A New Method of Dish Innovation Based on User Preference Multi-objective Optimization Genetic Algorithm

With the improvement of living level, people put forward new requirements for the diversification of diet and greater demand for new dishes. However, it is hard to make food collocation to meet specific requirement, since there are too many foodstuffs, while their nutrition ingredients and incompatibility are not well known to the ordinary people. To solve this problem, food collocation and dish creation to meet the user’s requirement or preference are studied in the paper. First, the data of food composition are collected, the different food guides are referenced and the food component incompatibility is studied. Second, a food nutrition evaluation model is constructed and an improved non-dominated sorting genetic algorithm is proposed. A probability operator is introduced, by analyzing the existing recipes, to control the number of foodstuffs of a dish. A strategy to model user preference is also proposed and the non-dominated solutions are filtered by using the preference model. Third, the experiments are carried out and the experiment results show that the proposed algorithm and nutrition evaluation model can meet the requirements of user preferred dish creation and multi-objective optimization, and has better convergence speed than the original algorithm.