|
In recent years,the development and applications of Taguchi''s e-
xperimental design method have been considered as a powerful to-
ol to enhance the products'' quality. However ,in the theoretical
aspect,Taguchi''s method is still contradictory. One of the exam-
ples is the assumption of quadratic loss function.Taguchi''s loss
function takes only the first and the second moment,i.e.,the ex-
pected value and variance, of the population distribution in its
expression, ignoring the higher-order moments from the Taylor''s
expansion. According to statistical theory, the third and fourth
moments also play important role to describe the population dis-
tribution.
This paper examines the relationship between Taguchi''s loss fun-
ction and population distribution by using simulation method. T-
he results show that for the large-the-better quality character-
istic, the value of loss function are affected by the higher-or-
der moments of the distribution; but for the smaller-the-better
and nominal-the-best quality characteristics, the loss functions
are only influenced by the expected value and variance of the d-
istribution.
|