11.07.2014 Views

Multiattribute acceptance sampling plans - Library(ISI Kolkata ...

Multiattribute acceptance sampling plans - Library(ISI Kolkata ...

Multiattribute acceptance sampling plans - Library(ISI Kolkata ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Proof:<br />

Denoting the regret function value of the 1 st plan and 2 nd plan by R 1 & R 2 respectively we get<br />

a −<br />

R1 − R2<br />

= γ<br />

2<br />

g ( a r −1<br />

, m′<br />

( r −2<br />

) ) ( 0,<br />

−1<br />

(<br />

′<br />

−1 ) )<br />

r<br />

ar<br />

g m<br />

r ∑ g<br />

( x , ′ )<br />

x = 0<br />

r<br />

mr<br />

a −a<br />

- γ<br />

1g<br />

( a r −1<br />

, m(<br />

r −2<br />

) ) ( r r<br />

g 0, m( r −1 ) )<br />

x = 0<br />

To show that R 1 –R 2 > 0 we show that,<br />

( γ / )<br />

2 γ − ( m′−m<br />

e<br />

) ⎡<br />

/<br />

⎤<br />

a<br />

m ′ m<br />

r−1<br />

. F 1 > 1<br />

1 ⎢⎣ ( r−2)<br />

( r−2)<br />

⎥⎦<br />

where,<br />

ar<br />

− ar−1<br />

ar<br />

− a<br />

x<br />

r−1<br />

1 = ∑ [ r<br />

x<br />

F<br />

m' !] / ∑ [ r<br />

r /x r<br />

m r /x r !]<br />

x = 0<br />

x = 0<br />

r<br />

r<br />

r<br />

∑ − 1<br />

r<br />

g ( x , )<br />

r m r<br />

In order to prove theorem 2.3.2, we first prove the lemma stated below.<br />

… (2.3.5)<br />

Lemma:<br />

c<br />

∑ ( m ' / x!<br />

) / ∑ ( m ′ / x!)<br />

> ∑ ( m / x!)<br />

/ ∑ ( m / x!)<br />

; for c ≥ 1 and m′ > m .<br />

x = 0<br />

x<br />

c −1<br />

x = 0<br />

x<br />

c<br />

x=<br />

0<br />

x<br />

c−1<br />

x=<br />

0<br />

x<br />

….(2.3.6)<br />

c−1<br />

x−c<br />

Proof of the lemma: The function ∑ m / x!<br />

c ≥ 1, decreases as m increases . Thus,<br />

x=<br />

0<br />

c<br />

x<br />

c−1<br />

x<br />

1<br />

∑ ( m / x!) / ∑ ( m / x!)<br />

= 1 +<br />

is an increasing function of m and<br />

x=<br />

0<br />

x=<br />

0<br />

c−1<br />

x−c<br />

c!<br />

∑ m / x!<br />

x=<br />

0<br />

therefore (2.3.6) holds.<br />

Let us come back to theorem 2.3.2. We continue with the proof in the following lines:<br />

( r 1)<br />

Since optimal a<br />

r −1<br />

must satisfy the inequality ∆ R1<br />

> 0 for the first optimal plan,<br />

( r −1)<br />

a r<br />

− a r −1<br />

where ∆ R1<br />

is the standard notation for the increase in R 1<br />

achieved when<br />

−1<br />

replaced by a 1<br />

r −1 + keeping all other parameters unaltered.<br />

The condition:<br />

a is<br />

r −1<br />

97

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!