节点文献

具自反馈和连续型信号函数的二元神经网络模型的定性研究

【作者】 刘艳青

【导师】 黄立宏;

【作者基本信息】 湖南大学 , 应用数学, 2001, 硕士

【摘要】 本文研究具自反馈的二元神经网络模型: x=-x+f(y(t-τ))+g(x(t-τ)), y=-y+f(x(t-τ))+g(y(t-τ)),的动力学性质,这里,信号函数f,g是R→R上的连续可微单调递增的有界函数(例如λtanh(x),(1/1+e-βx)都是这类函数)。我们分三章对模型(Ε)进行了定性研究。第一章研究了τ=0时模型方程收敛点的唯一性和全局吸引性,多个平衡点的稳定性以及收敛域的分界。在第二章中研究了τ>0时模型方程平衡点的存在性和稳定性问题以及稳定域。在第三章中我们研究了时滞对稳定性的影响。

【Abstract】 This thesis deals with the dynamics of the following neural networks of two neurons with self -feedback.~= —x + f(y(t — r))+ g(xQ — r)),(E)‘I. ~ —y + f(x(t — T)) + g(y(t —where, signal functions f and g are continuous, dfferentiable, increasing and bounded, such as 2. tanh(x),The thesis consists of three chapters. In chapter 1, we consider the model (F) with time delay r = 0 and study the unique and global attractability of stable equilibria of the system, furthermore , we discuss the stability of equilibrias and the boundary between the basins of attraction of two stable equilibria of the system. In chapte2, we consider the case where T >0 in the model (F), we study the existence, stability basin of attraction of the equilibrias of the system. In chapter 3, we obtain the result to show that the stability of neural network (F) may be influenced by time delay.

  • 【网络出版投稿人】 湖南大学
  • 【网络出版年期】2002年 01期
  • 【分类号】TP183
  • 【被引频次】1
  • 【下载频次】76
节点文献中: 

本文链接的文献网络图示:

本文的引文网络