What does A Convergence Criterion for the Self-Evolving Engine conclude, in numbers, about “Difference Equations and Stability: Pruning Gain, Audit Delay and Oscillation”?

Answer: Chapter 3 writes difference equations with audit delay: the stable limit of pruning gain is 1.0000 with a one-quarter delay and 0.6180 with two; with too little pruning capacity, conflicts grow 3.78-fold in twenty quarters, falling to 1.78-fold once obsolete rules retire.

Scope of the numbers: Its parameters are illustrative and contain no production information; all calculations are generated by its programs; it is general business analysis, not business advice. Chapter 4 sorts with Kahn's algorithm and finds and repairs cycles with strongly connected components; in a synthetic citation graph of three hundred volumes, 18,828 pairs are comparable, a share of 0.420.

中文版

Data table

The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 0.6180, 1.0000

Lclosed form g*simulated g*half-life at g*/2
02.00002.00001 quarters
11.00001.00002 quarters
20.61800.61804 quarters
30.44500.44506 quarters
40.34730.34737 quarters
50.28460.28469 quarters
each extra quarter of audit delay lowers the safe gain
Show the full program
# Chapter 3. p32: largest stable pruning gain for an audit delay of L.
# Model: x[t+1] = x[t] - g * x[t-L]. Its characteristic polynomial is
# z^(L+1) - z^L + g = 0; all roots lie inside the unit circle iff
# 0 < g < 2 sin(pi / (2 (2L + 1))) (a classical result for this
# delay equation). We check the closed form by bisection on simulated
# trajectories: g is stable if late peaks are smaller than early ones.
import math


def stable(g, lag, steps=4000):
    xs = [1.0] * (lag + 1)
    for _ in range(steps):
        xs.append(xs[-1] - g * xs[-1 - lag])
        if abs(xs[-1]) > 1e100:
            return False
    early = max(abs(v) for v in xs[1800:2000])
    late = max(abs(v) for v in xs[3800:4000])
    return late < early or late == 0.0


def bisect_gain(lag):
    lo, hi = 0.0, 2.5
    for _ in range(40):
        mid = (lo + hi) / 2
        if stable(mid, lag):
            lo = mid
        else:
            hi = mid
    return lo


if __name__ == "__main__":
    print("  L   closed form g*   simulated g*   half-life at g*/2")
    for lag in range(6):
        gs = 2 * math.sin(math.pi / (2 * (2 * lag + 1)))
        sim = bisect_gain(lag)
        xs = [1.0] * (lag + 1)
        t = 0
        while max(abs(v) for v in xs[-(lag + 1):]) > 0.5:
            xs.append(xs[-1] - gs / 2 * xs[-1 - lag])
            t += 1
        print("  {}   {:>14.4f}   {:>12.4f}   {:>8} quarters".format(
            lag, gs, sim, t))
    print("each extra quarter of audit delay lowers the safe gain")

Single source and citation

Fan, Y. (2026). Topological Ordering of the Theorem-Dependency Graph of Three Hundred Volumes and a Convergence Criterion for the Self-Evolving Engine: Halting Conditions between Proposal Noise and Pruning Gain (Version 1.0.0), Chapter 3: Difference Equations and Stability: Pruning Gain, Audit Delay and Oscillation. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-332 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026332,
  author    = {Fan, Yafu},
  title     = {Topological Ordering of the Theorem-Dependency Graph of Three Hundred Volumes and a Convergence Criterion for the Self-Evolving Engine: Halting Conditions between Proposal Noise and Pruning Gain},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book