00135 �0b8328312d99bc5a1f6952b3660550

136

Simpson & Keats

Table 6. Example 1 Using Smali and Large Process Shifts

Responses	Transfomn?	R^2	X	l I <f i i UJ |	Cj b Y A
Cost
Smali Shift		0.91	+	+	+
Large Shift		0.90	+	+	+
Reference Value
Smali Shift		0.88			+
Large Shift		0.80			+
Decision lnterval
Smali Shift		0.67			+
Large Shift		0.60			-
Sampling lnterval
Smali Shift	loge	0.64	-	-	+
Large Shift	log e	0.76	-		+ +
Sample Size
Smali Shift	loge	0.67		-	-
Large Shift		i	insufficient variability -		all but 8 runs had n = 1 as

optimum sample size

for the examples in the smali shift scenario, A was significant in each model that was developed for the large shift scenario. These results are not surprising, but they emphasize the importance of correctly specifying the size of the process shift.

Yerification Although the results for the two control chart variables (k and H), the sample size, and the sampling interval are interesting and can be beneficial in the decision making process, often the primary focus in economic design is the overall process cost. We decided to develop a third example, different from the other two, to test the results of the sensitivity analysis. A