ABSTRACT
Before the 1980s, American engineers had little knowledge on how to use statistical methods to improve processes. They were using the best guess and one-factor-at-a-time methods. Even though a very large number of statistical methods were available, these methods were complicated. Statisticians could not cross the bridge in finding applications of their sophisticated theoretical results to engineering. And many of the methods such as factorial designs and optimization, among others, were created in the West. Professor Taguchi was able to modify some of the existing statistical tools and made them easy enough to understand so that engineers could apply them. Even though there are a large number of statisticians who tend to disagree with most of Taguchi's contributions, there are many engineers who have reported great successes by using his modified statistical methods. To name one of his success stories, Bell Telephone Laboratories in New Jersey, USA, and another electrical communications laboratory based in Japan, where Taguchi held a position, were developing similar cross-bar telephone exchange systems. In addition to facing Bell's greater resources, the Japanese were constrained by inferior materials. Despite all the adversity, the new cross-bar system from Japan was rated superior and cost much less to produce. The effect was so dramatic that Western Electric stopped production and began importing systems from Japan. The emphasis of Dr. Taguchi's methodology is on functional variation, which can be measured in terms of product performance such as strength, pressure, shrinkage, response time, taste, and mean time to failure, among many others. Viewed as the enemy of the producer and its customer, functional variation can relate to the performance of the end product or to the process that manufactures the end result. The purpose of experimentation using Taguchi methodology is to identify those key factors that have the greatest contribution to variation and to ascertain those settings or values that result in the least variability. In developing methods to better understand the influences upon the functionality of products and associated processes, Professor Taguchi has been particularly recognized for three major contributions to the field of quality. 1. Loss function 2. Orthogonal arrays 3. Robustness
In the following, we will discuss Taguchi's contributions to the design of experiments field and illustrate the weaknesses of some of his methods.
