|
Draws a Monte Carlo
simulated curve, which is a histogram, approaching the density curve of the
quality index for an inverse Gaussian variable. (A proposed approximation to
the cdf is [provisionally] being used.)
In acceptance sampling
for this variable, the “quality index” is
Q = (xbar − L) ⁄
(xbar r √ L), with
rē = Σ(1⁄xi −
1⁄xbar).
For a suitable simulation,
the following steps are suggested: (a) find the
sampling
plan (with β 10 %) for λ unknown;
(b) insert above the data obtained ( n, L, μ,
AQL, α, LTPD); and (c) run the simulation. These are
the conditions to reproduce k. In the table beside, the
following values are given for each n: risk, α; k as
per ANSI/ASQC Z1.9; LTPD (in parentheses).
|
| n | α |
AQL: 1 (%) |
1.5 (%) |
| 5 | 10 |
k: 1.53 (28.5 %)* |
1.40 (31 %) |
| 7 | 10 |
1.62 (21 %) | 1.50 (23.5 %) |
| 10 | 10 |
1.72 (15.2 %) | 1.58 (18 %) |
| 20 | 8 |
1.82 (9.2 %) | 1.69 (11 %) |
| 25 | 7 |
1.85 (8 %) | 1.72 (9.8 %) |
| 50 | 5 |
1.93 (5.3 %) | 1.80 (6.7 %) |
* ANSI/ASQC LTPD value
|
