Free ENBIS Webinar by Ursula Garczarek: "Without Statistical DoE, 30% of Experimental Runs are Wasted"

27 September 2016; 12:30 – 13:30; Webinar

Ursula Garczarek will talk about Design of Computer Experiments in comparison to One-Factor-At-a-Time approaches. The webinar will be moderated by Winfried Theis.

This webinar has been postponed to late September, because of too few attendees.

Statisticians claim that applying statistical Design of Experiments (DoE) is better than its next best counterpart, the so-called One-Factor-At-a-Time (OFAT) approach. The prominent example consists of a quadratic function describing the outcome, with its maximum somewhere inside the spanned range of the two design factors. OFAT misses the optimum, DoE comes close. Proof done!? So why remains OFAT so popular, and why are those companies where research is still done without proper DoE not all going bankrupt?

Getting DoE in a company implemented costs efforts and it requires investments in statistical consultancy, training of employees and software. To justify those efforts, one convincing example is not good enough, a business case is needed that quantifies the cost reduction by DoE. In my presentation I will show the design and result of a simulation that I ran to deliver evidence to the rather bold statement: Without statistical DoE, 30% of experimental runs are wasted. Running the simulation with a clear goal in mind, I also learned a lot about those scenarios, where OFAT is actually not so bad.