离散神经网络模型平衡点的稳定性及分支.pdfVIP

中文摘要 本文主要研究动力系统中几类神经网络模型平衡点的稳定性及分支情况。全 文共分五章。 第一章简单介绍了神经网络的学科发展背景及离散神经网络模型的研究现 状，并给出了本文将要用到的一些符号和定义。 第二章，第三章分别研究了二元、四元离散神经网络模型的平衡点稳定性及 局部 Neimark-Sacker-分支情况，并利用 Matlab 数学软件，对文中讨论的模型进 行数值模拟，证明了本文结论的有效性和可行性。 第四章我们讨论了一类三元具时滞离散神经网络模型，利用数学分析技巧， 对线性化系统的特征根进行分析，获得了平衡点的局部稳定性及分支点，通过选 取适当的参数，讨论了该模型平衡点的稳定性，Pitchfork- 分支，Flip- 分支和 Neimark-Sacker-分支，并用中心流形理论得出了决定分支方向和稳定性的计算公 式。 第五章总结了本文所做的工作，得到的主要结果及对离散神经网络模型的进 一步研究方向。 关键词：离散系统；神经网络；时滞；稳定性；Pitchfork 、Flip 、Neimark-Sacker 分支 I Abstract The stability of the equilibrium point and bifurcation for a kind of discrete neural network model is concerned in this thesis. The thesis contains five chapters. As the introduction, in Chapter 1, the develop background and research status of discrete neural networks model are briefly addressed, and some notations and definitions are given in this thesis. The stability of the equilibrium point and Neimark-Sacker bifurcation for discrete neural networks of two 、four neurous are studied in Chapter 2、3, respectively. The calculating formulae of direction and stability of the bifurcation are obtained by using the normal form theory and the center manifold theorem. By using the mathematical software of Matlab, results of thesis are displayed graphically which show the validity and feasibility of the theory. In Chapter 4, we discussed a kind of three neurous delay discrete neural networks model. By using the skill of mathematical analysis, we analysis the characteristic roots of corresponding linearization system and obtain local stability of the equilibrium point and bifurcation point. Choosing an appropriate