Statistical design of experiments in industrial practice (Q1893511)

In chemical industry, statistical design is more and more widely used for the examination of industrial and biological processes and products. This way, the experimenter is supported in trying to develop products with optimal properties and reliable processes to produce such products with minimal effort. This paper demonstrates the practice of statistical experimental design in one big chemical and phamaceutical company, Ciba- Geigy in Basel, Switzerland. All the presented examples arose during consulting activities of the Mathematical Applications Group in Ciba. The plan of the paper is as follows. Section 2 will give an introduction into experimental design and general strategies for design choice. In Section 3 experimental design in both industrial and biological applications at Ciba will be presented. Finally, In Section 4 the designs used in the two fields will be compared. In the companion papers, \textit{M. Abt} and \textit{F. Pukelsheim} [ibid., 27- 33 (1995; see the following entry, Zbl 0818.62088)] and \textit{M. Abt} et al. [ibid., 35-47 (1995; see the entry below, Zbl 0818.62089)] demonstrate the theory and application of Taguchi methods in detail. In Section 3.3 of this paper the corresponding designs are briefly discussed, too.