00149 X58e44d67a8fa64ce25678ae9975d0f

150

Simpson & Keats

Table 12. Shewhart and CUSUM Control Chart Sensitivity Analysis for the Smali Shift Case (0.25-1.25)

Response	R^2			Input Variabłes
Expected Cost per Time	X	cusu M	X	o o O 1U	a	b	Y	5
LV	0.94	0.91	X c	X C X C				X c
Montgomery	0.93	0.91	X c	X C				X c
CUSUM Reference Value
LV	-	0.88						c
Montgomery	-	0.99						c
Decision lnterval
LV	0.85	0.67					c	X c
Montgomery	0.92	0.64			c	C	X c	X c
Sampling lnterval
LV	0.81	0.64	c	C X c		X c		X
Montgomery	0.76	0.58		C X c	X c	X c	c	X c
Sample Size
LV	0.87	0.67	X	X C X			X	X c
Montgomery	0.90	0.91		X C X c			X	X c

of variability explained by each model. For the expected cost per time unit response, the significant parameters are identical for the two control charts. The assignable cause ratę (X), the costs while producing in and out-of-control (Co and Ci), and the size of the process shift (A), are the most important inputs affecting cost. Comparison across examples reveals that, although the Co term is not significant in the Montgomery example, it should be considered important in estimation. Therefore, only four of the 12 input parameters