| 一种优化多层前向网络的IA-BP混合算法 |
| |
| 引用本文： | 吕岗,陈小平,赵鹤鸣. 一种优化多层前向网络的IA-BP混合算法[J]. 计算机工程与应用, 2003, 39(27): 27-28 |
| |
| 作者姓名： | 吕岗 陈小平 赵鹤鸣 |
| |
| 作者单位： | 苏州大学电子信息学院,苏州,215021 |
| |
| 基金项目： | 国家自然科学基金(编号:60172016)，江苏省高校自然科学研究计划(编号:02KJB51001) |
| |
| 摘 要： | 该文针对免疫算法(IA)在优化较大规模的多层前向神经网络时收敛速度慢的缺点,给出了一种综合免疫算法和BP算法优点的IA-BP混合算法,它首先采用免疫算法进行全局搜索,然后调用BP算法进行局部搜索,从而加快收敛速度。实验结果表明该算法在训练较大规模的前向神经网络时性能要优于免疫算法和BP算法。
|
| 关 键 词： | 神经网络 免疫算法 BP算法 |
| 文章编号： | 1002-8331-(2003)27-0027-02 |
| 修稿时间： | 2003-06-01 |
An IA-BP Hybrid Algorithm to Optimize Multilayer Feed-forward Neural Networks |
| |
| Lv Gang Chen Xiaoping Zhao Heming. An IA-BP Hybrid Algorithm to Optimize Multilayer Feed-forward Neural Networks[J]. Computer Engineering and Applications, 2003, 39(27): 27-28 |
| |
| Authors: | Lv Gang Chen Xiaoping Zhao Heming |
| |
| Abstract: | Aiming at the shortcoming that the speed of convergence will be slow when the size of feed-forward neural network increases,this paper brings up an IA-BP hybrid algorithm combining immune algorithm with BP algorithm.It first utilizes the immune algorithm to have a global search,and then uses BP algorithm to have a local search,which will improve the speed of convergence.Experiments show that the hybrid algorithm have better capability than immune algorithm and BP algorithm in training larger size of feed-forward neural network. |
| |
| Keywords: | Neural network Immune algorithm BP algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
