| 基于多Agent的并行Q-学习算法 |
| |
| 引用本文： | 周浦城,洪炳,韩学东,郭耸.基于多Agent的并行Q-学习算法[J].小型微型计算机系统,2006,27(9):1704-1707. |
| |
| 作者姓名： | 周浦城 洪炳 韩学东 郭耸 |
| |
| 作者单位： | 1. 哈尔滨工业大学,计算机科学与技术学院,黑龙江,哈尔滨,150001
2. 哈尔滨工程大学,计算机科学与技术学院,黑龙江,哈尔滨,150001 |
| |
| 摘 要： | 提出了一种多Agent并行Q-学习算法．学习系统中存在多个Agent，它们的学习环境、学习任务及自身功能均相同，在每个学习周期内，各个Agent在各自独立的学习环境中进行学习，当一个学习周期结束后，对各个Agent的学习结果进行融合，融合后的结果被所有的Agent共享，并以此为基础进行下一个周期的学习．实验结果表明了该方法的可行性和有效性。
|
| 关 键 词： | 强化学习 Q-学习 融合 |
| 文章编号： | 1000-1220（2006）09-1704-04 |
| 收稿时间： | 06 14 2005 12:00AM |
| 修稿时间： | 2005-06-14 |
Parallel Q-Learning Algorithm Based on Multiple Agents |
| |
| ZHOU Pu-cheng,HONG Bing-rong,HAN Xue-dong,GUO Song.Parallel Q-Learning Algorithm Based on Multiple Agents[J].Mini-micro Systems,2006,27(9):1704-1707. |
| |
| Authors: | ZHOU Pu-cheng HONG Bing-rong HAN Xue-dong GUO Song |
| |
| Affiliation: | 1.School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China;2.School of Computer Science and Technology, Harbin Engineering University, Harbin 150001, China |
| |
| Abstract: | Proposed a kind of multi-agent parallel Q-learning algorithm. There are multiple agents in the learning system. Each agent exists in an independent environment. However, these environments are identical, and the tasks and capability of each agent are also identical. In a learning episode, each agent learns independently. After a learning episode, the results of all agents are fused so as to achieve common result, which are shared by all agents in turn as the learning basis for the next learning episode. Experiments show the feasibility and validity of the given algorithm. |
| |
| Keywords: | reinforcement learning Q-learning fusion |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
