ISyE 3232C

Z. She, YL. Chang

Stochastic Manufacturing and Service Systems

Fall 2015

Homework 3 - Solutions
September 6, 2015
1. A camera store specializes in a particular popular and fancy camera. Assume that these cameras
become obsolete at the end of the month. They guarantee that if they are out of stock, they will
special-order the camera and promise delivery the next day. In fact, what the store does is to purchase
the camera from an out of state retailer and have it delivered through an express service. Thus, when
the store is out of stock, they actually lose the sales price of the camera and the shipping charge, but
they maintain their good reputation. The retail price of the camera is $500, and the special delivery
charge adds another $50 to the cost for each camera. At the end of each month, there is an inventory
holding cost of $50 for each camera in stock (for doing inventory etc). Wholesale cost for the store to
purchase the cameras is $300 each. (Assume that the order can only be made at the beginning of the
month.)
(a) Identify the understock cost per unit (Cu ) and overstock cost per unit (Co ).
Hint: Pay attention to understock cost. when the store is out of stock, they actually lose the
sales price of the camera and the shipping charge.
Data: unit retail price is 500, unit delivery charge is 50; consequently Cp = 500 and Pu = 50, unit
variable cost is Cv = 300, unit inventory holding cost is h = 50.
Understock cost: Cu = Cp + Pu = 500 + 50 = $550
Overstock cost: Co = h = $50
(b) Assume that the demand has a discrete uniform distribution from 25 to 30 cameras a month
(inclusive). If 26 cameras are ordered at the beginning of a month, what are the expected overstock
cost and the expected understock or shortage cost? What is the expected total cost?
The distribution of demand is:
D=d
P {D = d}

25

26

27

28

29

30

1
6

1
6

1
6

1
6

1
6

1
6

Expected understock cost:

Cu E[max{D 26, 0}] = Cu

30
X

(max{x 26, 0}P {D = x}

x=25

= 550(0 1/6 + 0 1/6 + 1 1/6 + 2 1/6 + 3 1/6 + 4 1/6)

= 5500/6 $916.67
Expected overstock cost:
Co E[max{26 D, 0}] = Co

30
X

(max{26 x, 0}P {D = x}

x=25

= 50(1 1/6 + 0 1/6 + 0 1/6 + 0 1/6 + 0 1/6 + 0 1/6)

= 50/6 $8.33
The expected total cost:
E[understock cost]+E[overstock cost]+cv q = 5500/6 + 300 26 = 925 + 7800 8725
(c) What is optimal number of cameras to order to minimize the expected total cost?
250
5
u Cv
y is the smallest number such that F (y) C
y = 27
Cu +Co = 600 = 12 0.417

(d) What is the expected total cost that corresponds to the optimal order quantity?
Expected understock cost:
Cu E[max{D 27, 0}] = Cu

30
X

(max{x 27, 0}P {D = x}

x=25

= 550(0 1/6 + 0 1/6 + 0 1/6 + 1 1/6 + 2 1/6 + 3 1/6)

= $550
Expected overstock cost:
30
X

Co E[max{27 D, 0}] = Co

(max{27 x, 0}P {D = x}

x=25

= 50(2 1/6 + 1 1/6 + 0 1/6 + 0 1/6 + 0 1/6 + 0 1/6)

= $25
The expected total cost:
E[understock cost]+E[overstock cost]+cv q = 575 + 300 27 = 575 + 8100 = $8675
(e) Assume that the demand can be approximated by a normal distribution with mean 800 and
standard deviation 100 cameras a month. What is the optimal number of cameras to order to
minimize the expected total cost?
Cu Cv
250
Cu +Co = 600
y 800
P (Z 100 )

Now y solves F (y ) = P (D y ) =
P ( D800
100

y 800
100 )

5
12

0.417

=
where Z is the standard normal random variable.
Note that
Also, P (Z 0.21) 0.417 from the standard normal table.
Thus, y = 0.21 100 + 800 = 779; so order 779 cameras.
2. Every four years, Blockbusters Publishers revises its textbooks. It has been three years since the bestselling book The Joy of OR has been revised. At present, 2000 copies of the book are in stock, and
Blockbusters must determine how many copies of the book should be printed for the next year. Each
copy of the book sold during next year will bring the publisher $45 in revenues. Any copies of that
are left at the end of the year cost $5 dollars per copy to get rid of. The cost of printing the book is
$50,000 (fixed one-time cost) plus $20 per book printed.
(a) Suppose that the estimated demand for the book is discrete with P r{D = 5000} = 0.20, P r{D =
6000} = 0.30, P r{D = 7000} = 0.30, and P r{D = 8000} = 0.20. How many copies of the book
should be printed?
When we have a fix cost of Cf = 50000, Cv = 20, Cu = 45, Co = 5 and x = 2000. The optimal
order quantity y satisfies:
F (y )

25
Cu Cv
45 20
=
=
= 0.5
Cu + Co
45 + 5
50

D=d
P {D = d}

5000
0.20

6000
0.30

7000
0.30

8000
0.20

Then y = 6000. Now, we check the expected costs for ordering 6000-2000=4000 and not ordering.

Scenario 1: Not to order

E[Cost] = Cu E[max{D 2000, 0}] + Co E[max{2000 D, 0}]
= 45E[D 2000] + 5 0

(as D > 2000)

= 45(E[D] 2000)
8000
X

= 45(

xP {D = x} 2000)

x=5000

= 45(5000 0.20 + 6000 0.30 + 7000 0.30 + 8000 0.20 2000)

= 202, 500
Scenario 2: To order 6000-2000=4000
E[Cost] = Cf + Cv (6000 2000) + Cu E[max{D 6000, 0}] + Co E[max{6000 D, 0}]
= 50000 + 20 4000 + 45 700 + 5 200

(calculated E below)

= 162, 500
8000
X

E[max{D 6000, 0}] =

(max{x 6000, 0}P {D = x}

x=5000

= (0 0.2 + 0 0.3 + 1000 0.3 + 2000 0.2)

= $700
8000
X

E[max{6000 D, 0}] =

(max{6000 x, 0}P {D = x}

x=5000

= (1000 0.2 + 0 0.3 + 0 0.3 + 0 0.2)

= $200
As the expected cost of ordering 4000 is less than that of not ordering, they should order 4000
books.
(b) How would the answer change if 4000 copies were already in stock?
Similar to part (a), we will calculate the expected costs for 2 scenarios. This time, we need to consider
the cost for ordering 6000 4000 = 2000 books.
Scenario 1: Not to order
E[Cost] = Cu E[max{D 4000, 0}] + Co E[max{4000 D, 0}]
= 45E[D 4000] + 5 0

(as D > 4000)

= 45(E[D] 4000)
= 45(

8000
X

xP {D = x} 4000)

x=5000

= 45(5000 0.20 + 6000 0.30 + 7000 0.30 + 8000 0.20 4000)

= 112, 500
Scenario 2: To order
E[Cost] = Cf + Cv (6000 4000) + Cu E[max{D 6000, 0}] + Co E[max{6000 D, 0}]
= 50000 + 20 2000 + 45 700 + 5 200

(calculated E in part (a))

= 122, 500
As the expected cost of ordering 2000 is more than that of not ordering, they should order nothing.
3

3. (a) Define the variables:

cf = 8000 (fixed cost)
cv = 50 (unit production cost)
cu = 100 (understock cost)
co = 15 (overstock cost)
The optimal order-up-to quantity is the smallest y such that:
FD (y)

100 50
10
cu cv
=
=
0.435
cu + co
100 + 15
23

Below is the CDF of the demand:

0.3 300

0.5 500
FD (y) =

0.9 700

1 900
Therefore, the optimal order-up-to quantity y is 500, as FD (y) 0.435.
Now, lets determine the optimal ordering policy for arbitrary initial inventory level x.
If we decide to order, we will order up to 500 liters of solvent. Suppose we have x inventory
currently. Evaluate the two scenarios:
Ordering: If we decide to order, we will place an order for (500 x) liters, as we already have
x liters in the inventory. We need to pay the fixed ordering cost and the variable purchasing cost
for the order. We also have to expect the costs from shortage and left-over situations. Note that
we will start the next month with 500 liters because we will order up to 500 liters. Hence,
E[Cost] = cf + cv (500 x) + cu E[(D 500)+ ] + co E[(500 D)+ ]
= 8000 + 50 (500 x) + 100 120 + 15 60

(by (1) and (2))

= 45900 50x

E[max{D 500, 0}] =

900
X

(max{x 500, 0}P {D = x}

x=300

= (0 0.3 + 0 0.2 + 200 0.4 + 400 0.1) = $120

E[max{500 D, 0}] =

900
X

(1)

(max{500 x, 0}P {D = x}

x=300

= 200 0.3 + 0 0.4 + 0 0.2 + 0 0.1) = $60

(2)

Not ordering If we decide not to order, we dont have to pay the ordering cost. We do have
to expect shortage and left-over costs. Now assume that x < 300. (We will check if our will be
indeed in this range, [0,300].) Note that we will start the next month with x liters because we
will not order in this scenario.
As we assume that x < 300, E[(D x)+ ] = E[D] x and E[(x D)+ ] = 0.
Given that E[D] = 300 0.4 + 500 0.3 + 700 0.2 + 900 0.1 = 560:
E[Cost] = cu E[(D x)+ ] + co E[(x D)+ ]
= 100 (560 x) + 15 0
= 56000 100x
4

When we have exactly x liters, we should be indifferent between ordering and not ordering. Hence,
56000 100x = 45900 50x

x = 202

We found that x is in [0,300] range, so our assumption holds. Below is the graph of the relationship between the cost and x . As you realize, the expected cost of ordering is less than the
expected cost of not ordering when x < 202. It means that if our inventory is below 202 liters,
we will order up to 500 liters. Similarly, if our inventory is above 202 liters, we will not order for
the next month.
E[Cost]
104
6

4
ordering

2
not ordering

100

200
300
x = 202

400

500

(b) As found in question 1, the optimal order-up-to quantity is 500. Also, we found that, we will
order if the initial inventory is less than 202.
Thus, for x = 0, we should order 500 liters. For other cases, we shouldnt order.

(c) For x = 0, we will order 500 liters, so the expected cost of ordering:

E[Cost] = cf + cv (500 x) + cu E[(D 500)+ ] + co E[(500 D)+ ]

= 45900 50x
= $45900
For x = 700, we will not order, so the expected cost of not ordering:

E[Cost] = cu E[(D x)+ ] + co E[(x D)+ ]

= 100E[(D 700)+ ] + 15E[(700 D)+ ]

(by (3) and (4))

= 100 20 + 15 200 = $5000

E[max{D 700, 0}] =

900
X

(max{x 700, 0}P {D = x}

x=300

= (0 0.3 + 0 0.2 + 0 0.4 + 200 0.1) = $20

E[max{700 D, 0}] =

900
X

(3)

(max{700 x, 0}P {D = x}

x=300

= 400 0.3 + 200 0.2 + 0 0.4 + 0 0.1) = $160

(4)