关于「冲刺长度:警觉度衰减、切换成本与九十分钟的来历」,《注意力残留与冲刺排程》给出的定量结论是什么?
结论:第三章求最优冲刺长度,它由警觉衰减常数与切换成本决定,近似为二者乘积两倍的平方根;在两千多种参数组合里,固定九十分钟只在 0.450 的格子里接近最优。
数据口径:书中参数为示意,全部计算由书中程序生成。 本书不是医疗、睡眠、营养或心理健康建议,不宣称任何功效。
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: 0.450
| A. linear decay, beta per minute | ||||
| beta | C | L_opt | E_opt | E90/Eopt |
| 0.004 | 3 | 35.8 | 0.857 | 0.926 |
| 0.004 | 8 | 55.7 | 0.777 | 0.969 |
| 0.004 | 15 | 72.9 | 0.708 | 0.992 |
| 0.008 | 3 | 24.5 | 0.804 | 0.771 |
| 0.008 | 8 | 37.4 | 0.701 | 0.839 |
| 0.008 | 15 | 48.0 | 0.616 | 0.891 |
| B. grid of 2,360 cells (half-life x cost) | ||||
| 90 min within 2 % of optimum: exp 0.450, linear 0.410 | ||||
| optimal length (exp): p10 42.0 | median 112.1 | p90 194.9 |
查看完整验算程序
# Chapter 3. p32: linear decay, and where a fixed 90 minutes is near-best
# Linear alertness a(t) = max(0, 1 - beta t). Efficiency
# E(L) = (L - beta L^2 / 2) / (L + C) for L <= 1 / beta,
# optimum L* = -C + sqrt(C^2 + 2 C / beta) (closed form).
# Region scan: half-life h = 20..600 min (exponential lam = h / ln 2;
# linear beta = 0.5 / h, same half-way point) and C = 1..40 min.
# Count grid cells where a fixed 90-minute sprint reaches >= 98 % of E*.
import math
def eff_lin(L, beta, c):
L = min(L, 1.0 / beta)
return (L - beta * L * L / 2.0) / (L + c)
def best_lin(beta, c):
return min(-c + math.sqrt(c * c + 2.0 * c / beta), 1.0 / beta)
def eff_exp(L, lam, c):
return lam * (1.0 - math.exp(-L / lam)) / (L + c)
def best_exp(lam, c):
lo, hi = 1e-6, 50.0 * lam
for _ in range(100):
mid = 0.5 * (lo + hi)
if math.exp(mid / lam) - 1.0 - (mid + c) / lam < 0.0:
lo = mid
else:
hi = mid
return 0.5 * (lo + hi)
print("A. linear decay, beta per minute")
print(" beta C L_opt E_opt E90/Eopt")
for beta in (0.004, 0.008):
for c in (3.0, 8.0, 15.0):
L = best_lin(beta, c)
e = eff_lin(L, beta, c)
print("{:>6.3f} {:>4.0f} {:>7.1f} {:>6.3f} {:>9.3f}".format(
beta, c, L, e, eff_lin(90.0, beta, c) / e))
halves = [20.0 + 10.0 * i for i in range(59)]
costs = [float(c) for c in range(1, 41)]
hit_exp = 0
hit_lin = 0
lens = []
for h in halves:
lam, beta = h / math.log(2.0), 0.5 / h
for c in costs:
le = best_exp(lam, c)
lens.append(le)
if eff_exp(90.0, lam, c) >= 0.98 * eff_exp(le, lam, c):
hit_exp = hit_exp + 1
ll = best_lin(beta, c)
if eff_lin(90.0, beta, c) >= 0.98 * eff_lin(ll, beta, c):
hit_lin = hit_lin + 1
cells = len(halves) * len(costs)
lens.sort()
print("B. grid of {:,} cells (half-life x cost)".format(cells))
print(" 90 min within 2 % of optimum: exp {:.3f}, linear {:.3f}".format(
hit_exp / cells, hit_lin / cells))
print(" optimal length (exp): p10 {:.1f} median {:.1f} p90 {:.1f}".format(
lens[cells // 10], lens[cells // 2], lens[9 * cells // 10]))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《多战线指挥官注意力残留清零、超日节律能量调配与跨场景秒级切换母典》第 3 章「冲刺长度:警觉度衰减、切换成本与九十分钟的来历」(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-362 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026362,
author = {Fan, Yafu},
title = {Eliminating Attention Residue in Multi-Arena Command, Ultradian Energy Pacing and Rapid Context-Switching Protocols},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}