关于「超几何操作特性曲线:一件一件精确算出接收概率」,《验收抽样与到货验货》给出的定量结论是什么?

结论:读这张表要从两端往中间看。不良率为零时,接收概率当然是 1.0000。不良率为 0.010,即五百件里有五件不良,接收概率是 0.9194,并没有达到 0.95;要让接收概率至少 0.95,不良率要降到 0.006,也就是三件不良。不良率升到 0.020,接收概率降到 0.7365,这意味着一个长期稳定在百分之二不良的供应商,每四批左右就会被判退一批。不良率到 0.050,接收概率是 0.2636,仍有四分之一以上的机会放行;到 0.100,接收概率 0.0278,才算基本拦住。

数据口径:书中批量、不良率与成本均为示意参数;全部计算由书中程序生成;本书属于非个人化经营建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(22 段程序)、原典与作者的假说。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.006, 0.010, 0.020, 0.0278, 0.050, 0.100

OC curve, N = 500, n = 50, c = 1
pDP(accept)
0.00001.0000
0.01050.9194
0.020100.7365
0.030150.5469
0.040200.3865
0.050250.2636
0.060300.1749
0.070350.1134
0.080400.0722
0.090450.0451
0.100500.0278
producer point (Pa >= 0.95): p = 0.006
indifference(Pa ~ 0.50):p = 0.032
LTPD(Pa <= 0.10): p = 0.074
查看完整验算程序
# Chapter 2. p22: the OC curve of one plan and its three landmarks.
# Model: lot N = 500, plan n = 50, c = 1, exact hypergeometric.
# Landmarks: the largest defect share accepted with prob >= 0.95
# (the producer's point), the share accepted with prob closest to
# 0.50 (indifference) and the smallest share accepted with prob
# <= 0.10 (LTPD, the consumer's point). Plan is illustrative.
import math
from fractions import Fraction


def hyper(N, D, n, c):
    num = 0
    for d in range(0, min(c, D, n) + 1):
        num += math.comb(D, d) * math.comb(N - D, n - d)
    return Fraction(num, math.comb(N, n))


if __name__ == "__main__":
    N, n, c = 500, 50, 1
    pa = [hyper(N, D, n, c) for D in range(0, 81)]
    print("OC curve, N = 500, n = 50, c = 1")
    print("  p       D    P(accept)")
    for D in range(0, 55, 5):
        print("  {:.3f}  {:2d}   {:.4f}".format(D / N, D, float(pa[D])))
    p95 = max(D for D in range(81) if pa[D] >= Fraction(95, 100))
    p50 = min(range(81), key=lambda D: abs(pa[D] - Fraction(1, 2)))
    p10 = min(D for D in range(81) if pa[D] <= Fraction(1, 10))
    print("producer point (Pa >= 0.95): p = {:.3f}".format(p95 / N))
    print("indifference   (Pa ~ 0.50):  p = {:.3f}".format(p50 / N))
    print("LTPD           (Pa <= 0.10): p = {:.3f}".format(p10 / N))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《验收抽样方案与操作特性曲线、买手到货验货抽检件数母典》第 2 章「超几何操作特性曲线:一件一件精确算出接收概率」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-329 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026329,
  author    = {Fan, Yafu},
  title     = {Acceptance Sampling Plans and Operating-Characteristic Curves: How Many Items a Buyer Must Inspect on Arrival},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论