抄録
Food ingredients' tastes were measured by a taste sensor and a optimization problem for food ingredients and their blend ratios was solved using a genetic algorithm(GA) to fit the food taste vector to a target taste one. First, the tastes of food ingredients were measured by a taste sensor. Next, chromosomes used by GA were set using genes defined by the serial numbers on the ingredients. Third, the problem was reduced to blend ratio optimization problems for the ingredients with the serial numbers given by the chromosomes, and the reduced problems were solved to fit the food taste vector to a target taste one. Using the fitness function of the chromosome defined by the taste error of the reduced problem, the chromosomes were improved using GA. Approximate sparse solutions of the taste optimization problem for more than thousand food ingredients were thus derived. Using this algorithm, optimum food ingredients and their blend ratios were obtained for a target taste vector.
