Who should use design of experiments? An interview with Douglas Montgomery
Apr 2, 2018 1:00 AM
Douglas Montgomery is an expert in design of experimentsDoug Montgomery, Regents’ Professor of Industrial Engineering and Statistics at Arizona State University, is an innovator in the field of experimental design. He has consulted with more than 200 leading businesses. On April 24, he will speak at the JMP Explorers seminar in Eindhoven, Netherlands. The event will be livestreamed. I asked him to tell me about applications of design of experiments and his thoughts on the field.
Who should use design of experiments?
I think that the potential applications of DOX [design of experiments, also abbreviated as DOE] are limited only by the imagination of the engineer, scientist or analyst.
I have seen this methodology used with great success in virtually all fields of engineering and science. I have also seen applications in health care, financial services, marketing, and many other business and transactional settings.
What is your favorite design and why?
I can’t really identify a “favorite” in the traditional sense. I have always been a big user of two-level factorial designs and the fractional versions of these designs because they are small and efficient, and they fit many types of application environments.
In recent years, I have conducted research on variations of these designs, space-filling designs for deterministic computer experiments, response surface designs, and mixture designs. I have come to appreciate the power of optimal designs as software has become available that implements these techniques in an appropriate way.
Definitive Screening Designs
What would you describe as an important recent development or improvement in design of experiments?
I think definitive screening designs are a very important recent development, perhaps the most important development, in the last decade. The ease with which optimal design methodology can be used by practitioners is another.
You have consulted with many businesses. Can you share what impressed you most? Or what was your biggest challenge?
What has always impressed me is how excited engineers and scientists become when that take ownership of the DOX toolkit. These are very smart people, and they always seem to be able to find applications that lead to great results.
I’ve had several situations where the problems were challenging. It’s hard to say which ones were the biggest. But certainly some of the applications with mixture experiments were very complex, as were applications in semiconductor manufacturing and large-scale chemical processes.
I currently have a research project sponsored by the Department of Defense where we are trying to solve some of the experimental design problems in operational testing of military systems. These can be very complex experiments with many factors, including some that are difficult to control or measure, have constraints on resources and nontraditional responses, among many other things. There are seven universities involved in this research consortium, and we are all working on a wide range of problems that should have high payback for the operational testing community.
What can we expect from the Explorers seminar in April in Eindhoven? What will we learn?
I think attendees will get a solid perspective on design basics. They will also get exposure to some new developments, such as definitive screening designs and optimal designs and their application.
If you would like to see Douglas Montgomery live, join us on April 24 at the seminar in Eindhoven, Netherlands, or join the event online.