| 一种基于演化算法的BP改进算法 |
| |
| 引用本文： | 李康顺,戴志晃.一种基于演化算法的BP改进算法[J].微计算机信息,2010(5). |
| |
| 作者姓名： | 李康顺 戴志晃 |
| |
| 作者单位： | 赣州江西理工大学;中国科学院自动化研究所; |
| |
| 基金项目： | 项目名称:基于粒子动力学演化算法的演化路由研究;基金颁发部门:江西省科学技术厅、江西省自然科学基金项目(2008GZS0028); 项目名称:基于神经网络的演化硬件研究;基金颁发部门:江西省教育厅、江西省研究生创新专项资金项目(YC08B029) |
| |
| 摘 要： | 本文在对BP神经网络算法分析的基础上,提出一种基于演化算法的BP改进算法(EBP)。该算法将演化算法运用到BP算法学习率的求解中,从而达到学习率的自适应、自组织的目的。实验结果表明,使用EBP算法进行求解函数逼近、优化和建模等BP神经网络应用问题,都要比传统的BP算法具有更好的精确度和收敛速度,并且能够克服传统BP算法易陷入局部最优解、学习过程出现震荡等缺点。
|
| 关 键 词： | 演化算法 自适应学习率 BP算法 |
An improved BP algorithm based on evolutionary algorithm |
| |
| LI Kang-shun DAI Zhi-huang.An improved BP algorithm based on evolutionary algorithm[J].Control & Automation,2010(5). |
| |
| Authors: | LI Kang-shun DAI Zhi-huang |
| |
| Affiliation: | LI Kang-shun DAI Zhi-huang (School of Information Engineering,Jiangxi University of Science & Technology,Ganzhou 341000,China) (Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China) |
| |
| Abstract: | In this paper,an improved BP algorithm based on evolutionary algorithm (EBP) is proposed.this algorithm greatly improves the traditional BP algorithm and has many good characteristics by using evolutionary algorithm to solve the learning rate of BP algorithm,so it can make the learning rate adaptive and self-organization.The experiment results show that it has not only better accuracy and convergency by using EBP algorithm to solve function approximation problems,optimization problems and modeling,compared ... |
| |
| Keywords: | evolutionary algorithm adaptive learning rate BP algorithm |
| 本文献已被 CNKI 等数据库收录！ |
