How to design highly informative and cost-efficient industrial experiments

Quality has to be built into products and processes during the design stage of their development. It is therefore crucial that any decision made in this early stage of development is based on accurate information. The information is typically obtained from a series of experimental tests, in which the effect of one or more factors on the quality of the product or process is investigated. The settings used for each of the experimental factors at every test run are crucial for the reliability of the information provided by the experiment. In the paper, it is shown how to compute the best factor settings for an industrial experiment. The split-plot type of experiment considered in the paper not only has the advantage that it is easy to carry out, but it has excellent statistical properties as well.status: publishe