To model cost against quality, quality can be measured
by the percentage of coverage and by the number of com-
plaints. By using the percentage of coverage to measure
quality, the costs of prevention and appraisal increase
when the quality increases as seen in Fig. 2. Costs of failure
are inversely proportional to costs of prevention and
appraisal, which decrease when quality increases. Further-
more, the company has spent around HK$10,000 to pro-
vide better vegetation coverage of hydro-seeding projects
for their clients, which shows significant quality
improvement.
With regard to quality measurement using the number
of complaints, Fig. 3 shows fewer complaints due to the
increase in cost quality for the project. The situation is sim-
ilar to the case obtained using the percentage of site cover-
age as shown in Fig. 2. When the costs of prevention and
appraisal decrease, the costs of failure increase and vice
versa. The costs of failure are also relatively lower than
the costs of prevention and appraisal.
From Figs. 2 and 3,it isdi?cultto show any break-even
point. Since the costs of prevention and appraisal are much
higher than the costs of failure, it suggests that this com-
pany may already use enough investments to prevent fail-
ures and to further improve quality.
As it is clear that the PAF modelling method cannot
reveal the correlation between the costs of prevention,
appraisal and failure, it is necessary to individually study
each of these costs by using other more e?ective methods.
One of them is the use of the hyperbolic and Gaussian dis- tributions employing the 6r concept which was reported by
Tam and Le [27]. In this paper, the Vandermonde interpo-
lation techniqueisproposed which cane?ectively showfine
details in the data, hence the location of the break-even or
optimal point. The main motivation of using this method is
because the Vandermonde interpolation technique has
been successfully used in the field of signal and image pro-
cessing to best-fit a given data set.