模糊联想推理及其实现  

Fuzzy Associative Inference and Its Implementation

作　　者：肖平[1] 邓达[1] 余英林[1] 

机构地区：[1]华南理工大学电子与通信工程系,广州510640

出　　处：《控制理论与应用》1999年第4期562-565,共4页Control Theory & Applications

基　　金：国家"攀登计划";国家自然科学基金!NSC92097

摘　　要：本文绘出了用模糊联想记忆网络实现肯定前件式，否定后件式或同时包含这两种模糊推理形式的充要条件，并提出了一个增加网络神经元的增强学习算法，这种学习算法能够可靠有效地用于任意多个基于规则的不同形式模糊推理．In this paper, the sufficient and necessary conditions for fuzzy inference with affirmative premise or negativeconsequence or both, to be implemented with fuzzy associative memory networks,is proposed. Based on these onditions, a kindof leaming algorithm by increasing neurons is also presented, which is called as augmenting leaming and is efficient and reliablemethod to implement the three sorts of fuzzy infereare with single system.

关 键 词：模糊推理 联想记忆 增强算法 神经网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]