What does Revenue Management and Slot Packing for Space Rentals conclude, in numbers, about “Nested Protection Levels: EMSR-a, EMSR-b and the Exact Optimum”?
Answer: Chapter 2 compares EMSR-a, EMSR-b, the exact optimum and fixed quotas for four fares: EMSR-b reaches 0.99778 of the exact optimum.
Scope of the numbers: Chapter 1 computes exact expectations for an illustrative month of 20 studio-days: protecting 10 days lifts monthly revenue from 47323.9 yuan to 57432.1 yuan, 21.36 % more, though simulation shows that protection earns less in 0.1900 of months. Marked boxes separate published evidence, runnable Python calculations (18 programs), classical sources and the author's hypotheses.
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: 0.99778
| policy | y1 | y2 | y3 | E[revenue] | share of optimum |
| optimal (DP) | 2 | 7 | 15 | 60086.2 | 1.00000 |
| EMSR-b | 2 | 7 | 14 | 59952.9 | 0.99778 |
| EMSR-a | 2 | 7 | 14 | 59952.9 | 0.99778 |
| FCFS | 0 | 0 | 0 | 52006.3 | 0.86553 |
| optimal protection levels re-evaluated: 60086.2 | |||||
| marginal value of the 15th day for classes 1..3: 1948.0 yuan |
Show the full program
# Chapter 2. p22: exact optimum versus EMSR-a, EMSR-b and FCFS.
# Classes of p21 arrive low-before-high (4, 3, 2, 1). Demand is the
# normal of p21 discretised to integers (mass at k = Phi(k+.5)-Phi(k-.5),
# all mass below 0.5 at 0). W_j(x) = expected revenue of classes 1..j
# with x days left; the optimum maximises over how many of class j to
# sell (Brumelle and McGill 1993: optimal nested protection levels
# exist). Every policy is evaluated with the same exact recursion.
import math
from statistics import NormalDist
C, K = 20, 40
FARE = [5600, 4200, 3000, 1800]
MU = [3.0, 5.0, 6.0, 10.0]
SD = [1.5, 2.0, 2.5, 3.5]
def pmf(mu, sd):
nd = NormalDist(mu, sd)
p = [nd.cdf(0.5)]
p += [nd.cdf(k + 0.5) - nd.cdf(k - 0.5) for k in range(1, K)]
p.append(1.0 - nd.cdf(K - 0.5))
return p
P = [pmf(m, s) for m, s in zip(MU, SD)]
def solve(prot=None):
# prot[j] = 第 j+1 类到来时为更高类保留的棚日;None 表示求最优
w = [0.0] * (C + 1)
marg = []
for j in range(4):
f, new = FARE[j], []
for x in range(C + 1):
terms = []
for d, p in enumerate(P[j]):
top = min(d, x)
if prot is None:
best = max(f * u + w[x - u] for u in range(top + 1))
else:
u = min(d, max(0, x - prot[j]))
best = f * u + w[x - u]
terms.append(p * best)
new.append(math.fsum(terms))
w = new
marg.append([w[x] - w[x - 1] for x in range(1, C + 1)])
return w[C], marg
def emsr(kind):
y = [0]
for j in range(1, 4):
f = FARE[j]
if kind == "a":
v = math.fsum(NormalDist(MU[k], SD[k]).inv_cdf(1 - f / FARE[k])
for k in range(j))
else:
s = math.fsum(MU[:j])
sd = math.sqrt(math.fsum(x * x for x in SD[:j]))
fb = math.fsum(FARE[k] * MU[k] for k in range(j)) / s
v = NormalDist(s, sd).inv_cdf(1 - f / fb)
y.append(max(0, round(v)))
return y
opt, marg = solve()
ystar = [0]
for j in range(1, 4):
m = marg[j - 1]
ystar.append(len([x for x in range(C) if m[x] > FARE[j]]))
rows = [("optimal (DP)", ystar), ("EMSR-b", emsr("b")),
("EMSR-a", emsr("a")), ("FCFS", [0, 0, 0, 0])]
print("policy y1 y2 y3 E[revenue] share of optimum")
for name, y in rows:
val = opt if name.startswith("opt") else solve(y)[0]
print("%-13s %3d %3d %3d %9.1f %.5f"
% (name, y[1], y[2], y[3], val, val / opt))
check = solve(ystar)[0]
print("optimal protection levels re-evaluated: %.1f" % check)
print("marginal value of the 15th day for classes 1..3: %.1f yuan"
% marg[2][14])Single source and citation
Fan, Y. (2026). Dynamic Revenue Management and Slot Packing for Space Rentals: Nested Protection Levels and the Elimination of Fragmented Idle Time (Version 1.0.0), Chapter 2: Nested Protection Levels: EMSR-a, EMSR-b and the Exact Optimum. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-282 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026282,
author = {Fan, Yafu},
title = {Dynamic Revenue Management and Slot Packing for Space Rentals: Nested Protection Levels and the Elimination of Fragmented Idle Time},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · The Cost of First Come, First Served: Littlewood's Two-Class Rule and the Full-Day Orders It Saves
- Chapter 3 · Bid Prices: How the Value of a Studio-Day Changes over the Booking Run
- Chapter 4 · Slot Packing: First-Fit Decreasing and Fragmented Idle Time
- Chapter 5 · Bookings at Fixed Times: Interval Partitioning and Weighted Interval Scheduling
- Chapter 6 · Price Ladders Posted in Full: Dynamic Pricing, a Promise Not to Cut, and Real Scarcity
- Chapter 7 · Classics of East and West: Season and Soil in the Qimin Yaoshu, and Ricardo's Differential Rent
- Chapter 8 · Closing the Loop: The Studio Price Table and a Pre-Registered Test