Modeling and Optimization of Rice Wine Fermentation Process Based on Hierarchical Adaptive Network-based Fuzzy Inference System
Dengfeng Liu, Guoqing Jiang, Wenjing Guo, Xibiao Xu, Haifeng Ding · 2023
Abstract. Rice wine fermentation process is a nonlinear, multi-input multi-output, complex biochemical reaction process. Modeling the rice wine fermentation process can realize the optimization control of the process and improve the quality of rice wine. Among various modeling methods, fuzzy system provides a reasonable framework for modeling by decomposing nonlinear systems into a set of local linear models. Multi-Output Adaptive Network-based Fuzzy Inference System (MOANFIS) is one of its branch methods, but it faces the problem of fuzzy rule explosion, which makes the model complex and reduces its interpretability. In this paper, a hierarchical structure of ANFIS, namely H-ANFIS, is proposed. Firstly, the number of fuzzy rules is reduced by using the Dropout method to discard nodes with small weights. Then, the Extreme Learning Machine (ELM) algorithm is used to expand the output dimension of ANFIS and solve the model parameters, creating the ANFIS-ELM module. Based on this module, a two-layer H-ANFIS model is proposed, where the output of the first layer is used as the augmented input feature of the second layer, expanding the input information of the second layer. The model achieves a good balance between interpretability, robustness, and accuracy.