Page 283 - Six Sigma Advanced Tools for Black Belts and Master Black Belts

OTE/SPH
OTE/SPH
August 31, 2006
JWBK119-18
Taguchi Methods
268 3:6 Char Count= 0
18.1 INTRODUCTION
A little over a decade ago, Genichi Taguchi, a Japanese engineer, spent some time
at AT&T Bell Laboratories in New Jersey, and demonstrated how statistical design
of experiments could enhance the efficiency of empirical investigations in industrial
research and development, rationalize the product realization process, and improve
the quality and reliability of manufactured products. Before long, a collection of ideas
and techniques under the label ‘Taguchi methods’ began to spread to other organi-
zations, and began to be hailed by their advocates as the long-awaited cure for the
faltering quality performance and declining competitiveness of American manufac-
turing industries. However, the Taguchi movement was soon met with criticisms and
challenges from many veterans in the quality field who, led by academics and ap-
plied statisticians, subjected the various Taguchi procedures to close scrutiny; not a
few of them have warned against ‘quick and dirty’ solutions to quality problems and
advised those seeking valid answers to depend more on methodologies built upon
firmer theoretical grounds.
The debate rages on today, although the different schools of thought have al-
ready been well aired in seminars, conferences, journals and books. 1−20 Although
descriptions and commentaries had initially appeared in a variety of non-academic
8
publications, 21−24 the paper by Kacker and the accompanying discussion marked the
first formal attempt to introduce Taguchi methods in a major journal. After several
years of debate in the USA, Box gave some comprehensive expositions of applied
statisticians’ views. 14,15 More recently, Pignatiello and Ramberg 19 summarized the
pros and cons of Taguchi methods for those who are familiar with the details, and
Nair and his panel of discussants 20 went deeper into the academic aspects of ‘pa-
rameter design’, the main pillar of Taguchi methods. Studies of Taguchi methods and
related matters can be expected to continue for years to come.
Management and technical decision makers in industry, who may only have a
cursory exposure to the subject but little inclination for procedural details or academic
debates, will have a need at this juncture to gain some essential insights into the issues
involved, so as to be able to make more informed judgments on the viewpoints and
techniques that they face from time to time. To this end this chapter presents, in a non-
mathematical language, an overview and analysis of Taguchi methods from several
newangles,encompassingtechnical,cultural,andpedagogicalconsiderations:itisnot
an attempt to arrive at definitive conclusions, but is meant to add further perspectives
to this important and practical subject.

18.2 GENERAL APPROACHES TO QUALITY

There are several dimensions to the quality improvement initiatives of an organiza-
tion. The most fundamental is a properly established quality management system,
which encompasses subjects ranging from corporate philosophy to policies, proce-
dures, employee motivation, and even supplier management and customer relations,
and can be typified by the ISO 9000 series descriptions. This should be complemented
by the requisite quality technology, entailing engineering capabilities and resources
commensurate with the technical performance required of the products or services