应用微粒群算法优化以变结构策略为学习算法的神经网络初始权值

来源期刊:中南大学学报(自然科学版)2011年第z1期

论文作者:夏志峰 韩宁 王亚慧

文章页码:169 - 174

关键词:人工神经网络;微粒群算法;滑模变结构控制

Key words:artificial neural network; particle swarm algorithm; quasi-sliding mode variable structure control algorithm

摘    要:应用变结构算法训练人工神经网络权值的问题,分析了初始权值对该算法训练效果的影响,在此基础上中提出了一种应用微粒群算法优化变结构训练的神经网络初始权值的方法,即将随机产生的初始权值通过经微粒群算法优化后再采用双滑模变结构算法进行训练从而提高网络的整体性能。应用此方法训练Adaline神经网络,并将训练后的神经网络用于MIMO系统辨识。研究结果表明:与未经优化的变结构学习算法相比,该方法显著提高了网络的训练精度和泛化能力。

Abstract: Variable structure learning strategy was used to train artificial neural network weights, the importance of initial weights in this process. Based on the analysis, in order to improve the performance of neural network, a method by which the weights were first initialized using particle swarm optimization and then trained by means of multi-objective sliding mode control algorithm. This method was applied to train the Adaline neural network which was used to identify the MIMO system. Comparing with the original algorithm whose initial weights were not optimized, the experimental results show that the proposed method efficiently improves the neural network’s precision and the ability of generalization.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号