For optimal use, please visit numiqo on your desktop PC!

Design of Experiments (DoE)

Screening design

You have many possible factors and little knowledge.

Full factorial design

You have only a few factors and want the full picture.

Optimization design

You already know the important factors and want to optimize.

Robust design

You already know the important factors and want stable settings that are less sensitive to noise.

Mixture design

Your factors are ingredients or proportions.

Optimal design

You have constraints or no standard design fits your situation.

Factors

No. of Factors
No. of Replicates
No. of Blocks
No. of Center Points
Factor Low High

Plan


Run Order Factor A Factor B
1-1-1
2-11
31-1
411

How-to

Custom Optimal Design Calculator

Here you can create a custom optimal design online when a standard factorial, screening, response surface, or mixture design does not fit your experiment. Define the candidate factor levels, choose the model and number of runs, select an optimality criterion, and exclude combinations that cannot be tested.

Custom Optimal Design

A custom optimal design is a design of experiments (DoE) method that searches a candidate set and selects an efficient subset of runs. It is useful for practical experiments with unusual factor levels, mixed numeric and categorical factors, hard constraints, or a fixed run budget.

Available Optimality Criteria

  • D-optimal: selects runs that are efficient for estimating the model coefficients.
  • I-optimal: selects runs that minimize the average prediction variance over the candidate set.

When to Use a Custom Optimal Design

  • A standard design does not match the available factor levels.
  • Some factor combinations are impossible, unsafe, or not useful.
  • You need a specific number of experimental runs.
  • You want to compare D-optimal and I-optimal designs for the same candidate set.

How to Create a Custom Optimal Design

  1. Select Custom Optimal Design in the DoE calculator.
  2. Add numeric or categorical factors and enter their available levels.
  3. Select the model: main effects, interactions, quadratic terms, or full quadratic.
  4. Choose D-optimal or I-optimal as the criterion.
  5. Add forbidden combinations if some candidate runs must be excluded.
  6. Generate the design and export the resulting test plan to Excel.

For more specific guidance, see the D-optimal design calculator or the I-optimal design calculator.

Cite numiqo: numiqo Team (2026). numiqo: Online Statistics Calculator. numiqo e.U. Graz, Austria. URL https://numiqo.com