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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

a<br />

1<br />

a<br />

2<br />

− x<br />

(1)<br />

a<br />

j−2<br />

− x<br />

( j−3)<br />

= ∑ ∑...<br />

∑ ∏ g ( x , m ) g ( a<br />

x<br />

x1 = 0 x2<br />

= 0 x<br />

j−2<br />

= 0<br />

a<br />

'<br />

j+<br />

1<br />

∑<br />

j+<br />

1<br />

= 0<br />

a<br />

'<br />

j+<br />

1<br />

x<br />

− x<br />

∑<br />

j+<br />

2<br />

j+<br />

1<br />

= 0<br />

....<br />

a<br />

'<br />

r<br />

−(<br />

x<br />

j −2<br />

i=<br />

1<br />

j+<br />

1<br />

x<br />

r<br />

i<br />

+ ..... + x<br />

∑<br />

= 0<br />

i<br />

r−1<br />

)<br />

r<br />

∏<br />

i=<br />

j + 1<br />

j<br />

− x , m ) g(0,<br />

m ) .<br />

g ( x<br />

( j −2)<br />

j −1<br />

j<br />

It becomes obvious by arguments similar to those used in theorems 2.3.2 and 2.3.3, from<br />

equation (2.3.13) that<br />

i<br />

, m<br />

i<br />

) .<br />

RA ( n;<br />

a , a<br />

2<br />

,..., a<br />

j 1<br />

= a<br />

j<br />

, a<br />

j<br />

,..., ar<br />

) − RA(<br />

n;<br />

a1<br />

, a<br />

2<br />

,..., a<br />

j − 1<br />

= a<br />

− 1, a<br />

1 − j j r ≤ 0<br />

if the given condition of the theorem is satisfied.<br />

Thus, we see that if the condition of the theorem is satisfied, then<br />

RA ( n ; a 1 …,a j-1 , a j , a j , a j+1 ,…, a r ) ≤ RA ( n ; a 1 ,…, a j-1 , a j-1 , a j , a j+1 ,…,a r ).<br />

... a<br />

)<br />

…(2.3.14)<br />

Let us consider the feasible interval for the (j-1) th <strong>acceptance</strong> number, denoted by c j-1 , viz.,<br />

a j-1 ≤ c j-1 ≤ a j (recall a j-1 < a j given) and keep all other <strong>acceptance</strong> numbers fixed as in the<br />

given optimal A plan.<br />

Then, RA ( n ; a 1 …,a j-1 , c j-1 , a j , a j+1 , …, a r ) treated as a function of a single variable c j-1<br />

keeping all other <strong>acceptance</strong> numbers fixed at their optimal values, has to exhibit one of the<br />

following three features in the interval a j-1 ≤ c j-1 ≤ a j :<br />

I. The function is monotonically decreasing in the interval.<br />

II. The function is monotonically increasing in the interval.<br />

III. The function first decreases monotonically and then increases monotonically.<br />

[In the definition monotonically decreasing (increasing) includes the possibility of equality<br />

also and does not mean strictly monotonically decreasing (increasing).]<br />

The only possibility ruled out is first increasing and then decresing or a feature of multiple<br />

waves. This is clearly not possible for a realistic regret function which can be supposed to<br />

possess one of the three properties stated for any feasible segment of values of c j-1 .<br />

Now because of (2.3.14), the only possibility is:<br />

RA ( n ; a 1 …,a j-1 , a j-1 , a j ,..,a r )<br />

= RA ( n ; a 1 …,a j-1 , a j-1 +1, a j , … ,a r )<br />

= RA ( n ; a 1 , …,a j-1 , a j , a j ,…, a r )<br />

103

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

Saved successfully!

Ooh no, something went wrong!