Transcript – Designing Optimal Diets via Mathematical Programming

This project is a collaboration between the colleges of engineering and agriculture and life sciences.

The objective is to design customized diet plans, aiming to optimize a given nutritional objective such as calories from fat, while satisfying constraints on various nutrients.

These constraints are based on dietary guidelines provided by the US Department of Agriculture, expert recommendations and personal preferences, the general dietary guidelines and databases outlining detailed nutrient data.

Thousands of food items are published by the US Department of Agriculture, as illustrated on the top third of the poster.

This information, together with personalized dietary goals and daily intake recommendations, is summarized in a mathematical optimization model.

The developed model belongs to a class of mixed integer linear programs.

The formulated model is then solved using a state-of-the-art optimization solver.

The output of the optimization model provides a personalized nutrition recommendation system that makes an optimal selection from the available set of foods.

It can also suggest new dishes and supplements based on the user’s conditions and preferences.

In addition, the output provides insights into how changes and intake recommendations affect the optimal diet plans.

Next, we are planning to evaluate several popular diets with respect to dietary guidelines of the US Department of Agriculture.

We are also planning to study novel network-based models of food systems expressing the relations between foods and diets as networks.

We would like to thank the T3 program for supporting this collaboration.