Page 321 - Six Sigma Advanced Tools for Black Belts and Master Black Belts

OTE/SPH
OTE/SPH
August 31, 2006
Char Count= 0
JWBK119-19
306 3:6 Economical Experimentation via ‘Lean Design’
very last observation from his experiment before he is in a position to announce any
findings at all from his work. Situations abound in industry where one may not have
the luxury of awaiting the full data set, owing to foreseen or unforeseen circumstances
similar to those highlighted in the previous section. In many cases, rapid diagnosis
of a troubled process is the best way to retain customer confidence, and in other
settings, early availability of indicative results is vital to competitiveness in research
and development: the time taken to generate full experimental data sets may simply
be unacceptable. Lean design can provide a way out, as long as the user is fully aware
of the possible pitfalls associated with the no-interaction assumption typically seen
in a Taguchi-type design -- and, better still, aided by prior knowledge concerning
the relative strengths of the factors under investigation. With this proviso, there are
several advantages that would make lean design attractive to both management and
technical personnel:

1. A lean design can be specified at the outset to suit given time and resource budgets.
2. A lean design takes maximum advantage of any prior technical knowledge about
the significance or otherwise of main and interaction effects, saving investigation
costs through reduction of experimental runs.
3. The use of lean design and incremental experimentation fits the Taguchi school
of thought regarding the handling of interaction effects, yet makes use of main-
stream statisticians’ favored approach of acquiring and updating information via
sequential data gathering.
4. Although quick results from lean experiments could be less reliable than results
associated with regular orthogonal designs, they can be subsequently improved
upon -- in fact, self-correct -- by incremental experimentation, to a point that is
identical to the results of regular designs. This, indeed, would be the clinching
argument for using lean design as the first stage in an experimental project.

REFERENCES

1. Kacker, R.N. (1985) Off-line quality control, parameter design, and the Taguchi method.
Journal of Quality Technology, 17, 176--188.
2. Goh, T.N. (1993) Taguchi methods: some technical, cultural and pedagogical perspectives.
Quality and Reliability Engineering International, 9, 185--202. See also Chapter 18, this volume.
3. Box, G.E.P., Bisgaard, S., and Fung, C.A. (1988) An explanation and critique of Taguchi’s
contributions to quality engineering. Quality and Reliability Engineering International, 4, 123--
131.
4. Pignatiello, J.J. and Ramberg, J.S. (1991) Top ten triumphs and tragedies of Genichi Taguchi.
Quality Engineering, 4, 211--225.
5. Goh, T.N. (1994) Taguchi methods in practice: an analysis of Goh’s paradox. Quality and
Reliability Engineering International, 10, 417--421.
6. Box, G.E.P., Hunter, W.G. and Hunter, J.S. (1978) Statistics for Experimenters. New York: John
Wiley & Sons, Inc.