[摘要]：Mathematical diet optimization models are used to create food plains that best resemble current eating habits while meeting prespecified nutrition and cost constraints. This study used linear programming to generate food plans meeting the key 2007 dietary recommendations issued by the World Cancer Research Fund/American Institute of Cancer Research (WCRF/AICR). The models were constructed to minimize deviations in food intake between the observed and the WCRF/AICR-recommended diets. Consumption constraints were imposed to prevent food plans from including unreasonable amounts of food from a single group. Consumption norms for nutrients and food groups were taken from dietary intake data for a sample of adult men and women (n = 161) in the Pacific Northwest. Food plans meeting the WCRF/AICR dietary guidelines numbers 3-5 and 7 were lower in refined grains and higher in vegetables and fruits than the existing diets. For this group, achieving cancer prevention goals required little modification of existing diets and had minimal impact on diet quality and cost. By contrast, the need to meet all nutritional needs through diet alone (guideline no. 8) required a large food volume increase and dramatic shifts from the observed food intake patterns. Putting dietary guidelines into practice may require the creation of detailed food plains that are sensitive to existing consumption patterns and food costs. Optimization models provide an elegant mathematical solution that can help determine whether sets of dietary guidelines are achievable by diverse U.S. population subgroups. J. Nutr. 139: 1541-1548, 2009.
