关于「残留衰减方程:目标记忆、收尾动作与时间常数 τ」,《注意力残留与冲刺排程》给出的定量结论是什么?
结论:第二章由目标记忆的激活模型推出残留衰减方程:敞口离场、写下一步与已收口三种方式的时间常数分别约为 22.33、8.68 与 6.45 分钟,并给出从台账估计个人 τ 的设计。
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: 22.33, 6.45, 8.68
| A. P(old goal intrudes), minutes after the switch | |
| case r | 1 2 5 10 20 40 |
| open | 1.00 0.956 0.930 0.825 0.642 0.414 0.251 |
| note | 0.25 0.927 0.876 0.723 0.357 0.110 0.031 |
| closed 0.00 | 0.927 0.876 0.651 0.273 0.052 0.007 |
| B. log-linear fit of P over 1..20 min: tau in minutes | |
| open tau | 22.33 ratio P40/P1 0.262 |
| note tau | 8.68 ratio P40/P1 0.034 |
| closed tau | 6.45 ratio P40/P1 0.007 |
查看完整验算程序
# Chapter 2. p21: goal activation after a switch, memory-for-goals style
# Base-level activation of a goal (ACT-R form used by Altmann & Trafton):
# A(t) = ln( sum_j (t - t_j)^(-d) ), d = 0.5, times in seconds.
# Old goal X was rehearsed every 30 s for 20 minutes before the switch.
# New goal Y is rehearsed every 30 s from the switch on. An unfinished X
# keeps intruding: extra rehearsals at r per minute (illustrative):
# open r = 1.0 (left mid-step, nothing written)
# note r = 0.25 (next step written down, intention offloaded)
# closed r = 0.0 (task finished or explicitly closed)
# Probability that a retrieval picks X instead of Y (logistic noise s):
# P(t) = 1 / (1 + exp((A_Y - A_X) / s)), s = 0.3
# Each table cell is P averaged over the minute that ends at that mark,
# sampled every 5 s, so single rehearsal spikes do not dominate.
import math
D = 0.5
S = 0.3
OLD = [-1200.0 + 30.0 * k for k in range(40)]
def act(times, t):
terms = [(t - tj) ** (-D) for tj in times if tj < t]
return math.log(math.fsum(terms))
def p_intrude(rate, t):
new = [30.0 * k for k in range(int(t // 30.0) + 1)]
extra = []
if rate > 0.0:
gap = 60.0 / rate
extra = [gap * k for k in range(1, int(t // gap) + 1)]
ax = act(OLD + extra, t + 0.5)
ay = act(new, t + 0.5)
return 1.0 / (1.0 + math.exp((ay - ax) / S))
MINUTES = (1, 2, 5, 10, 20, 40)
print("A. P(old goal intrudes), minutes after the switch")
print("case r " + "".join("{:>7}".format(m) for m in MINUTES))
curves = {}
for name, rate in (("open", 1.0), ("note", 0.25), ("closed", 0.0)):
ps = []
for m in MINUTES:
grid = [60.0 * (m - 1) + 5.0 * k for k in range(1, 13)]
ps.append(math.fsum(p_intrude(rate, t) for t in grid) / 12.0)
curves[name] = ps
print("{:<6} {:>4.2f} ".format(name, rate)
+ "".join("{:>7.3f}".format(p) for p in ps))
print("B. log-linear fit of P over 1..20 min: tau in minutes")
for name in ("open", "note", "closed"):
xs = [float(m) for m in MINUTES[:5]]
ys = [math.log(p) for p in curves[name][:5]]
mx = math.fsum(xs) / len(xs)
my = math.fsum(ys) / len(ys)
sxy = math.fsum((x - mx) * (y - my) for x, y in zip(xs, ys))
sxx = math.fsum((x - mx) ** 2 for x in xs)
slope = sxy / sxx
tau = -1.0 / slope if slope < 0.0 else float("inf")
print("{:<6} tau {:>7.2f} ratio P40/P1 {:>6.3f}".format(
name, tau, curves[name][5] / curves[name][0]))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 2 章「残留衰减方程:目标记忆、收尾动作与时间常数 τ」(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}
}