Sie sind auf Seite 1von 12

# Inventory Practice Problem Solutions

## EOQ and cycle inventory

Problem 1
Avitron, a large aviation maintenance and service firm, sources a GPS triangulation circuit board
from a Taiwanese electronics assembler. Avitron installs about 1000 of these boards a year into
its clients airplanes at a selling price of \$425 each. Each time Avitron orders the boards from its
supplier, it incurs 2000 dollars in administrative and shipping costs. These boards cost 150
dollars each, but the supplier offers a 5% discount from this price if the order quantity is 1000
units or more (the discount applies to all of the units in the order). Avitron estimates its annual
inventory holding cost as 35% of an items cost.
What should Avitrons ordering quantity be? How many orders should Avitron place per year?
Solution
R
S
h
C
H
Q
Annual
Order cost
Inv. cost
Purchase cost
total cost

1000
2000
0.35
150
52.5
276.026224

7,246
7,246
150,000
164,491

1000
2000
0.35
142.5
49.875
283.19693

7,062
7,062
142,500
156,624

1000
2000
0.35
142.5
49.875
1000

2,000
24,938
142,500
169,438

We need to find the order quantity that minimizes the sum of inventory, ordering, and
purchase costs. Recall that we solve this problem using the Economic Order Quanitity
formula
Q=

2SR
H

## Using this order quantity, the annual inventory cost is

ordering cost is

QH
, the annual
2

RS
, and the annual purchase cost is RC.
Q

## The calculations are given in the above spreadsheet. In that spreadsheet:

(i ) Column 1 gives costs w/out the discount using the undiscounted EOQ.
(ii ) Column 2 gives cost w/ discount using the discounted EOQ. Note that
this order quantity of 283 will not give the discount, so this option is
not feasible.
(iii ) Column 3 gives the cost w/ discount using the smallest possible order
quantity under which the discount will be given.
Thus, only columns 1 and 3 are feasible, and column 1 has the smaller total cost.
1000
Thus, order 276 units at a time, and order
= 3.6 times per year.
276

Problem 2
A retail store (such as Kmart) stocks portable phones. The following data are available:

a) What is the EOQ? What is the total annual holding and ordering costs?
b) The purchasing manager believes that the ordering cost for this item could be reduced to
\$20/order if an EDI (Electronic Data Interchange) link were established so that orders could
be transmitted electronically. Establishing this would require an initial investment so the
manager is interested in the savings per year the store could expect once EDI was
established. Compute these savings.
c) Suppose the purchasing manager misestimates the cost reduction due to EDI. The actual cost
is \$25 but he bases his quantity on his estimate of \$20. What does this error cost on an
annual basis? To do this suppose QS is the correct quantity for ordering cost of S and TCS(Q)
is the cost of using Q when S is the ordering cost. Compare TC25(Q20) to TC25(Q25). The
error in the cost estimate is 20%. What is the percentage error in the annual costs?
d) Suppose for reasons specific to shipping phones the vendor wants to ship in multiples of 60,
e.g., 60, 120, 180, etc. Of these choices which is the one with the least total annual cost of
holding inventory and ordering costs? What is the percentage change in Q from the optimal
you found in (a) and what is the percentage change in cost from that found in (a)?

Solution
The items are the portable phones. There is a selling process (task and resource), which sells
phones at a steady rate of 180 phones/month or 180*12 phones/year.
a)

b)

## So the change in Q is 13 units or a 12% change for a 20% change in S.

So the change in total annual ordering and holding costs, TC, is under 1% (\$6/\$900), when a
Q that is 12% different from the optimal, and this order quantity was computed with a
ordering cost that was off by 20%. Thus, the impact of small errors in estimates of the
parameters for this model is small.
d) The optimal quantity in part (a) was 152 units but the limitation here says we can either
purchase 120 or 180 at a time. We check the value of the annual ordering and holding costs
for these two values.

The Q = 180 is slightly better than Q = 120. Also note that again the 20% change in Q results
in only a 12% change in the costs.

## EOQ in the context of demand variability and lead time

Problem 1
The I-Scream firm purchases vanilla ice cream in high volume, mixes in various flavoring
ingredients and distributes the flavored ice cream to nearby restaurants and stores. Each gallon
of ice cream, which is purchased from a wholesaler for \$8, generates about \$20 in operating
margin. Daily demand for this ice cream is normally distributed with a mean of 400 gallons and
a standard deviation of 120. The wholesaler charges a \$90 delivery charge independent of order
size. The opportunity cost of capital to I-Scream is estimated to be 25% per year. Assume 360
working days in the year.
What is the optimal size of each order?

Assume that the delivery lead-time is 9 days. The owners would like ensure that the vanilla ice
cream inventory that is on hand when they place an order (using the order size calculated above)
will be sufficient to meet demand during this lead time with 95 percent probability.
How much safety stock do they require?

Suppose that I-Scream places an order as per your answers above. If the delivery is delayed by
two days, what is the probability that I-Scream will run out of ice cream before the delivery
actually arrives?
Solution:
What is the optimal size of each order?
3600 quarts from the EOQ formula. This is not affected by the demand uncertainty.
Assume that the delivery lead-time is 9 days. The owners would like ensure that the vanilla ice
cream inventory that is on hand when they place an order (using the order size calculated above)
will be sufficient to meet demand during this lead time with 95 percent probability.
How much safety stock do they require?
594 quarts. Standard deviation of lead time demand,

L r 3 120 360 .

## The z-value for 95% service = 1.65. Thus Is = z x = 594 quarts.

Suppose that I-Scream places an order as per your answers above. If the delivery is delayed by
two days, what is the probability that I-Scream will run out of ice cream before the delivery
actually arrives?
Lets first assume that the ROP is adjusted to account for the shift in mean lead time demand.
But the safety stock stays at 594 quarts. Standard deviation of lead time demand,

## L r 3.31 120 398 .

594 = z x z x
which corresponds to a service level of 93.2%
Now suppose ROP cannot adjust quickly enough to accommodate the change:
Old ROP = Mean Lead Time Demand + Safety Stock (using 9 days lead time) and this must
equal
New ROP = Mean Lead Time Demand + Safety Stock (using 11 days lead time)
Equating the 2, we get:
9(400) + 594 = 11(400) + z x
052 which corresponds to a service level of 30.15% (dramatic drop!)
Problem 2
In the past Wegmans reordered potato chips by determining a order quantity from an EOQ and
setting the reorder level, ROP, so that the cycle service level was 99.9%. The cost to Wegmans
for a bag of potato chips was \$0.90. Wegmans management computed an annual unit holding
cost as 20% of its cost for the good. For one particular brand, Chipo, monthly demand was 2000
bags. The standard deviation of monthly demand was 300 bags. The distribution of demand in a
month is normal. The ordering and delivery costs were \$15. The delivery leadtime was 2 weeks.
Assume 4 weeks in a month.
a) In what quantities did Wegmans purchase Chipo potato chips.

b) At what inventory level did Wegmans reorder Chipo potato chips to meet its service level?
c) Currently, Chipo manages the inventory of potato chips inside the Wegmans store, relieving
Wegmans management from this task. It visits the store each week with a truckload of potato
chips and restocks the shelf. What would the target inventory level need to be to maintain the
cycle service level of 99.9% ?

Solution
The items are the bags of chips, the tasks are the purchasing of chips and the sale of chips to the
consumer. There are no explicit resources mentioned and they are not constraints in EOQ/ROP
problems.
a)

b) The demand is given in bags/month. The lead time is given as 2 weeks so we need to
convert. The mean lead-time demand is = (2000 bags/month)(1 month/4 weeks)(2
weeks) = 1000 bags. The standard deviation of lead time demand is (300 bags/month)
2/4 = 212.1
Thus, ROP = + z0.999 = 1000 + 212.1(3.1) = 1657 bags
c) Chipo is using an order-up-to model with time period = 1 week between deliveries. The
lead time is zero. Chipo restocks the shelves with enough bags of chips so that the
probability of running out in the week (L + p = p = 1 week) is 0.1% or the probability of
not running out is 99.9% before the truck returns next week.
Since the time period is 1 week now L+p = 500 and L+p = 1/4 (300) so
Target Inventory level: S = L+p + z0.999 L+p = 500 + (3.1)150 = 965 bags

## Optimal safety stock (The Newsvendor problem)

Problem 1
The buyer for Needless Markup, a famous high-end department store must decide on the quantity
of a high-priced womens handbag to procure in Italy for the following Christmas season. The
unit cost of the handbag to the store is \$28.50, and the bag will sell for \$150. Any handbags not
sold by the end of the season are purchased by a discount firm for \$20.00
As a consultant, you first construct a demand forecast. Your proprietary methods, in which you
are most confident, reveal that demand does not follow a Normal distribution. Rather, you
believe that the sales of these bags are equally likely to be anywhere from 51 to 250 units; that is,
any sales number between (and including) 51 and 250 has the same probability of occurring.
How many handbags do you recommend the firm order?
Solution
You know from class that the order quantity Q should be set such that the probability that
demand is Q is the critical ratio of 0.935.
We now need to relate this to the non-Normal distribution that is described. Each of the 200
possible demand values has the same probability of occurrence, which is 1/200 = 0.005.
This means that, for example, if we choose Q = 100 then the probability that demand in Q is
0.005 x (100 50) = 0.25.
Thus, we want Q such that 0.005 x (Q 50) = CR = 0.935.
Solving this for Q gives Q = 237.
Problem 2
Natalie Attired is a buyer for a mens fashion retail store. She will be ordering a new cloth
overcoat from Paris for the fall fashion season. Based on her experience, she expects to sell at
least 100 coats, and at most 400, but she feels that any number of sales in between is equally
likely. Therefore, she estimates that her sales are uniformly distributed between 100 and 400:

The total cost to the store is \$100 per coat, and the retail price is set at \$180. Any coats left over
at the end of the season would be sold at \$60 each.
a) How many coats should Natalie buy if she wants to maximize profits?
For the rest of this question, assume that Natalie buys the number of coats suggested in part (a):
b) What is the probability that the coats sell out? Do not sell out?
c) Given that the coats do not sell out, what is the expected number of coats sold?
d) What is the expected profit from sales of this coat?
Solution
a) The marginal profit for each coat is \$180-100 = 80, and the liquidation cost is \$100-60 = 40.
The order level to maximize revenue is the quantity Q which satisfies the relationship:
Prob (D Q*) = 80/(80 + 40) = 2/3
where D is the unknown demand for coats. Since demand is uniformly distributed between
100 and 400, Q* is 2/3 of the way between these limits: Q* = 300.
b) Prob (coats do not sell out) = Prob(D 300) 2/3.
Prob (coats do sell out) = Prob (D > 300) 1/3.
c) Given that coats do not sell out, actual sales are uniformly distributed between 100 and 300.
Therefore, the expected number sold is 200 coats.
d) A probability tree helps:

The total expected profit is: (1/3* \$24,000) + (2/3 * (\$16,000 - \$4,000)) = \$16,000

Problem 3
Hollywood Video (HV) rents videos to retail customers for \$4.00 per night. If a video is not
rented, the video sits on the shelf for the next night. The store itself rents each videotape from a
studio distributor. For example, HV will rent the movie Shrek for \$1.50 per night from the
distributor when it is available on video. HV estimates that nightly demand for Shrek will
average 90 customers with a standard deviation of 32. You may assume that nightly demand is
normally distributed and that each customer returns the video the next morning.
a) Find the optimal number of Shrek videos that HV should rent from the distributor.
b) Given your answer in part (a), what is the probability that the store will run out of Shrek on
any given night?
c) Now suppose that when a customer finds a movie out of stock, there is a 20% chance that the
customer will decide to take all her future business to a competing store, with an expected
loss of \$30 (that's the NPV of all future business after the night of the stockout). There is an
Now find the optimal number of Shrek videos that HV should rent from the distributor.
Solution
a) Underage cost Cu = \$4.00 - \$1.50 = \$2.50. Overage cost Co = \$1.50. Note that there is no
salvage value here the video just sits on the shelf.
Critical ratio = Cu /( Cu + Co) = 2.50 / 4.00 = 0.625.

We can use our Normal table, since we know that demand is distributed as a Normal random
variable: z-value = 0.32.
Q* = 90 + (0.32)(32) = 100.24 videos. Rounding up gives us 101.
b) The critical ratio is the probability that we have a sufficient number of videotapes in an
evening to satisfy demand. Therefore, the probability that we do run out = 1 0.625 = 0.375.
c) The expected goodwill loss from a stockout is (0.2) (\$30) = \$6. Therefore, Cu is now \$4.00
- \$1.50 + \$6.00 = \$8.50. The new critical ratio is 8.50 / 10.00 = 0.85.
z-value = 1.04.
Q* = 90 + (1.04)(32) = 123.28 videos, so we buy 124.
Note that Q must go up (and cannot go down) since the opportunity cost of a lost sale is
higher.
Problem 4
As a supplier of seasonal goods to L. L. Bean, you will offer Bean one of two options for
purchasing your product. L. L. Bean will sell the product for \$100.

Option 1:

You offer the item at \$65 and agree to take back unsold merchandise at \$55 each.
You discard the returned merchandise because it has no value to you.

Option 2:

You offer the item at \$60 but refuse to take returns. L. L. Bean must simply throw
out unsold goods. There is no salvage value to L. L. Bean.

The demand for the item is known to have a normal distribution with a mean of 200 units per
season with a standard deviation of 40 (per season). Assume the cost of a lost sale to L. L. Bean
is merely the lost profit margin on this single item, i.e., there are no cross item effects.

## a) How much would L. L. Bean buy if given option 1?

b) How much would L. L. Bean buy if given option 2?
c) Now, consider the suppliers problem. Should it offer option 1 or option 2 to L. L. Bean? To
decide, compute the (expected) supplier's profit assuming L. L. Bean will respond to each

option by buying the quantities computed in (a) and (b). The supplier's cost for an item is
\$20.
The table below gives the expected number sold by L. L. Bean by the end of the season for
various possible original purchase quantities.

Quantity
Bought
by LL Bean
100
110
120
130
140
150
160
170
180
190

Expected
Quantity
Sold
by LL Bean
99.92
109.83
119.66
129.35
138.83
147.98
156.67
164.75
172.09
178.55

Quantity
Bought
by LL Bean
200
210
220
230
240
250
260
270
280
290
300

Expected
Quantity
Sold
by LL Bean
184.04
188.55
192.09
194.75
196.67
197.98
198.83
199.35
199.66
199.83
199.92

Solution
a) We are given that D, the seasonal demand, is normally distributed with a mean of 200 and a
std. dev. of 40. For L.L. Bean, under option 1:

## Co \$65 \$55 \$10, Cu \$100 \$65 \$35.

We want to find Q * such that :
35
35 7
P{D Q*}

0.778
10 35 45 9
Hence, z 0.765, the 77.8th percentile of the normal distribution.
Then : Q* 200 40(.765) 230
b) Under option 2,

## Co \$60 \$0 \$60, Cu \$100 \$60 \$40

40
P{D Q*}
0.4
60 40
Q* 200 40( .255) 190.

Note that since L.L. Bean wishes to stock out more than half the time the quantity to
order is less than the median.
c) Under option 1, L.L Bean will purchase 230 units, and will expect to sell 194.75. Therefore,
the supplier expects 230-194.75 to be returned. To the supplier, \$65-\$20 are made for each
item sold, and \$55 are paid out for each returned. Therefore,
Profit of the supplier = 230(\$65 \$20) (230 194.75)(\$55) = \$8411.25.
Under option 2, L.L. Bean purchases 190 units and there are no returns:
Profit of the supplier = 190(\$60 \$20) \$0 = \$7600.
Therefore, the supplier should offer Option 1 - it risks the possibility of returns, but has a
higher per-unit and expected total profit.