Complete Volume - Institute of Business Management
Complete Volume - Institute of Business Management
Complete Volume - Institute of Business Management
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Process Capability Analysis for Non Normal Data<br />
Research<br />
5. Simulation Results and Newly Proposed PCI<br />
The results <strong>of</strong> this section are based on a simulation study that<br />
we conducted to determine the effectiveness <strong>of</strong> our proposed<br />
measure <strong>of</strong> process capability. Details <strong>of</strong> the simulation procedure<br />
is listed below.<br />
1. Fifteen samples each <strong>of</strong> size 30 were drawn randomly<br />
from a Weibull distribution with different scale and<br />
constant location parameters.<br />
2. Regression model was fitted to equation (2) and<br />
estimators and standard errors for ln ( α ) and<br />
( 1 β ) were obtained.<br />
3. Variance <strong>of</strong> ln ( α ) and ( β )<br />
fitting least square regression model.<br />
1 were obtained while<br />
4. Delta method was used to determine variances <strong>of</strong><br />
α and β . Delta methods says that if g( γ )<br />
Var<br />
( γ ) [ g′<br />
2<br />
( γ )] [ Var( γ )]<br />
= .<br />
Cˆ<br />
p<br />
f = then<br />
5. Process capabilities ( USL − LSL) / 6σˆ<br />
estimated, replacing σˆ<br />
in step 4.<br />
= were<br />
with standard errors obtained<br />
6. Averages and standard errors <strong>of</strong> process capability<br />
indices are reported in Table 1.<br />
240<br />
PAKISTAN BUSINESS REVIEW JULY 2010
Change language