摘要
在样本的学习过程中引人了求解复杂优化问题比较有效的多级优化技术,将一个复杂的多层神经网络权值优化问题分解成多个相对简单的优化子问题,然后利用迭代的策略进行求解。并提出了一种改进的BP算法,提高了学习效率。
The paper makes research on the BP algorithm used in Multilayer Neural Networks Multievel optimization technique, Which is used to solve the complex cptimization problems effectively, is introduced to the problems of the trainning patterns Some optimization subproblems are built up based on the multievel optimization echnique. An improved BP algorithm is presented and it has more efficiency than the original BP algorithm.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
1994年第6期39-44,共6页
Journal of Harbin Institute of Technology
关键词
多层神经网络
神经网络
BP算法
Multilayer Neural Network
BP Algorithm
Artificial Neural Network