项目名称: 分数阶忆阻系统的动力学分析及其控制
项目编号: No.61463050
项目类型: 地区科学基金项目
立项/批准年度: 2015
项目学科: 自动化技术、计算机技术
项目作者: 周小兵
作者单位: 云南大学
项目金额: 46万元
中文摘要: 忆阻器,作为电子电路中的第四种基本元件,由于其巨大的应用潜力,引起了世界范围内的强烈关注。本项目借助非线性系统理论、分数阶微积分理论、泛函微分方程理论、现代控制理论等,研究分数阶忆阻系统的建模、动力学特性分析及其控制。结合非线性解析解分析、数值仿真,对分数阶忆阻系统的非线性动力学机理进行研究,以期发现尚未被充分揭示的动力学现象。将时滞引入系统,发现新的动力学现象以及分数阶阶次和时滞对系统动力学影响的机制。利用现代控制理论,构造有效的控制器对分数阶忆阻系统进行控制。将分数阶忆阻系统应用到递归神经网络中,结合一类基于分数阶忆阻的线性阈值神经网络,通过网络的局部抑制和局部不变集理论,拓展传统的递归神经网络的模式记忆理论。
中文关键词: 忆阻器;忆阻系统;分数阶;神经网络;系统动力学
英文摘要: Memristor, as the fourth element of electrical circuit, has arouse much focus for its tremendous potential applications. Based on the nonlinear theory, fractional calculus theory, functional diffential equation theory and modern control theory, this project investigates the modelling, dynamical characteristics analysis and control of fractional memristor systems. Combining the anlysis of nonlinear analytic solutions, numerical simulations, we investigate the nonlinear dynamical mechanism of fractional systems, in order to uncover some unknown dynamical behaviors. We introuduce time delay into fractional memristor system, aim to find some new dynamical phenomena and the affection mechanism on system dynamics of order and delay. With the help of modern control theory, effective controllers are contructed to stabilize fractional memristor systems. To apply fractional memristive system in recurrent neural networks, and combining fractional memristor based linear threshold neural networks, by the theory of topoinhibition and local invariant, to expand the traditional theory of pattern memory in recurrent neural networks.
英文关键词: memristor;memristive system;fractional order;neural networks;system dynamics