| 一种新颖的多agent强化学习方法 |
| |
| 引用本文： | 周浦城,洪炳殚,黄庆成. 一种新颖的多agent强化学习方法[J]. 电子学报, 2006, 34(8): 1488-1491 |
| |
| 作者姓名： | 周浦城 洪炳殚 黄庆成 |
| |
| 作者单位： | 哈尔滨工业大学计算机科学与技术学院,黑龙江,哈尔滨,150001;哈尔滨工业大学计算机科学与技术学院,黑龙江,哈尔滨,150001;哈尔滨工业大学计算机科学与技术学院,黑龙江,哈尔滨,150001 |
| |
| 摘 要： | 提出了一种综合了模块化结构、利益分配学习以及对手建模技术的多agent强化学习方法,利用模块化学习结构来克服状态空间的维数灾问题,将Q-学习与利益分配学习相结合以加快学习速度,采用基于观察的对手建模来预测其他agent的动作分布.追捕问题的仿真结果验证了所提方法的有效性.
|
| 关 键 词： | 多agent学习 Q-学习 利益分配学习 模块化结构 对手建模 |
| 文章编号： | 0372-2112（2006）08-1488-04 |
| 收稿时间： | 2005-12-01 |
| 修稿时间： | 2005-12-012006-03-15 |
A Novel Multi-Agent Reinforcement Learning Approach |
| |
| ZHOU Pu-cheng,HONG Bing-rong,HUANG Qing-cheng. A Novel Multi-Agent Reinforcement Learning Approach[J]. Acta Electronica Sinica, 2006, 34(8): 1488-1491 |
| |
| Authors: | ZHOU Pu-cheng HONG Bing-rong HUANG Qing-cheng |
| |
| Affiliation: | School of Computer Science and Technology,Harbin Institute of Technology,Harbin,Heilongjiang 150001,China |
| |
| Abstract: | A novel multi-agent reinforcement learning approach is proposed to learn the coordinated behaviors among cooperative agents team. The proposed approach combines advantages of the modular architecture, profit-sharing learning and opponent modeling technique in a single multi-agent framework. Simulation results on the pursuit problem show that the proposed learning approach has faster convergence speed and more optimal policy over conventional modular Q-learning algorithms. |
| |
| Keywords: | multi-agent learning Q-learning profit-sharing learning modular architecture opponent modeling |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《电子学报》浏览原始摘要信息 |
|
点击此处可从《电子学报》下载免费的PDF全文 |
