关于「两种风险的双重约束:搜索最小的样本量与接收数」,《验收抽样与到货验货》给出的定量结论是什么?
结论:第三章在两种风险约束下搜索最小方案,并指出零接收数方案让百分之一不良的正常批被误退 0.3638。
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。
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: 0.3638
| c | n | Pa(0.05) | risk@0.01 | risk@0.02 | extra n |
| 0 | 45 | 0.0994 | 0.3638 | 0.5971 | |
| 1 | 77 | 0.0973 | 0.1800 | 0.4573 | +32 |
| 2 | 105 | 0.0992 | 0.0888 | 0.3506 | +28 |
| 3 | 132 | 0.0992 | 0.0443 | 0.2719 | +27 |
| 4 | 158 | 0.0996 | 0.0219 | 0.2108 | +26 |
| 5 | 184 | 0.0982 | 0.0110 | 0.1652 | +26 |
| 6 | 209 | 0.0983 | 0.0054 | 0.1282 | +25 |
| c=0, n=45: a 1% lot is rejected 3.6 times in 10 |
查看完整验算程序
# Chapter 3. p33: what a zero-acceptance plan costs the good supplier.
# Model: binomial OC. For each acceptance number c, the smallest n
# that holds the consumer's risk at p2 = 0.05 to <= 0.10; then the
# producer's risk 1 - Pa at p1 = 0.01 and at p1 = 0.02 (lots that
# should normally pass). Quality levels are illustrative.
import math
def pa(p, n, c):
terms = [math.comb(n, d) * p ** d * (1 - p) ** (n - d)
for d in range(c + 1)]
return math.fsum(terms)
if __name__ == "__main__":
p2, beta = 0.05, 0.10
print(" c n Pa(0.05) risk@0.01 risk@0.02 extra n")
prev = None
for c in range(0, 7):
n = c + 1
while pa(p2, n, c) > beta:
n += 1
extra = "" if prev is None else "{:+d}".format(n - prev)
print(" {} {:4d} {:.4f} {:.4f} {:.4f} {}".format(
c, n, pa(p2, n, c), 1 - pa(0.01, n, c), 1 - pa(0.02, n, c),
extra))
prev = n
n0 = 45
print("c=0, n=45: a 1% lot is rejected {:.1f} times in 10".format(
10 * (1 - pa(0.01, n0, 0))))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 3 章「两种风险的双重约束:搜索最小的样本量与接收数」(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}
}