
反应扩散递归神经网络全局指数稳定的鲁棒分析
ROBUSTNESS ANALYSIS FOR GLOBAL EXPONENTIAL STABILITY OF REACTION DIFFUSION RECURRENT NEURAL NETWORKS
主要讨论反应扩散递归神经网络全局指数稳定的鲁棒分析. 给定反应扩散递归神经网络是全局指数稳定, 首先, 在此神经网络基础上考虑噪音扰动, 利用超越方程得到噪音密度的上界, 在上界范围内, 带噪音的反应扩散递归神经网络仍然是全局指数稳定. 进一步, 在反应扩散递归神经网络基础上同时考虑噪音扰动和连接权参数不确定, 利用超越方程得到连接权参数和噪音密度上界, 在两个参数描述的超越方程范围内, 带噪音和连接权参数不确定的反应扩散递归神经网络仍然是全局指数稳定. 最后给出数值算例证实相关理论的有效性.
In this paper, we analyze the robustness of global exponential stability of reaction diffusion recurrent neural networks. Given globally exponentially stable reaction diffusion recurrent neural networks, the upper bound of noise intensity is given by using the transcendental equation. The reaction diffusion neural networks with noise preserve global exponential stability if noise intensity is smaller than the upper bound. The upper bound of noise intensity and parameter uncertainty in connect weight is also obtained by using the transcendental equations. The reaction diffusion neural networks with uncertainty in connect weight and noise can remain to be globally exponentially stable when noise intensity and parameter uncertainty in connect weight are in the inner of the closed curve described by the transcendental equation. A numerical example is provided to illustrate the theory.
全局指数稳定 / 反应扩散递归神经网络 / 鲁棒性 / 噪音 / 超越方程. {{custom_keyword}} /
/
〈 |
|
〉 |