Mathematical programming models have been developed to find utility maximized nutritious diets. The utility maximized diets have been shown to be more practical, more realistic, and less costly than the non-utility maximized diets. In this study we further increase the desirability of diets by gradually softening those nutrients which have been supplied by the foods in the diet at a cost which is above the market price of the respective nutrient. Marginal analysis is used to find marginal costs of the nutrients. The parametric quadratic programming is used to find the optimum diet where the cost of nutrient is equal to the market price of the nutrient supplement. The resulting diet is more desireable and less costly than the previous diets.