关于「关系强度的指数衰减:半衰期、滑层时间与维护间隔」,《邓巴分层与私域半衰期》给出的定量结论是什么?

结论:第三章把关系写成指数衰减,给出守住每层的最长联系间隔,半衰期 60 天的客户守住贵宾层约需每 10.2 天联系一次。

数据口径:书中的层级下限、半衰期、权重与停止概率均为示意参数;全部计算由书中程序生成;本书是非个人化经营建议,合规规则以各地现行法规为准。 书中以标注框区分已发表的证据、可运行的 Python 计算(19 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: 10.2, 60

beta 0.5 friend (180 d) customer (60 d)
floor T* days per year T* days per year check
530.6 11.9 10.2 35.8 0.800
1574.7 4.9 24.9 14.7 0.600
50145.3 2.5 48.4 7.5 0.400
150285.3 1.3 95.1 3.8 0.200
customer T* (days) as beta varies
floor beta 0.3 beta 0.5 beta 0.7
56.3 10.2 14.0
1515.8 24.9 33.2
5032.2 48.4 62.1
15068.3 95.1 115.6

核心公式

\[s_{\min}(T) = \frac{\beta x}{1 - (1 - \beta) x}.\]
查看完整验算程序
# Chapter 3. p32: the longest contact interval that keeps a layer floor
# Model: between contacts s decays as exp(-lam * t); each contact
# closes a share beta of the gap to 1: s -> s + beta * (1 - s).
# With contact every T days the sawtooth settles at a minimum
#   s_min(T) = beta * x / (1 - (1 - beta) * x),  x = exp(-lam * T),
# so s_min >= f holds for T <= T* = ln((beta + f * (1 - beta)) / f) / lam.
# Floors as in p31; beta = 0.5 (illustrative); half-lives 180 days
# (friend) and 60 days (customer), illustrative. A step-by-step
# iteration of 40 cycles checks the closed form.
import math

FLOORS = [("5", 0.8), ("15", 0.6), ("50", 0.4), ("150", 0.2)]


def t_star(f, beta, lam):
    return math.log((beta + f * (1.0 - beta)) / f) / lam


def s_min_iter(T, beta, lam, cycles=40):
    s = 1.0
    low = 1.0
    for _ in range(cycles):
        s = s * math.exp(-lam * T)
        low = s
        s = s + beta * (1.0 - s)
    return low


def main():
    beta = 0.5
    print("beta 0.5   friend (180 d)      customer (60 d)")
    print("floor   T* days  per year    T* days  per year  check")
    for name, f in FLOORS:
        row = []
        for h in (180.0, 60.0):
            lam = math.log(2.0) / h
            t = t_star(f, beta, lam)
            row += [t, 365.0 / t]
        lam_c = math.log(2.0) / 60.0
        chk = s_min_iter(row[2], beta, lam_c)
        print("{:>5s} {:9.1f} {:9.1f} {:10.1f} {:9.1f}  {:.3f}".format(
            name, *row, chk))
    print("customer T* (days) as beta varies")
    print("floor   beta 0.3  beta 0.5  beta 0.7")
    lam_c = math.log(2.0) / 60.0
    for name, f in FLOORS:
        vals = [t_star(f, b, lam_c) for b in (0.3, 0.5, 0.7)]
        print("{:>5s} {:9.1f} {:9.1f} {:9.1f}".format(name, *vals))


main()

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《邓巴数分层、关系维护时间预算与私域客户衰减半衰期母典》第 3 章「关系强度的指数衰减:半衰期、滑层时间与维护间隔」(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-231 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026231,
  author    = {Fan, Yafu},
  title     = {Dunbar's Layered Numbers, Relationship-Maintenance Time Budgets and the Decay Half-Life of Private-Channel Customers},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论