Diet optimization allows us to navigate the complex trade-offs between nutritional adequacy, environmental impact, and cultural relevance. To support this, Optimeal Open Access is now available for free.
Below, you’ll find answers to common questions about using Optimeal Open Access. Please reac
Access to Optimeal Open Access requires an account and a license. To request access, please fill in the form and briefly describe how you plan to use Optimeal Open Access (e.g. academic research, consultancy work, or internal analysis), along with a short overview of the research or project context. You will then receive a license agreement to review and sign. Once the agreement is in place, you will be sent your account details and instructions to activate access, including setting up two-factor authentication (2FA).
Yes. Optimeal Open Access is free to use, with access granted under a license agreement. The license ensures appropriate use of the tool, proper attribution in publications, and responsible application in both academic and industry contexts. Users remain responsible for the quality, interpretation, and verification of results generated with the tool.
Optimeal Open Access is a web-based interface that allows users to upload data and generate results. It provides the diet optimization engine only.
Prepare your data. The most important step is preparing your data correctly. Start by downloading the empty template Excel file. The first worksheet contains detailed instructions and should be reviewed carefully before entering any data.
Upload the file via the web interface
Configure the optimization:
Select the objective function
Minimize or maximize a diet-level property, or
Minimize deviation from a reference diet
Choose linear or quadratic optimization
Run the model
Export results to Excel. All outputs are provided as structured tables for further analysis.
When minimizing deviation from a reference diet, Optimeal Open Access uses an objective function that minimizes absolute changes in food quantities (e.g. grams per person per day), rather than relative changes.
This choice is motivated by both numerical and conceptual considerations. From a numerical perspective, relative deviation requires dividing by baseline intake, which can lead to instability when foods are consumed in very small amounts or not consumed at all. Using absolute deviation avoids these issues and ensures robust behavior across datasets.
From a modeling perspective, relative deviation disproportionately penalizes foods with low baseline consumption. This makes it difficult for new or currently marginal foods to enter the optimized diet, and tends to reinforce the status quo by concentrating changes in foods already consumed in large quantities. Because many Optimeal applications are forward-looking and exploratory, absolute deviation provides a more neutral playing field in which alternative products can realistically replace existing ones.
When interpreting results, it is important to keep in mind that:
Deviations are minimized in absolute, not relative, terms
Foods consumed in larger quantities may absorb more of the adjustment
Minimum and maximum constraints at the food level play a key role in shaping realistic outcomes
This objective function supports transparent, stable, and interpretable diet optimization, but results should always be evaluated in the context of the chosen objective, constraints, and input data.
Optimeal® is built on mathematical optimization, using either linear or quadratic programming. In both cases, the model identifies the optimal combination of foods that satisfies all defined constraints while minimizing deviation from a reference diet. The key difference lies in how deviation from the reference diet is penalized.
When using linear programming, deviation from the reference diet is minimized as the sum of absolute changes in food quantities:
where xi is the optimized intake of food i and xi* is the reference intake.
Linear deviation is useful when the goal is to identify where the model finds the most leverage to meet constraints.
When using quadratic programming, deviation from the reference diet is minimized as the sum of squared absolute changes:
where xi is the optimized intake of food i and xi* is the reference intake.
Quadratic deviation is often preferred when exploring realistic dietary transitions rather than sharp adjustments.
Linear optimization is used when minimizing or maximizing a single diet-level property.
Optimeal Open Access produces one optimal solution per run, based on the selected objective function and the constraints defined in the template.
No. In its current form, Optimeal Open Access supports single-objective optimization only. If multi-objective optimization is of interest for your work, please reach out to us to discuss potential options or collaboration.
No. Optimeal Open Access uses continuous optimization, meaning food quantities can take any real (non-integer) values within the specified constraints. Integer or mixed-integer optimization, where foods are restricted to discrete units (e.g. servings) or on/off choices, is not supported.
Continuous optimization is well suited for diet modeling, as it allows smooth adjustments in food quantities and supports realistic exploration of dietary patterns. If integer constraints are important for your application, please contact us to discuss potential alternatives or custom solutions.
Diet optimization is data-intensive. At minimum, the template requires:
Decision variables: A list of food items or food groups included in the diet
Diet composition or reference diet*: Quantities per food item (e.g. current consumption)
Food-level properties: These can include nutrients, environmental indicators, price, or any other quantified attribute linked to the food list
Constraints: Minimum and/or maximum values at food, food group, nutrient, cost, or other diet-property levels
Although we focus on diets, the template is very flexible that you can focus on meals or composite foods, for which a list of foods or ingredients are listed and ingredient-level properties would be needed.
No. Optimeal Open Access does not explicitly model interdependencies between food products or nutrients. Each food contributes independently to diet-level properties based on the values provided in the input data, and nutrients are treated as additive across foods. This means that interactions such as nutrient bioavailability, synergistic or antagonistic effects between nutrients, or behavioral substitution effects between foods are not explicitly represented in the model.
If modeling nutrient interactions or more complex dietary dynamics is important for your application, these aspects would need to be addressed outside the optimization framework or explored through complementary analyses.
No, Optimeal Open Access does not include any data. Users must supply all required data via the template. This design allows full flexibility, but also means data quality and consistency are the user’s responsibility.
If you are looking for support with data sourcing, processing, or integration, you can contact us to explore collaboration or consultancy options. In addition, open-access datasets created by Mérieux NutriSciences | Blonk are available for certain contexts—for example, Dutch environmental impact data of food via the Dutch National Institute for Public Health and the Environment (RIVM), or European environmental impact data of food developed for the European Food Safety Authority.
Diet optimization searches for solutions that meet all constraints simultaneously. If constraints are too restrictive or conflicting, the model may not find a feasible solution.
In that case, you will get an export with no optimized solutions.
Typical causes include:
Very tight minimum and maximum constraints
Incompatible targets across multiple diet properties
Insufficient flexibility in the food list
Relaxing constraints or reviewing assumptions usually resolves the issue.
Optimeal Open Access can technically process individual-level diets, but each diet requires its own completed template. This makes large-scale or bulk processing inefficient.
At present, Optimeal Open Access does not support batch uploads or automated bulk runs. If bulk processing or large-scale applications are important for your work, please contact us to discuss possible solutions or collaboration.
No. Optimeal Open Access is currently accessed through a web interface and does not offer an API for automated or programmatic use. The tool is designed for template-based uploads and Excel-based result exports. If API access or automated workflows are important for your use case, please contact us to discuss potential options or collaboration.
Yes, subject to the license terms. Optimeal Open Access has a strong track record in peer-reviewed research and can be used for academic publications. Proper citation and acknowledgement are required and outlined in the license agreement.
Yes, subject to the license terms. Optimeal Open Access is suitable for consultancy and applied industry work, particularly where transparency, reproducibility, and control over assumptions are critical.
Optimeal Open Access exports results as an Excel file. The export includes:
The original input dataset
A list of food products with quantities for both the reference diet and the optimized diet
A list of diet-level properties with values for the reference diet and the optimized diet
Because results are delivered in Excel, users can easily perform additional analysis, summarize outcomes, and create their own tables or visualizations using standard Excel functions or external tools.
Optimeal Open Access is provided as an expert tool and does not include built-in user support or guided workflows. If you encounter a technical issue, please contact info@blonksustainability.nl. You are also welcome to reach out if you have questions that are not addressed in this FAQ. For methodological support, data preparation, or project collaboration, please contact us to discuss available options.
Optimeal Open Access may not be suitable if you are:
Looking for interactive dashboards or visualizations
New to diet optimization and data preparation
Seeking a plug-and-play tool with built-in datasets
In those cases, Optimeal Consultancy may be more appropriate. Please contact us, specifying “Optimeal consultancy” in the subject.
Optimeal® Open Access (Mérieux NutriSciences | Blonk, Rotterdam)
No. If this is of interest to you, please contact us to discuss possibilities.
No. Optimeal is made for individual runs. Each run requires you to upload your full data that is saved on your computer. User data and solutions are not retained.