论文部分内容阅读
首先将连续型双向联想记忆神经网络转化成一个特殊的 Ｈｏｐｆｉｅｌｄ网络模型．在此基础上，对连续ＢＡＭ神经网络的指数稳定性进行了新的分析，证明了神经网络连接权矩阵在给定的约束条件下有唯一平衡点．所做的分析可以用于设计全局指数稳定的神经网络．
First, the continuous bidirectional associative memory neural network is transformed into a special Hopfield network model. Based on this, a new analysis of the exponential stability of continuous BAM neural networks is given. It is proved that the neural network connection weight matrix has a unique equilibrium point under the given constraints. The analysis can be used to design a global exponential stable neural network.