摘要
本文对连接权矩阵为分块三角阵的神经网络的稳定性进行了研究,证明了连接权矩阵为严格三角阵的神经网络具有全局稳定性。最后构造了一种具有全局稳定性的人工神经网络A/D转换电路。
The stability of the neural networks with block triangular connections is considered, then the globally stability of the neural networks with strict triangular connections is proved, finally, a new kind of globally stable A/D converter is designed in this paper. (3 refs.)
出处
《南京邮电学院学报》
北大核心
1993年第2期73-76,共4页
Journal of Nanjing University of Posts and Telecommunications(Natural Science)
关键词
神经网络
稳定性
模-数变换
Artifical neural networks,Stability, Analog digital conversion
