基于神经网络的模糊综合评价

Neural network based fuzzy synthetical evaluation

在模糊系统理论与神经网络理论基础上，建立了一种基于神经网络的模糊综合评价模型，其合成运算采用加权平均算子“－·”，其隶属函数调整采用误差反传算法．并证明了这种算子的模糊综合评价神经网络能收敛到唯一的极小点．还给出了以三角模糊函数为隶属函数的模糊综合评价神经网络的隶属函数调整算法．

Based on the theory of fuzzy system and neural network, this paper constructs a kind of neural network based fuzzy synthetical evaluation model. its synthetic arithmetic adopts the operator “-·”, and the regulation of the active function adopts the error feedback arithmetic. It is proved that this kind of a network can converge to a unique point. When adopting thriangular fuzzy function as its membership function, it also gives a regulating algorithm of the membership function.