Abstract: A novel fuzzy-neural network (FNN) was proposed in this paper. A new type of neural network construction incorporating fuzzy rules was formed, in which the fuzzy sub - space was designed by homogenized design and could be adjusted. The ability of the fuzzy-neural network to simulate and predict was improved, and the time for training was shortened correspondingly. The kind of network could be used to simulate or optimize bioprocesses. In this paper, the fuzzy-network was used to predict the glutamic acid production by fed - batch fermentation, which gave satisfactory results.

摘要： 提出了一种子空间可调的模糊-神经网络(FNN),通过采用新型的网络结构设计、模糊子空间均匀化设计和模糊子空间可调性设计,提高了网络的模拟、预测精度,明显缩短了训练时间.通过可调模糊-神经网络可以对系统进行预测、优化及因素分析,并将所提出的模糊神经网络应用于谷氨酸流加发酵过程的模拟与预测,取得了满意的结果.