关于「单峰偏好与值域限制:循环在什么条件下消失」,《不存在完美裁决器》给出的定量结论是什么?

结论:\(w\) 为零时循环概率为零,与布莱克定理一致。\(w\) 只要升到 0.1,三人组的循环概率就是 0.0117,五人组是 0.0116;\(w\) 到 0.5 时,三人组是 0.0434,五人组是 0.0509;\(w\) 为一时回到不偏文化的 0.0556 与 0.0694。最后两行给出一个直观的比例:十分之一的离轴噪音,就把三人组完全无共识时循环风险的 0.2102 带了回来,五人组是 0.1677。

数据口径:多道检查意见不一时,经营者最常见的想法是:只要合成规则设计得足够聪明,总能得到一个公正、一致、无可挑剔的裁决。 阿罗 1950 年与 1951 年证明,选项不少于三个时,不存在同时满足无限制域、弱帕累托、无关选项独立与非独裁的排序合成规则。

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: 0.0116, 0.0117, 0.0434, 0.0509, 0.0556, 0.0694

wP(cycle) n=3P(cycle) n=5
0.00.00000.0000
0.10.01170.0116
0.20.02180.0229
0.30.03040.0333
0.40.03760.0427
0.50.04340.0509
0.60.04800.0576
0.70.05140.0628
0.80.05380.0665
0.90.05510.0687
1.00.05560.0694
n = 3: w = 0.1 restores 0.2102 of the full cycle risk
n = 5: w = 0.1 restores 0.1677 of the full cycle risk
查看完整验算程序
# Chapter 4. p43: a little off-axis noise brings cycles back.
# Reviewers judge three options on a strictness axis A < B < C. With
# probability 1 - w a reviewer holds a single-peaked ranking (one of
# ABC, BAC, BCA, CBA, uniformly); with probability w the reviewer
# ignores the axis and picks any of the 6 rankings uniformly. We
# compute P(no Condorcet winner) exactly for 3 and 5 reviewers by
# enumerating all 6 ** n profiles with Fraction weights. w is a free
# illustrative parameter; w = 1 is the impartial culture of chapter 2.
import itertools
from fractions import Fraction

OPTS = "ABC"
ORDERS = ["".join(p) for p in itertools.permutations(OPTS)]
SP = {"ABC", "BAC", "BCA", "CBA"}


def has_winner(prof):
    for x in OPTS:
        if all(2 * len([o for o in prof if o.index(x) < o.index(y)])
               > len(prof) for y in OPTS if y != x):
            return True
    return False


def p_cycle(n, w):
    prob = {o: (1 - w) * Fraction(1, 4) * (o in SP) + w * Fraction(1, 6)
            for o in ORDERS}
    tot = Fraction(0)
    for prof in itertools.product(ORDERS, repeat=n):
        if has_winner(prof):
            continue
        q = Fraction(1)
        for o in prof:
            q *= prob[o]
        tot += q
    return tot


if __name__ == "__main__":
    print("   w    P(cycle) n=3   P(cycle) n=5")
    for k in range(0, 11):
        w = Fraction(k, 10)
        print("  {:.1f}      {:.4f}         {:.4f}".format(
            float(w), float(p_cycle(3, w)), float(p_cycle(5, w))))
    for n in (3, 5):
        r = p_cycle(n, Fraction(1, 10)) / p_cycle(n, Fraction(1))
        print("n = {}: w = 0.1 restores {:.4f} of the full cycle risk".format(
            n, float(r)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《阿罗不可能定理与多评审规则聚合、五件套拒绝码冲突时为何不存在完美裁决器母典》第 4 章「单峰偏好与值域限制:循环在什么条件下消失」(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-310 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026310,
  author    = {Fan, Yafu},
  title     = {Arrow's Impossibility Theorem and the Aggregation of Multiple Reviewers: Why No Perfect Arbiter Exists When Kit Refusal Codes Conflict},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论