【摘要】 将负规则引入到模糊控制器规则基中的每一条正规则,扩展了削减矢量的类型,使之不再局限于单位矢量。同时将削减因子的可行域予以适当扩展,提出了广义削减因子反模糊化方法(GDA)。选择拥有折衷参数的BADD法计算训练数据的反模糊化输出期望值,将GDA与DA在最小二乘意义下进行了对比。数值结果显示由GDA所得的反模糊化输出比DA所得的能在更大程度上逼近期望值。三组训练数据的训练结果均能说明GDA优于DA。更多还原

【Abstract】 In this paper, the incorporation of negative rules is introduced into each positive rules of fuzzy inference system. At the same time, the interval in which the attenuation factors distribute is extended. A new defuzzification method which is called General dot attenuation (GDA) is presented here. This method is a generalization of the dot attenuation method (DA). The advantages of the GDA method over the DA method are validated by three numerical examples.更多还原