To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Multi-Agent Battlefield Game with Federated Reinforcement Learning |
|
Author |
Fardeen Hasib Mozumder |
| Citation |
Vol. 25 No. 11 pp. 29-34 |
|
Abstract |
This paper discusses the efficacy of federated learning in solving the Battlefield game from the petting zoo simulator for unseen and new environments utilizing multi-agent reinforcement learning. The game simulates a 12v12 battle in an open field with walls of different landscapes. The work incorporated federated learning by forming a global model by averaging the parameters of a few models trained locally on different landscapes utilizing a suitable reinforcement learning algorithm. In the simulation, the performance of the global model relative to local models within a new environment has been evaluated to assess the advantages of federated learning in these settings. The experimental results demonstrated that the global model consistently outperformed local models trained in a non-federated manner, thereby validating the effectiveness of federated learning in such environments. |
|
Keywords |
Reinforcement Learning, Federated Learning, Battlefield, Multi-Agent Learning. |
|
URL |
