History of Design of Experiments
A Brief History of Design of Experiments – DOE.
[leadplayer_vid id=”52594A4A370DB”]
Please Subscribe to receive further Tutorials.
There’s lots more to come.
SLIDE: The Scientific Method and the History of Experimentation (Aristotle, OFAT, Fisher, Box and Taguchi)
Let’s talk a little about the history of DOE. We like to start from many years ago, from 300 BC, and we can attribute Aristotle, the Greek philosopher, as the Father of Scientific Method, because he wrote about it. So he started the scientific method, basically it was…you put a hypothesis to it, and you test it, you validate your hypothesis or invalidate it, or so forth.
A lot of folks did that through what was called OFAT, One Factor At a Time, and basically you could constrain all the variables except one. You tweak one variable and see what the response is on your system or process, after you change that parameter. Of course, that was very popular and we were well taught in schools how to do that in our Science classes, and it prevailed until the early nineteen hundreds.
One joke that I have for Edison, one of my favorite inventor, is that his famous quote is, “one percent inspiration and ninety-nine percent perspiration.” My comment to that would be, if we had known these other experiments, he would have come with a fifty percent inspiration and fifty percent perspiration. So, basically, better (ROI) for his labor. Of, course, that’s a joke…but anyways, we move on to Sir Ronald Fisher, who is officially, probably, the person who actually started all this design of experiments craze. He was a…English, biostatistician and lived during the World War one. Living in UK he had to come up with a method, a better development… experimentation method for agriculture. His goal was to increase the yield of the crop to feed the soldiers and the citizens, and so forth, but as you can imagine his challenge was getting data. I mean, if you plant a crop, maybe in the spring, you get results by fall, so that’ll be at least six to nine months, right, he needed a better way to plan experiments and get data. And, that’s a common challenge that we have in the biologics or biotechnology cause all our campaigns may take up to many months to get some real data. That’s why he came up with these (DOE and Design of Experiments concepts). Now, you probably know him through ANOVA, which is fairly prevalent in our experiments analysis variance. If you use jump or even Excel or other…statistic techniques you’ll see this often, F-Ratio, that is, the F is for Fisher…Ratio. We’ll talk about these concepts in the ensuing modules.
Then we come to George Box, famous person for RSM, Response Surface Methodology, they are related to each other because Sir Ron Fisher’s daughter, Joan Fisher, married George Box who is also from England or UK, and so he’s the son-in-law. He is famous for, you probably know him through Box banking designs, if you’re familiar with the DOE world a little bit, (Box Cox transformations). He went to the University of Wisconsin, as a faculty member to teach. He recently passed away in 2013, March, I believe. So, we’re losing all our famous people in DOE. And, then we go on to Taguchi in the same area. He’s a Japanese statistician, more of an industry person, came with the Loss Functions. His contribution was in the robust design, (aftermath of the world war two) Japan was trying to develop better quality products, so he has done a lot of contribution in that area of helping the industry use statistical methods, and so forth. Some of his techniques may be controversial in the eyes of statisticians, but, nevertheless very practical advices he gave to the industry and the field.
Please Subscribe to receive further Tutorials.
There’s lots more to come.