关于「跨国供应链:等效电阻、最大流与一个港口的分量」,《谱间隙与瓶颈割集》给出的定量结论是什么?

结论:第五章用等效电阻与最大流比较三种线路数相同的供应网,交织设计把单季送达不足一半的概率压到 0.0203。

数据口径:书中参数为示意;全部计算由书中程序生成;本书是非个人化经营建议,任何生产系统或真实供应链的改动须由主事者人工复核放行。 第八章给出画图、体检、补边、复查的季度闭环,示意工作流三轮补边后代数连通度从 0.1824 升到 1.0000;蒸馏小工具「瓶颈割集体检器」;以及每组 74 个团队季度的预注册检验方案与推翻条件。

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: 0.0203

designnominalworst linkworst sitemean shareP(<=0.5)
A main port40.750.250.87090.0777
B port per country40.750.500.89090.0813
C cross-woven40.750.750.95500.0203
查看完整验算程序
# Chapter 5. p52: delivered share of the three supply designs of p51
# under random link failures and site closures, by max flow.
# Model: every link carries one unit (one tenth of nominal volume) in
# either direction; each supplier ships one unit, each market takes one
# unit, nominal flow 4 units. Per quarter each link fails with
# probability 0.05 and each port or factory closes with probability
# 0.02, independently (illustrative rates). Max flow by Edmonds-Karp;
# 4,000 quarters per design with common random numbers, seed fixed.
import collections
import math
import random

SITES = ["S1", "S2", "S3", "S4", "P1", "P2", "F1", "F2", "F3",
         "M1", "M2", "M3", "M4"]
IX = {s: k for k, s in enumerate(SITES)}
MARKET = [("F1", "M1"), ("F2", "M1"), ("F1", "M2"), ("F3", "M2"),
          ("F2", "M3"), ("F3", "M3"), ("F1", "M4"), ("F3", "M4")]
DESIGNS = [
    ("A main port", [("S1", "P1"), ("S2", "P1"), ("S3", "P1"),
                     ("S4", "P1"), ("S4", "P2"), ("P1", "F1"),
                     ("P1", "F2"), ("P1", "F3"), ("P2", "F3"),
                     ("F1", "F2"), ("F2", "F3"), ("F1", "F3")] + MARKET),
    ("B port per country", [("S1", "P1"), ("S2", "P1"), ("S3", "P2"),
                            ("S4", "P2"), ("S2", "P2"), ("S3", "P1"),
                            ("P1", "F1"), ("P1", "F2"), ("P2", "F2"),
                            ("P2", "F3"), ("F1", "F2"), ("F2", "F3")]
     + MARKET),
    ("C cross-woven", [("S1", "P1"), ("S1", "P2"), ("S2", "P1"),
                       ("S3", "P2"), ("S4", "P1"), ("S4", "P2"),
                       ("P1", "F1"), ("P1", "F2"), ("P1", "F3"),
                       ("P2", "F1"), ("P2", "F2"), ("P2", "F3")]
     + MARKET),
]
SRC, SNK = 13, 14


def maxflow(links, closed):
    res = collections.defaultdict(int)
    adj = collections.defaultdict(list)

    def arc(u, w, c):
        res[(u, w)] += c
        adj[u].append(w)
        adj[w].append(u)
    for a, b in links:
        if a in closed or b in closed:
            continue
        arc(a, b, 1)
        arc(b, a, 1)
    for s in range(4):
        arc(SRC, s, 1)
    for m in range(9, 13):
        arc(m, SNK, 1)
    flow = 0
    while True:
        prev = {SRC: None}
        dq = collections.deque([SRC])
        while dq and SNK not in prev:
            u = dq.popleft()
            for w in adj[u]:
                if w not in prev and res[(u, w)] > 0:
                    prev[w] = u
                    dq.append(w)
        if SNK not in prev:
            return flow
        w = SNK
        while prev[w] is not None:
            u = prev[w]
            res[(u, w)] -= 1
            res[(w, u)] += 1
            w = u
        flow += 1


rng = random.Random(20261113)
draws = [([rng.random() for _ in range(20)], [rng.random() for _ in
                                              range(5)])
         for _ in range(4000)]
print("design              nominal  worst link  worst site  mean share"
      "  P(<=0.5)")
for name, links in DESIGNS:
    es = [(IX[a], IX[b]) for a, b in links]
    nom = maxflow(es, set())
    wl = min(maxflow([e for e in es if e != x], set()) for x in es)
    ws = min(maxflow(es, {v}) for v in range(4, 9))
    shares, low = [], 0
    for lu, su in draws:
        alive = [e for e, u in zip(es, lu) if u >= 0.05]
        closed = {4 + k for k, u in enumerate(su) if u < 0.02}
        f = maxflow(alive, closed) / nom
        shares.append(f)
        if f <= 0.5:
            low += 1
    shares.sort()
    mean = math.fsum(shares) / len(shares)
    print("%-19s %7d  %10.2f  %10.2f  %10.4f  %8.4f"
          % (name, nom, wl / nom, ws / nom, mean, low / len(draws)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《图拉普拉斯谱间隙、奇格常数与多智能体工作流及跨国供应链单点瓶颈割集母典》第 5 章「跨国供应链:等效电阻、最大流与一个港口的分量」(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-243 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026243,
  author    = {Fan, Yafu},
  title     = {Graph-Laplacian Spectral Gaps and Cheeger Constants: Single-Point Bottleneck Cuts in Multi-Agent Workflows and Cross-Border Supply Chains},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论