Application of design of experiment methods from the industrial environment to crop science and medicine
By Jens Harbers
In all scientific and engineering disciplines, the design of experiment is used to create experimental plans. Particularly in industrial experimentation, methods are available that are hardly or not at all used in agriculture. For example, the surface response method in agricultural sciences is only used in mechanical engineering and the adjustment of agricultural machinery.
The work project therefore tries to combine ideas from both fields in order to improve experiments in crop science so that more accurate results can be obtained with fewer plots.
For example, methods such as fractional factorial designs and Plackett-Burman types are a possible extension when it comes to crop production systems. If there are already literature values for questions, optimal designs (D, I, A, G, T, E optimal) are also a possibility to achieve more precise results with fewer plots.
The challenge is to transfer the experimental designs from the industry into a plot form, as it is used in agricultural sciences and medicine. Especially the apsect of spatial influence has to be taken into account, as this phenomenon does not appear in this form in industrial trials, but is omnipresent in crop science due to natural inhomogeneous soil conditions.
Attachment: html_doe_industry.pdf (424 KB)