our approach achieves a 66.8% win rate against HELIOS agent, which employs a state-of-the-art heuristic strategy. In addition, we provided an in-depth analysis of the policy behavior and interpreted the agent's intention using the critic network. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_20326 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MARLadona -- Towards Cooperative Team Play Using Multi-Agent Reinforcement Learning Li, Zichong Bjelonic, Filip Klemm, Victor Hutter, Marco Multiagent Systems Robot soccer, in its full complexity, poses an unsolved research challenge. Current solutions heavily rely on engineered heuristic strategies, which lack robustness and adaptability. Deep reinforcement learning has gained significant traction in various complex robotics tasks such as locomotion, manipulation, and competitive games (e.g., AlphaZero, OpenAI Five), making it a promising solution to the robot soccer problem. This paper introduces MARLadona. A decentralized multi-agent reinforcement learning (MARL) training pipeline capable of producing agents with sophisticated team play behavior, bridging the shortcomings of heuristic methods. Furthermore, we created an open-source multi-agent soccer environment. Utilizing our MARL framework and a modified global entity encoder (GEE) as our core architecture, our approach achieves a 66.8% win rate against HELIOS agent, which employs a state-of-the-art heuristic strategy. In addition, we provided an in-depth analysis of the policy behavior and interpreted the agent's intention using the critic network. |
| title | MARLadona -- Towards Cooperative Team Play Using Multi-Agent Reinforcement Learning |
| topic | Multiagent Systems |
| url | https://arxiv.org/abs/2409.20326 |