Operations Research 2025
Abstract Submission

2388. Towards application maturity of the classical diet model by integrating accurate nutrition data, personalization aspects and enhanced dietary variety

Invited abstract in session WC-5: Multiobjective Decision Making and Integer Programming, stream Decision Theory and Multi-criteria Decision Making.

Wednesday, 13:30-15:00
Room: H7

Authors (first author is the speaker)

1. Mick Stewart Wörner
2. Taieb Mellouli
Business Information Systems and Operations Research, Martin-Luther-University Halle-Wittenberg

Abstract

The classic Stigler-Dantzig diet model optimizes a mix of ingredients or meals to meet specified nutrient requirements. Previous studies of OR and nutritional scientists lack of realistic pricing and nutrient data and/or of fulfilling diverse, user-specific criteria. We formulate a linear model with both meal and ingredient levels, integrate real-world recipe and ingredient-to-nutrient data, and account for nutritional adequacy, user preferences, and balanced meal solutions. Ingredient price and nutrient data are taken from Open Food Facts database with prices regionally limited to Germany and France and nutrient values cross-validated using the USDA FoodData Central. Health-oriented nutrient constraint bounds are based on intake recommendations from the U.S. National Institutes of Health (NIH) and the World Health Organization (WHO), tailored by gender, age, and activity level.
User preferences are modelled at both ingredient and recipe levels, distinguishing between exclusion-based constraints, e.g. veganism, and inclusion-based preferences, e.g. fish consumption or cultural cuisines. Piecewise linear penalties to discourage excessive repetition of recipes and cuisines, promoting dietary variety. Besides their individual level, nutrients are also constrained at group-level, e.g. limiting total fat to less than 30% of total caloric intake, saturated fats to less than 10%, and trans fats to less than 1%, as recommended by the WHO.
Experimentation using accurate real-world data show how different dietary preferences, such as vegan, pescatarian, and Mediterranean diets, affect cost and nutritional quality of generated meal plans. Trade-offs in practical diet planning between dietary restrictions, cultural preferences, and financial feasibility are deduce

Keywords

Status: accepted

Back to the list of papers