Genetic Algorithm Predicts Blended Oil Formulations with Improved Nutrition, Prolonged Frying Life, and Low Cost

A good frying oil is selected based on several criteria including thermal stability, fatty acid nutrition, and cost among many other parameters. The industry uses oil blending to achieve these diverse parameters. However, methods that also include maximizing frying life are rare. Here, we used a database comprising thermal stability, fatty acid profile, and cost of oils to build a genetic algorithm that generates the fittest oil that also maximizes frying life. Optimal solutions from the genetic algorithmic model had frying stability similar to palm oil but with reduced saturated fat. We validated the frying life with negative controls and palm olein through heating and frying experiments; our results confirmed the accuracy of the modeled solutions. The utility of the model to predict blended oil formulae with better frying performance, better fatty acid nutrition requirement, and lower cost can be meaningful in modern societies where health and economic factors are important considerations.