| 基于改进的遗传算法神经网络控制器 |
| |
| 引用本文： | 陈华艳. 基于改进的遗传算法神经网络控制器[J]. 兵工自动化, 2007, 26(6): 66-67 |
| |
| 作者姓名： | 陈华艳 |
| |
| 作者单位： | 广东药学院医药信息工程学院,广东广州510006 |
| |
| 摘 要： | 基于遗传算法的神经网络控制器,对神经网络中连接权值采用实值编码,通过随机选择,可变变异率和Elitism方法形成改进遗传算法.以倒立摆系统神经网络控制器参数寻优为例,对二进制编码和实值编码的遗传算法进化过程进行仿真,结果表明本算法寻优效果优,在多参数、多目标、多约束条件下,寻优效果更佳.
|
| 关 键 词： | 遗传算法 神经网络控制器 实值编码 Elitism方法 |
| 文章编号： | 0006-0576（2007）06-0066-02 |
| 收稿时间： | 2007-04-12 |
| 修稿时间： | 2007-04-122007-05-17 |
Neural Network Controller Based On Genetic Algorithm |
| |
| CHEN Hua-yan. Neural Network Controller Based On Genetic Algorithm[J]. Ordnance Industry Automation, 2007, 26(6): 66-67 |
| |
| Authors: | CHEN Hua-yan |
| |
| Affiliation: | College of Medical Information Engineering, Guangdong Pharmaceutical University, Guangzhou 510006, China |
| |
| Abstract: | Based on genetic algorithm, NN controller uses real coding, stochastic sampling, invariable mutation rate and elitism method to improve genetic algorithm. Take inverse pendulum system as an example, it uses binary coding genetic algorithm and real coding genetic algorithm to simulate. The results show that the method has good optimization effect, and under multi-parameter, multi-object and multi-constraint conditions, it is more feasible. |
| |
| Keywords: | Genetic algorithm(GA) Neural nets controller Real coding Elitism method |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
