一类离散反馈型时滞细胞神经网络模型及其数值模拟

A Class of Discrete-feedback Delayed Cellular Neural Network Model and its Numerical Simulation

细胞神经网络是一个大规模非线性系统,具有高速并行计算且易于硬件实现等特点,已被广泛应用于模式识别,图像处理等领域。本文基于时滞方程和反馈控制,对离散反馈型细胞神经网络进行设计,给出了满足相应功能的模板及其动力学分析,提出了一类离散反馈型时滞细胞神经网络,并借助模拟平台Matlab 7.0对基础的随机布朗运动进行了初步的数值模拟,数值模拟结果充分显示了布朗运动的随机性,证实了理论分析的有效性。该模型经过一定时间后能对粒子的随机状态作出反馈,可用于模拟二维平面上的随机行走。

Cellular neural network is a large-scale nonlinear system, which has the advantages of high-speed computation and easily adapting hardware to realize, and which was applied to image processing and pattern recognition. Based on the time delay differential equations and feedhack control, this paper designed a class of discrete feedback cellular neural networks, given the corresponding function template and its dynamics analysis, proposed a class of discrete feedback delayed cellular neural networks. With the help of simulation platform Matlab 7.0, this paper conducted a preliminary numerical simulation. The simulation results show the random of Brownian motion fully confirmed the validity of the theoretical analysis. The model can make particles random state feedback after certain time and can be used to simulate the random walk of the two-dimensional plane, which has important significance for the de- velopment of CNN and relevant subjects.