Optimization of Theanine-Glucose Maillard Reaction Conditions Based on Artificial Neural Network Coupled Genetic Algorithm (BP-GA)

In order to comprehensively evaluate the Maillard reaction system and optimize the conditions of theanine-glucose Maillard reaction, an artificial neural network coupled genetic algorithm (BP-GA) model for the comprehensive evaluation value of theamine-glucose Maillard reaction was herein constructed, and the best reaction conditions were optimized. Firstly, the weight among five evaluation indexes, fluorescence value, A₂₉₄, A₄₂₀, glucose and theanine residue were obtained by entropy method. Next, a BP-GA artificial neural network model with experimental factor parameter as the input and the comprehensive evaluation value of Maillard reaction as output was established in combination with the results of the single factor experiment and orthogonal experiment. Finally, the optimal condition for the theanine-glucose Maillard reaction and the corresponding comprehensive evaluation value of the Maillard reaction were obtained with the established model. The results showed that the optimal conditions for the theanine-glucose Maillard reaction obtained through the global optimization by the established BP-GA model were: Reaction temperature 117.6 ℃, reaction time 1.8 h, pH7.3, and carbonyl ammonia ratio 1:2. In such case, the comprehensive evaluation value of Maillard reaction was 93.22. Compared with the results obtained by the orthogonal experiment, it was found that the BP-GA model had good predictive performance. In this study, the optimum conditions for the model Maillard reaction was obtained by the BP-GA model, which would provide a reference for the establishment of the Maillard reaction simulation system and prediction of the reaction.