论文部分内容阅读
提出一种新的神经网络模型——标准神经网络模型(SNNM),并给出基于线性矩阵不等式(LMI)的SNNM平衡点的全局渐近稳定性定理。通过状态的线性变换,将推广的离散BAM神经网络转化为SNNM,利用SNNM的稳定性结论,判定该离散BAM的全局渐近稳定性。该方法扩展了以前的稳定性结果。保守性低,容易验证,同时也适用于其它类型的递归神经网络的稳定性分析。
A new neural network model - Standard Neural Network Model (SNNM) is proposed. The global asymptotic stability theorem of SNNM equilibrium points based on linear matrix inequality (LMI) is given. Through the linear transformation of state, the generalized discrete BAM neural network is transformed into SNNM, and the global asymptotic stability of the discrete BAM is judged by using the stability of SNNM. This method extends the previous stability results. Conservative low, easy to verify, but also applies to other types of recursive neural network stability analysis.