论文部分内容阅读
运用云理论,建立自适应神经-云推理系统(adaptive neuro-cloud inference system,ANCIS)的控制模型,并证明该模型具有全局逼近性质。首先将Zadeh模糊推理神经网络变为云推理网络,建立一个多输入单输出的T-S型ANCIS模型;然后设计系统变量的属性函数和推理规则,确定输入输出关系以及系统输出结果的计算表达式;最后通过证明所建模型的输出结果计算式满足Stone-Weirstrass定理的3个假设条件,完成该模型的全局逼近性证明。
By using cloud theory, a control model of adaptive neuro-cloud inference system (ANCIS) is established and its global approximation property is proved. Firstly, the Zadeh fuzzy reasoning neural network is transformed into a cloud reasoning network, and a multi-input single-output TS-type ANCIS model is established. Then the attribute functions and inference rules of the system variables are designed to determine the input-output relationship and the output expression of the system. Finally, The global approximation of the model is proved by proving that the output of the model is satisfied with the three assumptions of Stone-Weirstrass theorem.