Practice Set 8 Newsvendor Model

Question 1
University of Florida football programs are printed 1 week prior to each home game. Attendance averages
90,000 screaming and loyal Gators fans, of whom two-thirds usually buy the program, following a normal
distribution, for $4 each. Unsold programs are sent to a recycling center that pays only 10 cents per
program. The standard deviation is 5,000 programs, and the cost to print each program is $1.
a) What is the cost of underestimating demand for each program?
Cu = $4 – $1 = $3

b) What is the overage cost per program?

Co = $1 – $0.1 = $0.9

c) How many programs should be ordered per game?

P = Cu/ (Cu + Co) = 3/ (3 + 0.9) = 0.7692
F(0.73) = 0.7673, F(0.74) = 0.7704, z = 0.74
Hence, Q* = 𝜇 + z𝜎 = 90000*2/3 + 0.74*5000 = 63,700
(Only two-thirds of the fans would buy the program on average)

d) What is the stockout risk for this order size?

P = Prob(D ≤ Q*), which is the probability of no stock-out and demand can be fulfilled
Hence, stockout risk = 1 – 0.7692 = 23.08%

Question 2
Elite Couture, a high-end fashion goods store has to decide on the quantity of Luella Bartley handbags to
sell during the Christmas season. The unit cost of the handbag is $28.5 and the handbag sells for $150.
All handbags remaining unsold at the end of the season are purchased by a discounter for $20 each.
Further, there is a significant inventory holding cost incurred for each unsold bag, which is 40% of the
unit cost. Demand for bags is distributed normally with mean 150 and standard deviation 20. How many
bags should be purchased to maximize expected profit?
Cu = $150 – $28.5 = $121.5
Co = $28.50 – $20 + $28.5*0.40 = $8.5 + $11.40 = $19.9
P = Cu/(Cu + Co) = 121.5/(121.5 + 19.9) = 0.8593
F(1.07) = 0.8577, F(1.08) = 0.8599. Hence, z = 1.08
Q* = 𝜇 + z𝜎 = 150 + 1.08*20 = 171.6
Hence, order 172 handbags

Question 3
The local supermarket buys lettuce each day to ensure fresh produce. Each morning any lettuce that is left
from the previous day is sold to a dealer that resells it to farmers who use it to feed their animals. This
week the supermarket can buy fresh lettuce for $4.00 a box. The lettuce is sold for $10.00 a box and the
dealer that sold lettuce is willing to pay $1.50 for a box of unsold lettuce at the end of the day. Past
history says that demand for lettuce is normally distributed with a mean of 250 boxes with a standard
deviation of 34 boxes. How many boxes of lettuce should the supermarket purchase?
Cu = $10 – $4 = $6
Co = $4 – $1.5 = $2.5
P = Cu/(Cu + Co) = 6/(6 + 2.5) = 0.7059
Hence, z = 0.54
Q* = 𝜇 + z𝜎 = 250 + 0.54*34 = 268.36
Hence, purchase 268 boxes of lettuce
Question 4
Goop Inc needs to order a raw material to make a special polymer. The demand for the polymer is
forecasted to be normally distributed with a mean of 250 gallons and a standard deviation of 125 gallons.
Goop sells the polymer for $25 per gallon. Goop’s purchases raw material for $10 per gallon and Goop
must spend $5 per gallon to dispose all unused raw material due to government regulations. (One gallon
of raw material yields one gallon of polymer.) If demand is more than Goop can make, then Goop sells
only what they made and the rest of demand is lost.
a) How many gallons should Goop purchase to maximize its expected profit?
Cu = $25 – $10 = $15
Co = $10 + $5 = $15
P = Cu/(Cu + Co) = 15/(15 + 15) = 0.5
z=0
Q* = 𝜇 + z𝜎 = 250 + 0*125 = 250

b) Suppose Goop wants to ensure that there is a 92% probability that they will be able to satisfy the
customer’s entire demand. How many gallons of the raw material should they purchase?
Lookup 0.92 in the normal distribution table.
F(1.40) = 0.9192, F(1.41) = 0.9207, z = 1.405
Q* = 𝜇 + z𝜎 = 250 + 1.405*125 = 425.625 = 426

c) Suppose Goop purchases 150 gallons of raw material. What is the probability that they will run
out of raw material?
z = (150 – 250)/125 = -0.8.
From the Distribution Function Table, F(-0.8) = 0.2119
P = Prob(D ≤ Q*), which is the probability of no stock-out and demand can be fulfilled
Hence, the risk of stocking out is 1 – 0.2119 = 78.81%

Question 5
Tom owns a small firm that manufactures “Tom Sunglasses.” He has the opportunity to sell a particular
seasonal model to Land’s End. Tom offers Land’s End two purchasing options:
§ Option 1. Tom offers a price of $55 for each unit, but returns are no longer accepted. In this case,
Land’s End throws out unsold units at the end of the season.
§ Option 2. Tom offers to set his price at $65 and agrees to credit Land’s End $53 for each unit
Land’s End returns to Tom at the end of the season.
This season’s demand for this model will be normally distributed with mean of 200 and standard
deviation of 125. Land’s End will sell those sunglasses for $110 each.
a) How much would Land’s End buy if they choose option 1?
Cu = $110 – $55 = $55
Co = $55
P = Cu /(Cu + Co) = 55/(55 + 55) = 1/2
z=0
Q* = 𝜇 + z𝜎 = 200 + 0*125 = 200

b) How much would Land’s End buy if they choose option 2? What is the probability that Land’s
End will return sunglasses to Tom at the end of the season?
Cu = $110 – $65 = $45
Co = $65 – $53 = $12
P = Cu /(Co+ Cu) = 45/(45 + 12) = 0.7894
F(0.80) = 0.7881, F(0.81) = 0.7910. Hence, z = 0.805
Q* = 𝜇 + z𝜎 = 200 + 0.805*125 = 300.625
 Hence, purchase 301 sunglasses.
Prob(Land’s End will return sunglasses) = Prob(D ≤ Q*) = 0.7894

Question 6
Dan McClure owns a thriving independent bookstore in artsy New Hope, Pennsylvania. He must decide
how many copies to order of a new book, Power and Self-Destruction. The book’s retail price is $20 and
the wholesale prices is $12. The publisher will buy back the retailer’s leftover copies at full refund, but
McClure Books incurs $4 in shipping and handling costs for each book returned to the publisher. Dan
believes his demand forecast can be represented by a normal distribution with mean 200 and standard
deviation 80.
What would be the optimal order quantity for Dan?
Cu = $20 – $12 = $8
Co = $12 – ($12 – $4) = $4
(The salvage value is $8)

P = Cu/ (Cu + Co) = 8/ (8 + 4) = 0.667

F(0.43) = 0.6664, hence, z = 0.43
Q* = 200 + 0.43*80 = 234.4
Hence, order 234 copies

Question 7
A retailer sells a fashion product during a short sales season. Since there is a long production lead time,
the retailer needs to purchase the product in advance and no further inventory replenishment is allowed.
The product costs $70 per unit and the retail price is $100. For units that are not sold by the end of the
main sales season, the retailer can sell the leftover units at a discounted price $30 through clearance sales.
The demand is uncertain and the demand distribution is forecasted as follows.

Demand (units) 500 350 250 150

Probability 0.2 0.4 0.25 0.15

a) What is the underage cost?

Cu = $100 – $70 = $30

b) What is the overage cost?

Co = $70 – $30 = $40

c) How many units should the retailer purchase in order to maximize the expected profit?
P = 30/(30 + 40) = 0.4286
Next, compute the cumulative probability and find closest one that is greater than the critical
fractile, i.e. 0.8

Cumulative
Demand Probability
Probability
150 0.15 0.15
250 0.25 0.40
350 0.40 0.80
500 0.20 1.00

Q* is 350
Question 8
A product is priced to sell at $100 per unit, and its cost is constant at $70 per unit. Each unsold unit has a
salvage value of $20. Demand is expected to range between 30 and 40 units for the period. The demand
probabilities and the associated cumulative probability distribution (P) for this situation are shown below.

# of units demanded Probability Cumulative probability

35 0.10 0.10
36 0.15 0.25
37 0.25 0.50
38 0.25 0.75
39 0.15 0.90
40 0.10 1.00

How many units should be ordered?

Cu = $100 - $70 = $30
Co = $70 - $20 = $50
P = Cu/ (Cu + Co) = 30/ (30 + 50) = 0.375
From the table, we will look for the cumulative probability and find closest one that is greater than the
critical fractile, i.e. 0.5
Q* is 37 units.

To illustrate the problem, here is a full marginal analysis. Note that the cost is minimized when 37 units
are purchased.
# purchased
# demanded Probability 35 36 37 38 39 40
35 0.1 0 50 100 150 200 250
36 0.15 30 0 50 100 150 200
37 0.25 60 30 0 50 100 150
38 0.25 90 60 30 0 50 100
39 0.15 120 90 60 30 0 50
40 0.1 150 120 90 60 30 0
Expected total cost 75 53 43 53 83 125

Question 9
Sally’s Silk Screening produces specialty T-shirts that are primarily sold at special events. She is trying to
decide how many to produce for an upcoming event. During the event itself, which lasts one day, Sally
can sell T-shirts for $20 a piece. However, when the event ends, any unsold T-shirts are sold for $4 a
piece. It costs Sally $8 to make a specialty T-shirt. Using Sally’s estimate of demand that follows, how
many T-shirts should she produce for the upcoming event?

Demand Probability
300 0.05
400 0.10
500 0.40
600 0.30
700 0.10
800 0.05
Cu = $20 - $8 = $12
Co = $8 - $4 = $4
P = Cu/ (Cu + Co) = 12/ (12 + 4) = 0.75
Next, compute the cumulative probability and find closest one that is greater than the critical fractile, i.e.
0.85
Demand Probability Cumulative
probability
300 0.05 0.05
400 0.10 0.15
500 0.40 0.55
600 0.30 0.85
700 0.10 0.95
800 0.05 1.00

Q* is 600 units.

Question 10
Famous Albert prides himself on being the Cookie King of the West. Small, freshly baked cookies are the
specialty of his shop. Famous Albert has asked for help to determine the number of cookies he should
make each day. From analysis of past demand he estimates demand for cookies as the following.

Demand Probability of demand

1,800 0.05
2,000 0.10
2,200 0.20
2,400 0.30
2,600 0.20
2,800 0.10
3,000 0.05

Each dozen sells for $0.69 and costs $0.49, which includes handling and transportation. Cookies that are
not sold at the end of the day are reduced to $0.29 and sold the following day as day-old merchandise.
What is the optimal number of cookies to make?
Cu = $0.69 – $0.49 = $0.2
Co = $0.49 – $0.29 = $0.2
P = Cu/ (Cu + Co) = 0.2/ (0.2 + 0.2) = 0.5
Next, compute the cumulative probability and find closest one that is greater than the critical fractile, i.e.
0.65
Demand Probability of demand Cumulative
probability
1,800 0.05 0.05
2,000 0.10 0.15
2,200 0.20 0.35
2,400 0.30 0.65
2,600 0.20 0.85
2,800 0.10 0.95
3,000 0.05 1.00

Q* is 2400 units