基于离散泰勒级数的新型联想记忆系统

Associative memory system based on discrete taylor series

基于离散泰勒级数提出一种对任意阶多维函数可实现无差逼近的新型联想记忆系统,详细讨论了该系统的插值算法、训练规则及寻址机制.新型联想记忆系统相对于CMAC具有逼近精度更高、占用存储单元较少、学习速度较快等优点.该系统在信号处理、模式识别、过程建模及高精度实时智能控制等领域具有广泛的应用价值.仿真实例表明了系统的可行性与有效性.

This paper proposes a novel highorder associative memory system based on the Discrete Taylor Series(DTSAMS),which includes the interpolation algorithm and training algorithm. It is capable of implementing errorfree approximations to multivariable polynomial functions of arbitrary order. The advantages are:highprecision of learning and much smaller memory requirement. The novel associative memory system has great potential in the application areas of signal processing, pattern recognition, process modeling and implementation of highprecision realtime intelligent controllers. The simulation results verify that DTSAMS is feasible and efficient.