What does One Commander and the Serial Bottleneck conclude, in numbers, about “The Commander Is a Single Server: Queueing Theory, the Finite-Source Model and Little's Law”?

Answer: Chapter 3 treats the commander as a single server; in the finite-source model lowering the serial share releases 9.667 items an hour, while adding agents only adds backlog.

Scope of the numbers: With illustrative parameters, cutting the serial share from a fifth to a twentieth raises the speedup by 3.325, while raising the agents from ten to a hundred adds only 1.236. Its parameters are illustrative; all calculations are generated by its programs.

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: 9.667

caseitems/hspeedupL queueW min
base N=10 s=0.204.9734.9736.0272.6
A N=100 s=0.205.0005.00096.001152.0
B N=10 s=0.059.6679.6670.825.1
A+B N=100 s=0.0520.00020.00081.00243.0
Little check, base: L = X W = 4.973 x 1.2107 h = 6.02

Key formula

\[W_q=\frac{\rho}{1-\rho}\cdot\frac{1+c_s^2}{2}\cdot E[S],\]
Show the full program
# Chapter 3. p33: Two levers in the same queue, and Little's law.
# Model: the finite-source queue of p32 (agents work 1/lam, then
# wait for one commander's review 1/mu, exponential times). Base:
# ten agents, 48 min work, 12 min review. Lever A: ten times more
# agents. Lever B: review cut to 3 min (agent self-checks absorb the
# difference, so work becomes 57 min and s = 0.05). Little's law
# L = X W reads the number of agents stuck at the commander from the
# throughput X and the response time W. Illustrative parameters.
import math
from fractions import Fraction as F


def queue(n, work, review):
    rho = F(review, work)
    terms = [F(1)]
    for k in range(1, n + 1):
        terms.append(terms[-1] * (n - k + 1) * rho)
    total = math.fsum(float(t) for t in terms)
    p = [float(t) / total for t in terms]
    x = (1.0 - p[0]) * 60.0 / review
    length = math.fsum(k * pk for k, pk in enumerate(p))
    w = length / x * 60.0
    return x, length, w


cases = [("base N=10 s=0.20", 10, 48, 12),
         ("A N=100 s=0.20", 100, 48, 12),
         ("B N=10 s=0.05", 10, 57, 3),
         ("A+B N=100 s=0.05", 100, 57, 3)]
print("%-18s %8s %8s %9s %10s" % ("case", "items/h", "speedup",
                                  "L queue", "W min"))
for name, n, work, review in cases:
    x, length, w = queue(n, work, review)
    spd = x * (work + review) / 60.0
    print("%-18s %8.3f %8.3f %9.2f %10.1f" % (name, x, spd, length, w))
x, length, w = queue(10, 48, 12)
print("Little check, base: L = X W = %.3f x %.4f h = %.2f" % (
    x, w / 60.0, x * w / 60.0))

Single source and citation

Fan, Y. (2026). Constitutional Engineering for a One-Commander Silicon Empire: Serial Decision Bandwidth and Breaking Amdahl's Bottleneck (Version 1.0.0), Chapter 3: The Commander Is a Single Server: Queueing Theory, the Finite-Source Model and Little's Law. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-253 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026253,
  author    = {Fan, Yafu},
  title     = {Constitutional Engineering for a One-Commander Silicon Empire: Serial Decision Bandwidth and Breaking Amdahl's Bottleneck},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book