论文部分内容阅读
当混沌神经网络的输入发生较大变异时,网络混沌运动偏离了原有混沌吸引域,从而丧失了对原有被储存样本模式的记忆。本文针对神经网络的时空特征,采用混沌控制的钉扎反馈方法,使网络重新恢复记忆。通过对应用例的仿真实验表明,对神经网络时空系统的混沌控制,钉扎反馈控制是一种值得推荐的方法;神经网络的混沌控制增强了网络的容错能力及其鲁棒性,进而提高了混沌神经网络的实用性。
When the input of chaotic neural network changes greatly, the chaotic motion of the network deviates from the original chaotic attracting domain, thus losing the memory of the original stored sample pattern. In this paper, according to the spatio-temporal characteristics of neural network, we use the pinned feedback method of chaos control to restore the network to memory. The simulation experiments show that the chaotic control and pinning feedback control of neural network spatio-temporal system is a recommended method. The chaotic control of neural networks enhances the network’s fault tolerance and robustness, Practicality of Chaos Neural Network.