Page 304 - Six Sigma Advanced Tools for Black Belts and Master Black Belts
P. 304

OTE/SPH
 OTE/SPH
          August 31, 2006
                              Char Count= 0
 JWBK119-18
                         3:6
                                  Training and Learning                      289
           Performance














                                                                       Time
               Screening  Breakaway  Marginal  Characterization  Optimization  Tracking
               procedures  moves    analysis  studies     applications  routines


                    Information analysis  ‘Parameter  Mathematical modeling
                                    design’
                  Figure 18.9 Choice of statistical tools along the learning curve.


      methods, as already pointed out, offer a prescribed set of techniques, whereas statisti-
      cians represented by Box encourage `a la carte selections used according to the progress
      of investigation: screening, characterization or optimization. Thus the actual design
      of an experiment is shaped by the existing knowledge about the subject of study, and
      the size of experiment is defined according to the expectation on the quality of the
      resulting information (e.g. level of resolution of parameter effects). In this respect the
      Box approach is much more versatile and flexible, as it offers opportunities for quick
      screening of a large number of parameters, followed by detailed mathematical model-
      ing of the critical ones that remain or, in some cases, a quick breakaway from current
      ranges of parameter values for better results, culminating in a final fine-tuning of
      parameter settings for optimized performance. Indeed on-line optimality seeking (by
      response surface methodology) and tracking (by evolutionary operation) are strate-
      gies that are sorely absent from Taguchi methods; no statistical quality engineering
      training or application would be complete without their inclusion. Figure 18.9 shows
      the sequence of major tools to be applied as performance progresses along the learning
      curve.


      18.7.3 Directions
      In discussing the technicalities of statistical quality engineering with management
      and operations personnel, measures should be taken to ensure that they do not fail
      to see the forest for the trees. Hence, it would be useful to explain the historical
      background of the various schools of thought that exist in the profession today. An
      awareness should be brought about of the richness of experimental design literature,
      scope of applications, and practical conditions to be encountered in practice, 60,61  and
   299   300   301   302   303   304   305   306   307   308   309