关于「把收费站、代购与出版写成臂:毛利、单位成本、先验强度与停线表」,《业务线周度配额》给出的定量结论是什么?

结论:按转化率排,代购第一(0.5000),收费站第二(0.3095),出版最后(0.2500)。按净均值排,收费站第一(5.2857),代购第二(5.0000),出版最后(2.0000):代购转化率最高,但每单毛利太薄。按净指数排,出版第一(7.1851),收费站第二(5.8142),代购最后(5.1241)。

数据口径:核对吉廷斯 1979 年的论文与后续文献以后,本书修正了立项蓝图:吉廷斯指数的最优性只在单线经营、闲置线不变、换线无成本、几何贴现时成立,周度拆分预算、换线成本与会变的市场都在定理之外;锁死是一个概率说法,两条示意线上按累计回报全押有 0.2133 的重复在下半年完全押错,但在每单位贴现 0.9 的短视界里贪心相对指数只损失 0.0050 到 0.0166。 书中参数为示意,毛利与成本用记账点数而非价格;全部计算由书中程序生成。

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: 0.2500, 0.3095, 0.5000, 2.0000, 5.0000, 5.1241

line  record margin    rate   index net-mean net-index
toll   12/40   30.0  0.3095  0.3271   5.2857    5.8142
buy    45/90   14.0  0.5000  0.5089   5.0000    5.1241
pub      0/2   40.0  0.2500  0.3796   2.0000    7.1851
rank by rate     : buy > toll > pub
rank by net-mean : toll > buy > pub
rank by net-index: pub > toll > buy
查看完整验算程序
# Chapter 6. p61: margins and unit costs turn rates into money indices.
# Model: a funded unit of line i converts with unknown rate p_i (Beta
# posterior from its record, uniform prior); a conversion earns margin
# m_i points and every unit costs c_i points (illustrative points, not
# prices). The index is affine-equivariant: for reward m X - c it is
# m nu - c, with nu the Bernoulli index (p21, beta 0.9). Rank the lines
# by rate, by expected net per unit, and by net index.
import math


def gittins(a, b, beta, depth=40, tol=1e-6):
    lo, hi = a / (a + b), 1.0
    while hi - lo > tol:
        lam = 0.5 * (lo + hi)
        r = lam / (1.0 - beta)
        n = a + b + depth
        v = [max(r, (a + k) / n / (1.0 - beta)) for k in range(depth + 1)]
        for d in range(depth - 1, -1, -1):
            n = a + b + d
            nv = []
            for k in range(d + 1):
                p = (a + k) / n
                c = math.fsum((p * (1.0 + beta * v[k + 1]),
                               (1.0 - p) * beta * v[k]))
                nv.append(c if c > r else r)
            v = nv
        if v[0] > r + 1e-12:
            lo = lam
        else:
            hi = lam
    return 0.5 * (lo + hi)


LINES = (("toll", 12, 40, 30.0, 4.0),
         ("buy", 45, 90, 14.0, 2.0),
         ("pub", 0, 2, 40.0, 8.0))
print("%-5s %6s %6s %7s %7s %8s %9s" % ("line", "record", "margin",
                                        "rate", "index", "net-mean",
                                        "net-index"))
rows = []
for name, w, n, m, c in LINES:
    a, b = 1 + w, 1 + n - w
    mu = a / (a + b)
    nu = gittins(a, b, 0.9)
    rows.append((name, mu, m * mu - c, m * nu - c))
    print("%-5s %6s %6.1f %7.4f %7.4f %8.4f %9.4f" % (
        name, "%d/%d" % (w, n), m, mu, nu, m * mu - c, m * nu - c))
for k, label in ((1, "rate"), (2, "net-mean"), (3, "net-index")):
    order = sorted(rows, key=lambda r: -r[k])
    print("rank by %-9s: %s" % (label, " > ".join(r[0] for r in order)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《多业务线算力与预算动态裁决、吉廷斯指数探索与收费站代购出版线周度配额母典》第 6 章「把收费站、代购与出版写成臂:毛利、单位成本、先验强度与停线表」(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-264 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026264,
  author    = {Fan, Yafu},
  title     = {Dynamic Allocation of Compute and Budget across Business Lines: Gittins-Index Exploration and Weekly Quotas for Tollbooth, Buying and Publishing Lines},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论