关于「响应曲线的四种形状与火力收敛系数」,《存量资产七日点火与火力收敛》给出的定量结论是什么?

结论:第二章定义火力收敛系数,示意书目里只推前三件仅得最优的 0.805。

数据口径:书中参数为示意,价格用无币种示意单位;全部计算由书中程序生成;本书是非个人化经营建议,不承诺任何收入。 第二章定义火力收敛系数,示意书目里只推前三件仅得最优的 0.805。

English version

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: 0.805

A. best plan per response shape (X120, 30 candidates)
shape value active HHI1/HHI
concave144.81 13 0.131 7.61
fixed cost96.85 4 0.259 3.87
switching124.25 5 0.219 4.56
S-shape153.88 6 0.170 5.88
B. forced plans: top 3 only vs all 30 equal
shape top3 all30_equal top3/best
concave116.51 91.50 0.805
fixed cost92.51 -148.50 0.955
switching113.99 0.00 0.917
S-shape121.59 13.96 0.790
C. S-shape check, unrestricted DP over all 30153.88
查看完整验算程序
# Chapter 2. p22: concentration (Herfindahl) against seven-day cash
# 30 candidate assets with ceilings M_i = 60 * i^-0.7 (illustrative), a
# weekly firepower budget X = 120 units, the four shapes of p21.
# For k active assets (the top k, since shapes are equal) we find the
# best split: water-filling in 0.25-unit steps for the concave family,
# a knapsack-style DP in 1-unit steps for the S shape. HHI = sum of
# squared firepower shares; 1/HHI is the "effective number" of assets.
import heapq
import math

X, S, F, H, K, DELTA = 120.0, 15.0, 8.0, 3.0, 12.0, 0.05
MS = [60.0 * (i + 1) ** -0.7 for i in range(30)]


def conc(m, x):
    return m * (1.0 - math.exp(-x / S))


def sfun(m, x):
    return m * x ** H / (K ** H + x ** H) if x > 0 else 0.0


def waterfill(k, budget, step=0.25):
    cnt = [0] * k
    x = [0.0] * k
    heap = [(-(conc(MS[i], step) - conc(MS[i], 0.0)), i) for i in range(k)]
    heapq.heapify(heap)
    for _ in range(int(round(budget / step))):
        _, i = heapq.heappop(heap)
        cnt[i] += 1
        x[i] = cnt[i] * step
        gain = conc(MS[i], x[i] + step) - conc(MS[i], x[i])
        heapq.heappush(heap, (-gain, i))
    return x


def hhi(x):
    tot = math.fsum(x)
    return math.fsum((v / tot) ** 2 for v in x if v > 0)


def best_concave(fixed, delta, kmax=30):
    best = None
    for k in range(1, kmax + 1):
        budget = X * max(0.0, 1.0 - delta * (k - 1))
        x = waterfill(k, budget)
        val = math.fsum(conc(MS[i], x[i]) for i in range(k)) - fixed * k
        if best is None or val > best[0] + 1e-12:
            best = (val, k, hhi(x))
    return best


def best_s(kmax=30):
    nb = int(X)
    dp = [0.0] * (nb + 1)
    for i in range(kmax):
        new = dp[:]
        for b in range(nb + 1):
            for x in range(1, b + 1):
                v = dp[b - x] + sfun(MS[i], x)
                if v > new[b]:
                    new[b] = v
        dp = new
    return dp[nb]


def s_alloc_topk(k):
    # exact DP restricted to the top k, returns value and allocation
    nb = int(X)
    dp = [0.0] * (nb + 1)
    choice = []
    for i in range(k):
        new, ch = dp[:], [0] * (nb + 1)
        for b in range(nb + 1):
            for x in range(1, b + 1):
                v = dp[b - x] + sfun(MS[i], x)
                if v > new[b]:
                    new[b], ch[b] = v, x
        dp = new
        choice.append(ch)
    x, b = [0] * k, nb
    for i in range(k - 1, -1, -1):
        x[i] = choice[i][b]
        b -= x[i]
    return dp[nb], x


print("A. best plan per response shape (X = 120, 30 candidates)")
print("shape        value   active  HHI    1/HHI")
for name, fx, dl in (("concave", 0.0, 0.0), ("fixed cost", F, 0.0),
                     ("switching", 0.0, DELTA)):
    v, k, h = best_concave(fx, dl)
    print("{:<11} {:6.2f}  {:6d}  {:.3f}  {:5.2f}".format(name, v, k, h,
                                                         1.0 / h))
sv, sx = s_alloc_topk(30)
act = [v for v in sx if v > 0]
hs = hhi([float(v) for v in act])
print("{:<11} {:6.2f}  {:6d}  {:.3f}  {:5.2f}".format(
    "S-shape", sv, len(act), hs, 1.0 / hs))
print("B. forced plans: top 3 only vs all 30 equal")
print("shape        top3    all30_equal  top3/best")
for name, fx, dl in (("concave", 0.0, 0.0), ("fixed cost", F, 0.0),
                     ("switching", 0.0, DELTA)):
    x3 = waterfill(3, X * (1.0 - dl * 2))
    t3 = math.fsum(conc(MS[i], x3[i]) for i in range(3)) - fx * 3
    eq = X * max(0.0, 1.0 - dl * 29) / 30.0
    a30 = math.fsum(conc(m, eq) for m in MS) - fx * 30
    print("{:<11} {:6.2f}  {:11.2f}  {:9.3f}".format(
        name, t3, a30, t3 / best_concave(fx, dl)[0]))
s3, _ = s_alloc_topk(3)
s30 = math.fsum(sfun(m, X / 30.0) for m in MS)
print("{:<11} {:6.2f}  {:11.2f}  {:9.3f}".format("S-shape", s3, s30,
                                                 s3 / sv))
print("C. S-shape check, unrestricted DP over all 30 = {:.2f}".format(
    best_s()))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《重装数字资产库首周现金流点火、火力收敛系数与饱和变现突击母典》第 2 章「响应曲线的四种形状与火力收敛系数」(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-415 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026415,
  author    = {Fan, Yafu},
  title     = {Seven-Day Cashflow Ignition from Large Digital Asset Catalogs: Firepower Concentration Ratios and Saturation Monetisation Sprints},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论