Design of Experiments
In the 1930s, statistical design of experiments was developed as a method for optimizing agricultural processes. Based on this, DoE - derived from the English term for statistical experimental design "Design of Experiments" - has developed into a standard procedure for process and product optimization in almost all industrial sectors. This applies to pharmaceutical applications, developments in the automotive sector and also especially to the semiconductor industry. Furthermore, the design of experiments is nowadays used in the planning of psychological experiments or also in the optimization of marketing campaigns.
The core idea of experimental design is to perform exactly the right experiments to solve a particular problem by concretely planning experiments. That is, statistical design ensures that the right experiments are performed - not too many and not too few.
We support the design of experiments by taking care of the systematic without sticking too much to a standardized procedure. The statistics adapt to the respective circumstances as well as possible, your question will not be adapted to the statistical methodology. We take care that the analyzability of your data remains optimal.
Of course, we assure ourselves of the qualities of the various measured values, which are used within the framework of a test program. The evaluation of the data can be carried out by us. The usual graphics in the field of experimental design, such as Pareto diagrams, main effect and interaction plots, are produced as a matter of course, as are variance and regression analyses. In summary, we prepare a report which we discuss in detail with your experts.
We have access to practically all software products available on the market for evaluating the data. We can therefore always use exactly the software that is also used in your company. In this way, we ensure the greatest possible learning effect for your employees.
Design of Experiments: 4 steps to success
Classical design of experiments
The classical design of experiments still represents the most frequently used variant of Design of Experiments today. The classical design of experiments is characterized by the fact that the experimental plans used are clearly structured and easy to understand. For a large part of all applications, the classical plans are the suitable solution.
Classical design of experiments is a cornerstone of more general paradigms, such as Six-Sigma or Quality-by-Design.
Optimal design of experiments
In certain areas, however, the classical plans turn out to be too inflexible. The school of optimal experimental design therefore follows the principle not to adapt the problem to the experimental design, but to adapt the experimental design to the problem. For this purpose, statistical optimality criteria (usually D-optimality and I-optimality) - some of them complex - are used to create experimental designs with a higher flexibility.Mixtures and formulations
Mixing experiments are a special application. These are applications in which each individual experiment consists of combining different components into a mixture. This type of experimental setup has certain implications both for the setup of a DoE and for the statistical evaluation of the resulting data.
Therefore, special variants of classical and optimal experimental designs were developed to address exactly this issue.
Split-Plot-Designs
Another important special field of experimental design are so-called split-plot designs. These are problems with factors that are difficult to adjust. In a classical or optimal experimental design it can happen that the experimenter often has to investigate a certain influencing factor in different settings. For example, in the case of temperatures, it is often difficult to carry out the many independent experiments of a DoE at different temperatures.
Split-plot designs are one answer for this common challenge.