摘要
利用饱和域的特性,和拟对角列支配矩阵与M-矩阵之间的关系,获得了非对称细胞神经网络(简称CNNS)的稳定平衡点存在的两个充分条件;在此基础上,通过定义一个更高阶的神经网络模型,推广了该网络存在稳定平衡点的结果;通过大量的模拟仿真,提出了非对称细胞神经网络完全稳定的充分条件,并就二细胞神经网络的情况给予了证明;最后,将本文的结果与已取得的结果进行了比较,并给出一个实例说明本文取得的结果优于已有文献取得的结果。
Two sufficient conditions of the existence of stable equilibrium points of nonsymmetric cellular neural networks （CNNC for short） are obtained by using the characteristics of saturation region and the relationship between quasidiagonally column dominant and M-matrix. In this basis, through defining a much higher order cellular neural networks model, the result of the existence of of stable equilibrium points of the networks is extended. The paper given some sufficient conditions of complete stability of nonsymmetric cellular neural networks based on plenty of simulation and proves the case of two-cell CNNS. Finally the paper shows with comparison and examples that its results are better than others＇.
出处
《电子测量与仪器学报》
CSCD
2005年第3期25-29,共5页
Journal of Electronic Measurement and Instrumentation
基金
湖南省自然科学基金资助项目(01JJY3029)
