Design of Experiments, or DOE, has been in practice since the early 1900's. Developed in England, R.A. Fisher led the initial applications in the field of agriculture. Other contributors to Designed Experiments have included Rao, Plackett, Burman, Box, Taguchi, Barker, Derringer and many others. Until the early 1980's, Designed Experiments were usually set up and run by specialists within an organization. Today, with the advent of readily available software, such as DOE Wisdom, the non-statistical person can successfully set up and analyze simple but powerful experiments.
A designed experiment is a well-controlled family of tests. Each test is run one or more times. For each controlled test, outputs, also known as responses, are measured. From test to test controlled changes will be made. Once the tests are completed, the data can then be analyzed. As you will see with further understanding of designed experiments, simple graphs can be of great value in furthering your understanding of the results of the experiment. As the data is analyzed, conclusions will be reached as to the the best setting for the inputs, or factors. The final step in a designed experiment is to actually try these settings and see if they produce the predicted result based on the analysis. If the results are close to the prediction, and better than any baseline we might have, the experiment will be declared a success.
The challenge
facing corporations is to cut design and development time while
producing low-cost quality products that are ready to perform. Many
organizations are being challenged to cut delivery time in half
or more.
Demanding that
people work harder is not the solution. Providing proper tools aids
people in working more efficiently and effectively. Experimental
Design is a premier tool in helping meet these challenges. By
understanding and applying Experimental Design techniques,
scientists, researchers, and engineers typically obtain a 50% reduction
in the time required to conduct tests. This translates into enhanced
understanding of technologies, reduced design and development time,
and decreased costs. If you are looking for a little more background on design of experiments, we have an example DOE, "creating the perfect lego" that can be followed here: