节点文献
一类梯形激活函数的细胞神经网络的静态解及Smale马蹄
The stationary solutions of a class of cellular neural networks with trapezoidal activation function and Smale horseshoe
【摘要】 研究了一类以梯形激活函数为输出函数的细胞神经网络.通过将a-z平面分解成25个不交的区域,分析了这类细胞神经网络的静态解.其中,a和z(即反馈系数和阈值)被视为参数.当给定参数时,每个区间内该类细胞神经网络的静态解的范围被确定.针对一类特殊的梯形激活函数的细胞神经网络,研究了其静态解诱导的映射产生的Sm ale马蹄的充分条件,说明了该映射的复杂性.
【Abstract】 A class of cellular neural network(CNN) with trapezoidal activation function(CNNwTAF) was discussed.With the a-z plane been partitioned into 25 disjoint open regions,where parameters a and z stand for the output synaptic weight and system threshold respectively,the stationary solutions of CNNwTAFs were discussed and some useful results were obtained.Moreover,for a special kind of CNNwTAF,the sufficient conditions of the existence of Smale horseshoe induced by the stationary solutions′ maps were studied,which in part implied the complexity of the maps.
【Key words】 cellular neural network; trapezoidal activation function(TAF); parameters space; stationary solutions; Smale horseshoe;
- 【文献出处】 浙江师范大学学报(自然科学版) ,Journal of Zhejiang Normal University(Natural Sciences) , 编辑部邮箱 ,2008年01期
- 【分类号】O175;O193
- 【被引频次】1
- 【下载频次】73