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.

The individual Poisson term is denoted as g( x,<br />

m)<br />

= Exp ( −m)<br />

m / x!<br />

and the cumulative<br />

c<br />

x<br />

x<br />

Poisson as G( c,<br />

m)<br />

= ∑ Exp ( −m)<br />

m / x!<br />

. We also introduce h ( x,<br />

m)<br />

= m / x!<br />

. Further the<br />

x=<br />

0<br />

Poisson probability of <strong>acceptance</strong> at ( p , p ,..., p ) , for the Plan A with sample size n and<br />

1 2 r<br />

<strong>acceptance</strong> number a , a ,..., a is denoted as: PA , ,..., ; , ,..., )<br />

1 2 r<br />

( a1 a2<br />

ar<br />

m1<br />

m<br />

2 mr<br />

and the<br />

corresponding probability of rejection at ( p<br />

1,<br />

p2,...,<br />

pr<br />

) as QA(<br />

a , a ,..., a ; m , m ,..., m )<br />

1 2 r 1 2 r<br />

= 1 − PA(<br />

a , a ,..., a ; m , m ,..., m ) .<br />

1 2 r 1 2 r<br />

x<br />

For the D kind plan with sample size n and <strong>acceptance</strong> number k the probability of<br />

<strong>acceptance</strong> at ( p , p ,..., p ) is denoted as PD ( k,<br />

m)<br />

. The corresponding probability of<br />

1 2 r<br />

rejection at ( p , p ,..., p ) is denoted by QD ( k,<br />

m)<br />

.<br />

1 2 r<br />

For given values of γ 1 , γ 2 the regret function defined in equation ( 2.2.9 ) for lots of size N<br />

using an A kind, with sample size n and with <strong>acceptance</strong> numbers ( a , a ,..., a ) is written<br />

as :<br />

1<br />

2<br />

r<br />

RA(<br />

n;<br />

a<br />

, a<br />

,..., a<br />

1 2 r<br />

) =<br />

n + ( N − n)[<br />

γ QA(<br />

a , a ,..., a ; m , m , a , a ,..., a m ) + γ PA(<br />

a<br />

, a<br />

,..., a<br />

; m'<br />

, m'<br />

,..., m'<br />

1 1 2 r 1 2 1 2 r r 2 1 2 r 1 2 r<br />

Similarly the regret function of the plan D with sample size n and <strong>acceptance</strong> number k :<br />

RD ( n,<br />

k ) = n + N − n)[<br />

γ QD(<br />

k,<br />

m)<br />

+ γ PD(<br />

k,<br />

')]<br />

(<br />

1 2<br />

m<br />

)]<br />

2.3.2 Situation 1<br />

Theorem (2. 3.1)<br />

Assume γ<br />

1<br />

, γ<br />

2<br />

, N , p, p'<br />

to be given. Given an optimal D plan with parameters (k, n), the<br />

regret function value of the plan A with <strong>acceptance</strong> criterion:<br />

x() i<br />

≤ k −1;<br />

for i = 1, 2,…,r-1; is less than the regret function value of the plan D (k, n) if<br />

k<br />

k + 1<br />

[ p '( r − 1) / p(<br />

r −1)<br />

] > ( p'<br />

/ p)<br />

…. (2.3.1)<br />

Proof:<br />

Since D (n, k) is an optimal plan RD (n, k+1) - RD (n, k) ≥ 0<br />

⎛ γ<br />

k + 1<br />

⎞<br />

⎜ 2 ⎟<br />

− ( m′−m)<br />

⎛ m′<br />

⎞<br />

e ⎜ ⎟ > 1 …. (2.3.2)<br />

⎝ γ<br />

1 ⎠ ⎝ m ⎠<br />

Now RD (n; k) - RA (n; a 1 = k-1, a 2 = k-1 ,..., a r-1 = k-1, a r = k) / (N-n) =<br />

γ2 [ g ( k,m′<br />

( r 1)<br />

) g (0,m′<br />

r<br />

)] − γ1<br />

[ g ( k,m(<br />

r 1 )<br />

) g (0,mr<br />

)]<br />

−<br />

−<br />

95

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

Saved successfully!

Ooh no, something went wrong!