相空间压缩法实现混沌神经网络联想记忆

Realization of associative memory using chaotic neural network via phase space compression

给出了利用相空间压缩法控制混沌神经网络,使得网络能够收敛于存储的目标模式的充分条件和必要条件.通过数学分析,得到了相空间压缩控制方法中对应参数的上下限;并通过对仿真结果的分析,提出了通过改变相空间压缩控制方法中对应的参数来实现混沌神经网络联想记忆的新方法.以上结果均通过仿真得到验证.

We develop the necessary and sufficient conditions for a chaotic neural network to converge to the stored target patterns. The upper and lower limits of the corresponding parameters in the phase space compression control method are obtained. An associative memory scheme is proposed through the analysis of simulation results. The scheme is realized by changing the corresponding parameters in the phase space compression control method. These results are validated by simulations.