摘要
本文提出了一种实数值联想记忆模型，可以处理实数值的训练模式，并能确保训练模式的稳定性．对于记忆离散双值模式，该模型相当于一个高阶神经元网络，并指出Ｈｏｐｆｉｅｌｄ模型只是它的特例．本文并将该模型应用于多属性目标决策和等级划分中，其结果令人满意．
Abstract This paper presents a new real associative memory model(RAM model)that can deal with the training patterns and that is sta-bk. In cases of storing binary patterns, the RAM model is equivent to a highorder neural network,and it includes the Hopfield net.This paper also discusses the application of RAM model in multi attribute decision making and grade differentiating, and the satisfactory results are obtained.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
1995年第5期56-62,共7页
Systems Engineering-Theory & Practice
