DECIDING THE TEMPERATURE COURSE DURING SAKE MASHING USING A GA-FNN FOR QUALITY-CONTROL OF SAKE

Simulation models for Baume and alcohol concentration from the 11th da y to the end of the sake mashing were constructed using a fuzzy neural network (FNN). The models could simulate the time courses of Baume an d alcohol concentration in 17 actual sake mashings. Average errors at the ends of the mashings were 0.22 and 0.40% for Baume and alcohol con centration, respectively. By applying a genetic algorithm (GA) with th e simulation models, temperature time courses were calculated with goo d accuracy, and the target values for Baume and alcohol concentration on the final day could be achieved. To make a variety of sakes with di fferent qualities, temperature courses were calculated against 3 targe t values: higher (+0.3), ordinary (0.0), and lower (-0.3) final day Ba umes. The calculated temperature courses were found to be similar to a Toji's (expert's) strategy for making decisions on temperature. By ap plying this procedure, quality control of sake can be realized.