关于「五件套拒绝码冲突:示意场景、死锁概率与优先级顺序」,《不存在完美裁决器》给出的定量结论是什么?

结论:表中有两个方向相反的趋势。检查越多,死锁越容易:拒绝率 0.10、两个动作时,三道检查的死锁概率是 0.0734,五道检查升到 0.1677。动作越多,死锁越难:同样是三道检查、拒绝率 0.10,两个动作时 0.0734,四个动作时只有 0.0054。可是第二章刚刚算过,动作越多,多数决越容易转圈。增加备选动作在减少死锁的同时增加了循环,两者无法同时压到零。

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: 0.0054, 0.0734, 0.10, 0.1677

pkmP(deadlock)after relaxing 1after relaxing 2
0.05320.02030.00950.0025
0.05330.00290.00090.0001
0.05340.00040.00010.0000
0.05520.05120.03440.0203
0.05530.01160.00640.0029
0.05540.00260.00120.0004
0.10320.07340.03610.0100
0.10330.01990.00690.0010
0.10340.00540.00130.0001
0.10520.16770.11830.0734
0.10530.06870.04070.0199
0.10540.02810.01400.0054
0.20320.23810.12960.0400
0.20330.11620.04670.0080
0.20340.05670.01680.0016
0.20520.45200.34860.2381
0.20530.30390.20580.1162
0.20540.20430.12150.0567
more guards raise the deadlock risk; more options lower it
查看完整验算程序
# Chapter 6. p62: when every option draws at least one refusal.
# k guards each refuse each of m options independently with probability
# p (illustrative; real refusal rates must be read from the kit logs).
# Under "any refusal blocks", no option is left when every option is
# refused by someone:  P(deadlock) = (1 - (1 - p) ** k) ** m.
# A written priority lets the arbiter relax the lowest-ranked guard
# first; after relaxing one guard the deadlock risk becomes the same
# formula with k - 1. Exact Fraction arithmetic, printed as decimals.
from fractions import Fraction


def dead(p, k, m):
    return (1 - (1 - p) ** k) ** m


if __name__ == "__main__":
    print(" p     k  m   P(deadlock)  after relaxing 1  after relaxing 2")
    for p in (Fraction(1, 20), Fraction(1, 10), Fraction(1, 5)):
        for k in (3, 5):
            for m in (2, 3, 4):
                print("{:.2f}  {:>2} {:>2}     {:.4f}         {:.4f}"
                      "           {:.4f}".format(
                          float(p), k, m, float(dead(p, k, m)),
                          float(dead(p, k - 1, m)),
                          float(dead(p, k - 2, m))))
    print("more guards raise the deadlock risk; more options lower it")

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 6 章「五件套拒绝码冲突:示意场景、死锁概率与优先级顺序」(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