基于代数神经网络的置换多项式判定网络学习算法  

Based on Algebra Neural Networks of Tests for Permutation Polynomials and Learning Algorithms

作　　者：周永权[1] 马武瑜[1] 

机构地区：[1]广西民族学院数学与计算机科学系,南宁530006

出　　处：《计算机工程与应用》2001年第17期102-103,146,共3页Computer Engineering and Applications

摘　　要：从神经元的运算特性入手，对神经元的激发函数，网络结构，学习目标三方面进行了推广，设计出了一类用于有限域上置换多项式判定的多项式神经网络模型，它们是单输入单输出的3层神经网络。给出了两类置换多项式判定的离散网络模型学习算法，该算法简单可行，易于实现。： Start with the neural of operation characteristic，the neurons of stimulate function，networks construction and learning objective were popularized in finite field．A kind of neurons model of apply to tests for permutation polynomials is designed．It is single input and single output three layers polynomials neural networks．Two kinds of hopfield networks and learning algorithms are applies to test for permutation polynomials are designed．The algorithms are efficiency and simple．

关 键 词：代数 神经网络 置换多项式 学习算法 有限域 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]