关于「从两站到多站:流量天花板、最优站数与五站矩阵的裁剪」,《站群竞争相图》给出的定量结论是什么?
结论:第四章推广到多站,最优站数大致是系数的倒数,五个不等的站在自由竞争下只留下两个,合计 13126 名访客。
数据口径:立项命题经核对后修正:系数大于一时,只有被经营者撑住的内部平衡点低于单站,自由竞争的结局是只剩一个站;两个系数之积小于一只是共存的必要条件,示意网格中满足它的格子仍有 0.627 以一方出局告终。 书中站、账号、访客、成本与经营者均为示意;全部计算由书中程序生成;本书是非个人化经营建议,只讨论以真实主体运营、彼此关系公开的合规品牌矩阵,不涉及马甲账号与刷量。
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: 13126
| feasible subsets: 16 of 31 | |||
| subset | total | closed | traffic per open site |
| ML | 13126 | yes | L=1993 M=11134 |
| MT | 12504 | no | M=11311 T=1193 |
| M | 12000 | no | M=12000 |
| ST | 10279 | no | S=4689 T=5590 |
| STL | 10201 | no | L=567 S=4689 T=4945 |
| TDL | 9612 | no | D=6773 L=2672 T=166 |
| DL | 9593 | no | D=6823 L=2770 |
| TD | 9577 | no | D=6089 T=3488 |
| all five held open: infeasible | |||
| sum of capacities if no overlap: 35000 | |||
| simulated month 600: M=11134 S=0 T=0 D=0 L=1993 |
核心公式
\[N^* = \frac{K}{1 + (n-1)a}, \qquad T(n) = \frac{nK}{1 + (n-1)a} .\]
查看完整验算程序
# Chapter 4. p43: five unequal sites; which subset should stay open?
# Overlaps a_ij from audience profiles (MacArthur-Levins, as in p21),
# capacities K_i in visitors a month; all illustrative. For every one
# of the 31 subsets S we solve N_i + sum_{j in S} a_ij N_j = K_i
# exactly (fractions). S is feasible if every N_i > 0; it is closed
# to re-entry if no site outside S could grow back, that is
# K_i - sum_{j in S} a_ij N_j <= 0. A feasible, closed subset is
# where free competition can settle. We print the best totals and
# check the answer with a direct simulation of all five sites.
import math
from fractions import Fraction as Fr
from itertools import combinations
P = {"M": (20, 18, 16, 14, 12, 10, 6, 4),
"S": (40, 35, 15, 10, 0, 0, 0, 0),
"T": (0, 0, 5, 10, 15, 20, 25, 25),
"D": (22, 18, 15, 14, 12, 9, 6, 4),
"L": (5, 0, 0, 0, 5, 10, 30, 50)}
K = {"M": 12000, "S": 5000, "T": 6000, "D": 8000, "L": 4000}
NAMES = ("M", "S", "T", "D", "L")
def alpha(i, j):
num = Fr(0)
den = Fr(0)
for x, y in zip(P[i], P[j]):
num += x * y
den += x * x
return num / den
AL = {(i, j): alpha(i, j) for i in NAMES for j in NAMES}
def solve(sub):
m = [[AL[(i, j)] for j in sub] + [Fr(K[i])] for i in sub]
k = len(sub)
for c in range(k):
p = next(r for r in range(c, k) if m[r][c] != 0)
m[c], m[p] = m[p], m[c]
for r in range(k):
if r != c:
f = m[r][c] / m[c][c]
m[r] = [x - f * y for x, y in zip(m[r], m[c])]
return {sub[i]: m[i][k] / m[i][i] for i in range(k)}
rows = []
for size in range(1, 6):
for sub in combinations(NAMES, size):
n = solve(list(sub))
if all(v > 0 for v in n.values()):
closed = True
for i in NAMES:
if i not in sub:
press = Fr(0)
for j in sub:
press += AL[(i, j)] * n[j]
closed = closed and K[i] - press <= 0
tot = Fr(0)
for v in n.values():
tot += v
rows.append((tot, "".join(sub), closed, n))
rows.sort(key=lambda r: (-r[0], r[1]))
print("feasible subsets: %d of 31" % len(rows))
print("subset total closed traffic per open site")
for tot, name, closed, n in rows[:8]:
print("%-6s %6.0f %-6s %s" % (name, float(tot), "yes" if closed
else "no", " ".join(
"%s=%.0f" % (k, float(v)) for k, v in sorted(n.items()))))
full = [r for r in rows if r[1] == "MSTDL"]
print("all five held open: %s" % ("feasible" if full else "infeasible"))
cap = 0
for i in NAMES:
cap += K[i]
print("sum of capacities if no overlap: %d" % cap)
# Check by simulation: all five open with 1000 visitors each, r = 0.3
# per month, Euler steps of 0.05 month for 600 months.
af = {key: float(v) for key, v in AL.items()}
n = {i: 1000.0 for i in NAMES}
for _ in range(12000):
g = {i: 0.3 * n[i] * (1 - math.fsum(af[(i, j)] * n[j]
for j in NAMES) / K[i])
for i in NAMES}
n = {i: n[i] + 0.05 * g[i] for i in NAMES}
print("simulated month 600: " + " ".join(
"%s=%.0f" % (i, n[i]) for i in NAMES))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《洛特卡沃尔泰拉竞争相图、多子站多账号矩阵内耗与共生壁垒判定母典》第 4 章「从两站到多站:流量天花板、最优站数与五站矩阵的裁剪」(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-251 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026251,
author = {Fan, Yafu},
title = {Lotka-Volterra Competition Phase Portraits: Cannibalisation within Multi-Site, Multi-Account Matrices versus Symbiotic Barriers},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}