| 自适应RBF网络Q学习控制 |
| |
| 引用本文： | 徐明亮,须文波.自适应RBF网络Q学习控制[J].控制与决策,2010,25(2):303-306. |
| |
| 作者姓名： | 徐明亮 须文波 |
| |
| 作者单位： | 江南大学信息工程学院,江苏无锡214122; |
| |
| 基金项目： | 国家自然科学基金项目(60703106) |
| |
| 摘 要： | 利用RBF网络逼近连续空间的Q值函数,实现连续空间的Q学习.RBF网络输入为状态-动作对,输出为该状态-动作对的Q值.状态由系统的状态转移特性确定,动作由优化网络输出得到的贪婪动作与服从高斯分布的噪声干扰动作两部分叠加而成.利用RNA算法和梯度下降法自适应调整网络的结构和参数.倒立摆平衡控制的实验结果验证了该方法的有效性.
|
| 关 键 词： | RBF网络 自组织 Q学习 连续空间 优化 |
| 收稿时间： | 2009/3/24 0:00:00 |
| 修稿时间： | 2009/6/1 0:00:00 |
Q-learning Control Based on Self-organizing RBF Network |
| |
| XU Ming-liang,XU Wen-bo.Q-learning Control Based on Self-organizing RBF Network[J].Control and Decision,2010,25(2):303-306. |
| |
| Authors: | XU Ming-liang XU Wen-bo |
| |
| Affiliation: | School of Information Technology/a>;Jiangnan University/a>;Wuxi 214122/a>;China. |
| |
| Abstract: | The radial basis function (RBF) neural network is used to approache the Q-value function. The information learnt is generalized by learning agent in continuous state space and action space. The input of RBF network is the pair of state and action,and the output is the Q-value of the pair of state and action. The state is decided by the transfer characteristic of system. The act of the input is consisted of the greedy act,which can be calculated with the Q-value optimization and noise act which has a normal ... |
| |
| Keywords: | RBF network Self-organization Q-learning Continuous space Optimization |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《控制与决策》浏览原始摘要信息 |
|
点击此处可从《控制与决策》下载全文 |
|
