| 一种基于自适应遗传算法的CMAC的学习率优化方法 |
| |
| 引用本文： | 林旭梅,梅涛.一种基于自适应遗传算法的CMAC的学习率优化方法[J].系统仿真学报,2005,17(12):3081-3084,3088. |
| |
| 作者姓名： | 林旭梅 梅涛 |
| |
| 作者单位： | 1. 中国科学院合肥智能机械研究所,安徽,合肥,230031;中国科学技术大学工程科学学院精密机械与精密仪器系,安徽,合肥,230026
2. 中国科学院合肥智能机械研究所,安徽,合肥,230031 |
| |
| 基金项目： | 国家自然科学基金（50275141） |
| |
| 摘 要： | CMAC（cerebellar model articulation controller）是一种局部逼近神经网络，它发展了近30年，但是关于其学习率的确定仍缺乏好的方法。基于CMAC的控制系统如果没有好的学习率，那么系统就会不稳定或者收敛速度很慢。在传统方法的基础上提出利用遗传算法（GA）来确定其学习率，通过自适应遗传算法（GA）其他传统方法相比较，表明利用自适应遗传算法（GA）不仅使系统稳定，而且收敛速度更快，并进行了仿真验证。
|
| 关 键 词： | 小脑模型 遗传算法 学习率 收敛性能 |
| 文章编号： | 1004-731X（2005）12-3081-04 |
| 收稿时间： | 2004-10-27 |
| 修稿时间： | 2004-10-272005-04-04 |
A kind of Optimization Based on Adaptive GA for Cerebellar Model Articulation |
| |
| LIN Xu-mei,MEI TAO.A kind of Optimization Based on Adaptive GA for Cerebellar Model Articulation[J].Journal of System Simulation,2005,17(12):3081-3084,3088. |
| |
| Authors: | LIN Xu-mei MEI TAO |
| |
| Institution: | 1. Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei 230031, China; 2. Department of Precision Machinery and Precision Instrttmentation, School of Engineering Science, University of Science and Technology of China, Hefei 230026, China |
| |
| Abstract: | Cerebellar model articulation (CMAC) was developed about three decades ago, but yet it lacks good learning rate parameter. Without proper learning rate parameter, the control system based on CMAC will become unstable or its learning will become slowly converged after a period of real time runs. A new kind of optimization based on adaptive GA about learning rate parameter was proposed. The performance of the proposed CMAC was compared with those of conventional CMAC. The experimental results show that performance of the CMAC based the proposed learning rate parameter is stable and more effective than that of the conventional CMAC. |
| |
| Keywords: | Cerebellar model articulation genetic algorithms learning rate convergence property |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
