Operations Management (OPMG),

Sample End-term exam

Term 3, 2021-22

Indian School of Business

Basic Information

There are 25 questions in the exam (either qualitative or quantitative). All questions carry
equal weight.

Choose the best answer by marking one of the options amongst those given. If none of the
choices are an exact match to your answer, pick the one that is closest. Only clearly marked
choices on the exam will be evaluated.

Please submit your rough work at the end of the exam using the designated link. Your rough
work will be used only for cross-verification and will NOT be graded.

This exam is an open notes exam. Open notes include all material including handouts,
documents, course pack and excel sheets uploaded on LMS and other notes you may have
created. You can use Excel or calculator for your calculations.

You cannot use any other sources on the worldwide web or any standard textbook including
the recommended textbook.

The exam is completely self-explanatory. The instructors and the academic associates will not
be available to clarify doubts.

You are expected to abide by the ISB honor code while taking this examination.

1
Q1. A shipment of laptops takes three months on average to arrive from Shanghai to San
Francisco. After arrival, the laptops sit in the warehouse in San Francisco for one month
on average. The optimal order size is 500,000 laptops. The demand for laptops is 1.2
million per year. What is the average number of laptops in transit on a ship anytime during
the year? (Find the closest answer)
a. 250,000
b. 300,000
c. 400,000
d. 500,000

Solution: The delivery lead time L = 3 months, average demand rate R = 1.2 million / year.
Using Little’s Law the average inventory in transit = R x L = 1.2 million / year x 3 months =
100,000 laptops / month x 3 months = 300,000.

Notice that the average number of laptops in transit on the ship during the year does not
depend on the order size. Also, note that the time spent in the warehouse is irrelevant to
calculate inventory on a ship.

Q2. If all units in a sample fall within the product specification limits, the process must be in
control. Do you agree?
a. Yes, since all the units in the sample are acceptable.
b. Yes, since there is a very small probability of producing a defect.
c. No, since control limits are always within the specification limits.
d. No, since control limits do not have any direct relation with specification limits.

Solution: The correct answer is No, since control limits do not have any direct relation with
specification limits. The determination of whether the process is in control or not depends
on how the sample mean compares with the control limits, not the specification limits
(Recall specifications are ``voice of customer'', whereas control limits are ``voice of
process''). If the sample is large or the sigma capability is high, the control limits can be
much tighter than the specification limits. Hence it is possible that all units meet the
customer's requirements even though the sample mean falls outside the control limits and
signals that the process is out of control.

Q3. For a certain process at Wondrous Widgets, the mean of the output is 25 centimeters and
the standard deviation is 0.3 centimeters. The process manager has suggested that they
collect samples of 16 units per hour instead of 9 units per hour to increase the process’s
sigma capability. Is the manager correct?
a. Yes, the sigma capability will increase because the standard deviation of the
distribution of the sample mean will reduce with increase in sample size.
b. No, increasing sample size will have no impact on the sigma capability of the
process.

2
 c. No, the sigma capability will decrease because a larger sample size will make it
more difficult to identify defective units.
d. You cannot say because it depends on the process’s current sigma capability.

Solution: The sigma capability of a process depends only the process mean, its standard
deviation and the specification limits. It has no relation to the process control mechanism
in place.

Q4. A petrol pump sells 1000 litres of petrol every day. Arranging for one tanker of petrol
delivery costs the petrol pump around INR 4000. The cost of capital for the pump owner
is around 12% per year. Assume the pump pays INR 60 per litre of petrol. What is the
optimal order quantity for the petrol pump? (Assume 360 days per year and the tanker is
large enough to accommodate the required delivery quantity.)
a. 20,000 litres.
b. 10,000 litres.
c. 5,000 litres.
d. 1,000 litres.

Solution: This is a straight forward application of the EOQ formula with R = 1000 per day, S
= 4000 and H = (0.12/360)60 = 0.02 INR per litre per day. Substituting, we get the optimal
quantity as sqrt(2*1000*4000/0.02) = 20,000.

Q5. Frizzz operates two ice-cream stores in the city, which offer identical product mix.
Demand at both stores is stable (constant) but demand at the Hi-Tech City store is twice
that at the Banjara Hills store. However, the average monthly inventory holding cost and
ordering cost is the same for both stores. Assume that the cost of capital as well as the
fixed cost of ordering is the same for both stores. Which of the following statements
CANNOT BE TRUE?
a. Hi-Tech City store follows “Economic Order Quantity” but Banjara Hills store does
not.
b. Banjara Hills follows “Economic Order Quantity” but Hi-Tech City store does not.
c. Both stores do not follow “Economic Order Quantity”.
d. Both stores follow Economic Order Quantity.

Solution: Everything else being the same, if both stores were following their respective
EOQ, the Hi-Tech city store’s order quantity would be sqrt(2) times the Banjara Hills store’s
order quantity. This would mean the average monthly inventory holding cost for the Hi-
Tech city store would be less than twice the average monthly inventory holding cost of the
Banjara Hills store, which is not the case as mentioned in the question.

3
Great American Pajama Company

Great American Pajama (GaP) company has seen a dramatic increase in demand for high
quality lounge and sleep wear as more people work from home during the pandemic. Their
supplier base in Asia provides a lead time of 4 months. Shifting their supplier base to suppliers
in Mexico, who have lost business due to the pandemic, can offer a lead time of 1 month and
will reduce fixed cost of ordering and shipping from $2000 to $1000. However, procurement
cost due to this move will go up from $10 per piece to $11.25 per piece. GaP’s cost of capital
is 12% per year and monthly demand follows a normal distribution with a mean of 100 pieces
and standard deviation of 40 pieces.

Q6. If GaP wants to provide the same service level to their customers as before (95%), what is
the expected impact on their inventory holding and ordering costs due to the relocation
of their supplier base?
a. Reduction of $50 per month
b. Reduction of > $ 50 per month
c. Reduction of < $ 50 per month
d. Increase of $ 50 per month

Solution: The problem parameters and the EOQ, average cost of ordering and inventory
holding cost due to batch ordering, and inventory holding cost because of safety stock
requirements for the two options are shown in the table below.

Old New
S ($ / order) 2000 1000
c ($ / unit) 10 11.25
Opportunity cost
of capital (/$ /
month) 0.01 0.01
H (/unit / month) 0.1 0.1125
R (units / month) 100 100
sigma_R 40 40
L (months) 4 1
CSL 0.95 0.95
C_EOQ ($ /
month) 200.00 150.00
C_SS (Is x H) 13.16 7.40
Total ($ / month) 213.16 157.40

By shifting procurement to Mexico, the company can save 213.16 – 157.40 > $ 50 per
month.

4
Q7. GaP places an order of 1500 units at its supplier in Mexico. To meet this order in its
entirety within the stipulated lead time, the supplier must maintain a finished goods
inventory of 500 units. However, the inventory of finished goods follows a normal
distribution with mean 667 pieces and a standard deviation of 100. What is the supplier’s
ability to meet the order requirement, i.e., it will ship a full order?
a. 99%
b. 95%
c. 50%
d. Close to 0%

Solution: The supplier is unable to meet GaP’s order in full whenever inventory of finished
goods is less than 500. In other words, 500 is the LSL for the inventory maintained by the
supplier. For the finished goods inventory process, we have mean = 667 and standard
deviation = 100. Thus, the probability of shipping a non-defective order = Prob (Inventory
> 500) = Prob (Z > (LSL – mean)/SD) = Prob (Z > (500 – 667)/100 = -1.67) = 95.2%.

Q8. Without increasing her average inventory, what is the maximum standard deviation of
finished goods inventory that the supplier can tolerate if the supplier needs to meet GaP’s
order requirement with probability of 99.8%?
a. The supplier can never achieve the desired service level.
b. 270
c. 55
d. Close to 0

Solution: To meet the order requirement in full 99.8% of the time, we required Prob
(Inventory > 500) = Prob (Z > (500-667)/sigma) = 99.8%, or Prob(Z < (500-667)/sigma) =
0.2%. We have Prob (Z < -3) = 0.2% and therefore, (500 – 667)/sigma = -3, or sigma = 55.67.
___________________________________________________________________________

Honest-To-Core Bank (HTCB)

A loan officer at HTCB takes 10 hours on average (with a standard deviation of 2 hours) to
process a home loan application and dispose it with a final decision. Senior management,
based on an extensive customer study, has arrived at the conclusion that application
processing time should be between 9 and 11 hours to balance between speed and quality.

Q9. If HTCB follows standard practice and sets control limits three standard deviations of the
sample mean from the mean of the process and takes a random sample of 16 loan
applications every 3 days, what should the upper control limit be?
a. 10.5
b. 11.0
c. 11.5
d. 12

5
 Solution: With a sample size of 16, the standard deviation of the sample mean is 2/sqrt(16)
= 0.5. The upper control limit is then 10+3 x 0.5 = 11.5.
___________________________________________________________________________

Aventis Flu Vaccine

Aventis is a major manufacturer of the flu (influenza) vaccine in the U.S.. It produces one dose
of the vaccine for $6 and sells it for $24 per dose. Vaccine doses that remain unsold (due to
mild flu season) are worthless at the end of the season and must be thrown away. Aventis
produces the number of doses that maximizes its expected profit.

Q10. What is the approximate probability that Aventis should target for running out of
vaccines in any given flu season?
a. 25%
b. 75%
c. 20%
d. 80%

Solution: In this case, the cost of under stocking is Cu = 24-6= $18, and the cost of over-
stocking is Co = $6. The optimal service level (at which Aventis' expected profit is
maximized) is given by the Newsvendor formula Cu/(Cu +Co ) = 18/(18+6) = 0.75. Recall
that service level is defined as probability of being able to satisfy demand, or not stocking
out. Thus, with optimal number of dozes, there is 75% probability that Aventis will not run
out of stock. In other words, there is 25% probability that Aventis will run out of stock.

Q11. The U.S. Department of Health is considering a buy-back program, wherein it will
purchase all unsold doses of vaccine from Aventis at the end of the season. For example,
if Aventis manufactures 50 million doses and sells 40 million, the government would buy
back each of the remaining 10 million for some price. What should be the buy-back price
per unit so that the probability of running out of vaccines is reduced to 10%?
a. $2
b. $4
c. $5
d. $10

Solution: Suppose the price per unit of unused dose that the government offers is $b. In
this case, the cost of under stocking is Cu = 24-6= $18, and the cost of over-stocking is Co =
$(6 - b). The optimal service level (at which Aventis' profit is maximized) is given by the
Newsvendor formula Cu/(Cu +Co ) = 18/(18+6 - b). We need the probability of stock-out to
be no more than 10 %. Thus, the service level has to be 90 %. That is, 18/(24 - b) = 0.9 or, b
= $4.

6
Q12. Health activists believe that the current supply chain planning does not account for
health benefits of vaccination. They would like the probability of running out of vaccines
in the season to be reduced to 1%. What is the monetary value of health benefit implied
in this demand of the activists? Assume that there is no buy-back program.
a. $576
b. $594
c. $600
d. $612

Solution: The optimal service level according to the health activists is equal to 100 – 1 = 99
%, when the health benefits of vaccination are accounted for. Thus, the implied cost of
understocking after accounting for health benefits is such that Cu / (Cu + Co) = 0.99. With
Co = $6, we require Cu = 0.99 x (Cu + 6), or Cu = 99 x 6 = $ 594. This Cu includes the
margin from sale of vaccine and therefore the health benefit = $594 – $18 = $576.

___________________________________________________________________________

Rent-A-Suit and ECHO

Rent-a-Suit (RAS) maintains an average inventory of 7500 suits, each of which has an average
useful life of 25 weeks. They donate all old suits to a non-profit, ECHO, which further
distributes them among underprivileged youth seeking job opportunities in the formal sector.
ECHO finds it optimal to pick the old suits from RAS's premises 10 times per year on average.
Assume that the cost of holding one dress in ECHO's inventory is INR 300 per year. (Assume
that 1 year = 50 weeks.)

Q13. What must be the fixed cost incurred per pick-up by ECHO?
a. INR 15,000
b. INR 20,000
c. INR 22,500
d. INR 25,000

Solution: For RAS, I = 7500 and T = 25 weeks. Therefore, the rate at which old dresses are
becoming available is I/T = 7500/25 = 300 per week, using Little's law. As ECHO picks up all
old suits from RAS, the average rate at which dresses are being picked up (and used) by
ECHO, R = 300 per week.
We are given that it is optimal for ECHO to pick up 10 times per year. Thus, the optimal
batch size in which ECHO picks up dresses is R x 50 / 10 = 300 x 50 / 10 = 1500. We know
the inventory holding cost per dress per year for ECHO, He = 300. Substituting in the EOQ
formula, QEOQ = sqrt(2 R S/He), we get S = 22,500 per pick-up.

Q14. RAS also incurs a holding cost of INR 500 per year on the old dresses. It proposes to
increase the shipment frequency to 20 times per year and share half of ECHO's fixed cost
per pick up. Which of the following is MOST LIKELY to happen if this new arrangement was
implemented?

7
 a. RAS will reduce its cost but ECHO would not.
b. ECHO would reduce its cost but RAS would not.
c. Costs will reduce for both.
d. Costs will increase for both.

Solution: Currently, RAS incurs only holding cost on the old dresses. As ECHO picks up in
batches of 1500 dresses, the average inventory of old dresses at RAS is also equal to 1500/2
= 750. RAS's inventory holding cost is therefore equal to 750/2 x Hr = 750 x 500 = 375,000
INR per year. ECHO's cost currently is equal to (R x 50/Q) x S + (Q/2) x H e = 10 x 22,500 +
750 x 300 = 450,000 INR.

With the proposal, Q = R x 50 / 20 = 15000/20 = 750. As RAS shares half the cost of pick up,
RAS's annual costs are now equal to (R x 50)/Q x S/2 + (Q/2) x Hr = 20 x 22500/2 + 750/2 x
500 = 412,500 INR. On the other hand, ECHO's cost becomes 20 x 22500/2 + 750/2 x 300 =
337,500 INR.

CreamStone Ice Cream Parlor

Daily demand at the CreamStone Ice Cream Parlor follows a Normal distribution with a mean
of 100 tubs and a standard deviation of 50 tubs. The ice cream is supplied by a wholesaler
who charges $2 per tub and a $90 delivery fee independent of the order size. The opportunity
cost of capital for CreamStone is 25% per year. Currently, the owner places an order for 2000
tubs of ice cream each time she has 600 tubs on hand. Assume 50 weeks in the year.

Q15. If the delivery lead time is 4 days, what is CreamStone's service level? (Choose the
closest option if no option is a perfect match.)
a. 99.3 %.
b. 97.72 %.
c. 84.13 %.
d. 50 %.

Solution: The average daily demand R = 100 tubs. With an order quantity Q = 2000 tubs,
the average cycle inventory because of batch ordering is equal to Q/2 = 1000 tubs.

The re-order point ROP = 600 tubs, while the mean demand during the lead time is equal
to R x L = 100 x 4 = 400 tubs. Thus, the safety inventory Is = ROP - R x L = 600 - 400 = 200
tubs. The standard deviation of lead-time demand is equal to sigma x sqrt(L) = 50 x sqrt(4)
= 100. Thus, the z value corresponding to the service level is equal to Is / (sigma x sqrt(L) )
= 200 / (100) = 2. Thus, the implied service level is Prob(Z <= 2) = 0.977.

8
Q16. What is the average time that a tub of ice cream spends on Creamstone’s shelves?
a. 6 days
b. 12 days
c. 16 days
d. 26 days

Solution: For an order size of Q and safety stock of Is tubs, the total average inventory of
ice cream tubs held by Creamstone is = Q/2 + Is. We have Q = 2000, Is = 200 and therefore
average inventory = 2000/2 + 200 = 1200. The average daily demand is 100 tubs and
applying Little’s Law, the average “flow time” for a tub of ice cream at Creamstone =
average inventory / average demand = 1200 / 100 = 12 days.

Q17. A predictive analytics vendor claims that it can reduce the standard deviation of daily
demand to 25 tubs. If CreamStone wants to maintain the same service level and nothing
else about the business changes, what is the maximum amount it should be willing to pay
for this service per year?
a. $50
b. $250
c. $1250
d. $2500

Solution: As nothing else about the business has changed, the reduction in demand
uncertainty impacts the safety stock held to achieve the 97.72 % service level determined
in the answer to the previous question. We know Is = Zsl x sigma x sqrt(L). As sigmanew =
sigma/2, the new safety stock requirement is half of what it was and is equal to 200/2 =
100 tubs. The inventory holding cost per tub per year is equal to 0.25 x $2 = $0.5. Thus, the
benefit of using the predictive analytics is the inventory cost saving on 100 tubs, i.e., 100 x
0.5 = $50.

Q18. CreamStone has been getting complaints from disgruntled customers on their
Instagram page, and would like to halve its probability of stockout. What is the maximum
amount that it should be willing to offer to the predictive analytics vendor now if the
impact of their solution on standard deviation of the demand is the same as before?
a. < $50
b. $50
c. $50 - $100
d. $100

Solution: Creamstone’s current probability of stockout is equal to 100 – 97.72 = 2.28 %.

Halving the probability of stockout would require increasing the service level to 98.86% and
would require a safety stock of more than 200. With sigmanew = sigma/2, the difference
between the safety without and with the improved demand forecast would be greater than
200/2. Thus, the saving in inventory holding cost would be higher than 100 x 0.5 = 50.

9
 To justify paying $ 100 to the vendor, the safety stock requirement should reduce by 200
tubs because of the better demand predictions. This implies the safety stock required to
maintain at 98.86% service level at current variability should be equal to 400, which is not
the case as Z98.86 < 2 x Z97.72.
___________________________________________________________________________

Sunshine Energy

Sunshine Energy (SE) assembles and installs solar panels for industrial customers. Currently,
SE has no limits on the amount of work-in-process inventory, and often there are more than
200 parts between some stages. The manager, Ritesh, considers limiting the amount of work-
in-process inventory between stations to no more than 10 parts at any time.

Q19. Do you agree with Ritesh that this change will take the process closer to a pull system?
a. No, pull system should have zero inventory.
b. No, pull system must use Kanban cards.
c. Yes, upstream resource will stop processing if inventory reaches the limit.
d. Yes, limiting inventory will balance capacity of resources.

Solution: By having a limit on the amount of work-in-process inventory between two stages,
the upstream station will not produce unless the inventory drops below the limit, which
results from the demand pull by the downstream station.

Q20. Which of the following is MOST LIKELY to be a consequence in the short run because
of this change?
a. Variability of resource capacity will increase.
b. Throughput will not change as capacity will be unchanged.
c. Throughput will increase due to lower buffer between stations.
d. Capacity utilization may reduce due to lower buffer between stations.

Solution: This will likely reduce the effective output of the system in the short run due to
variability. This is because the process will lose capacity when downstream machines are
starved of input due to lower inventory levels and remain idle.

10
Seton Hospital

Seton is considering building a new CABG (Coronary Artery Bypass Graft, commonly known
as open heart surgery) facility. Annual demand for CABG is estimated as follows:

CABG Demand
Probability
(# of procedures)
400 25%
700 50%
1,000 25%

Seton’s relevant capacity costs and contribution margin (= price – variable cost) are as follows:

Costs & Margin Incremental

Fixed Cost of Contribution Margin
CABG Cost of Capacity
Capacity per Procedure
per Procedure
(Million $) (Thousand $)
(Thousand $)
1.5 6 12

Q21. What is the optimal capacity that Seton should plan for the CABG facility? (Find the
closest answer)
a. 400
b. 500
c. 700
d. 800

Solution: The tradeoff involved in the capacity investment decision is the same as the
tradeoff in the inventory decision for a perishable product. The cost of capacity underage,
Cu = $6,000 while the cost of excess capacity Co = $6, 000. Thus, the optimal service level
to target = 6 / (Cu + Co) = 6/12 = 1/2. Thus, the optimal capacity investment should be the
smallest capacity that has a probability of meeting demand equal to or greater than 1/2,
that is, optimal capacity = 700.

Q22. What is Seton’s optimal expected net profit after accounting for the cost of CABG
capacity?
a. $ 1.2 million
b. $ 1.4 million
c. $ 1.6 million
d. $ 1.8 million

Solution: The cost and profitability assessment of the optimal capacity investment to
handle a demand of up to 700 procedures is shown in the table below:

11
 CABG Cost of Capacity (Million $)
Capacity Fixed Incremental Total Cost
700 1.5 4.2 1.5 + 4.2 = 5.7

CABG CABG Probability Contribution Profit Expected

Demand Performed Margin (Million $) Profit
(Million $) (Million $)
400 400 25% 4.8 4.8–5.7= - 0.9
700 700 50% 8.4 8.4–5.7 = 2.7 1.8
1,000 700 25% 8.4 8.4–5.7 = 2.7

Always Accurate Labs

A large national diagnostic laboratory chain receives blood samples at a constant rate of 10
per min which are first assembled in batches of 100. The batches are then processed one at
a time in a large machine, whose processing time is equal to 12 minutes irrespective of the
batch size.

Q23. What is the average time that a blood sample spends in the overall process?
a. 22 minutes.
b. 17 minutes.
c. 12.1 minutes.
d. 12 minutes.

Solution: As samples arrive at the rate of 10 per minute, it takes 100/10 = minutes to build
a batch at the assembling step. The inventory of samples between the assembling and
processing steps increases at a steady rate from 0 to 100, and then goes to zero once
processing starts. Therefore, the average inventory of samples between the two stages is
100/2 = 50 and average time spent by a sample = 50/(10/minute) = 5 minutes, using Little’s
Law. The testing takes 12 minutes, yielding a total flow time of 5 + 12 = 17 minutes.

12
Smartcom

Smartcom is an upcoming indigenous manufacturer of smartphones. It imports components

and assembles them into two smartphone models, Vajra and Vayu. Demand for the two
models are independent of each other. Their mean demand is 1000 and 1440 units per month,
respectively and standard deviation is 300 and 400 units per month, respectively. Both models
use the same processor chips, which are imported and have a lead time of about 4 months.
Smartcom would like to maintain a service level of 98% for its processors, i.e., only 2% chance
that assembly operations have to stop due to inadequate inventory of processors. Each
processor chip costs INR 10,000 and Smartcom’s annual cost of capital is 10%. Fixed cost of
ordering and shipping processors from the supplier is INR 45,000.

Q24. What is the safety stock of processors that Smartcom needs to maintain?
a. 1232
b. 1643
c. 2054
d. 2875

Solution: To calculate safety stock requirement for processors, we need to consider the
sum of demand from the two models as the same processor is used in both. Note that if
you calculated the safety stock separately for the two models and added them up, you
would get a higher requirement for safety stock.

Vajra Vayu Total

R 1000 1440 2440
standard
dev 300 400 500
L 4 4 4.00
Service
level 0.98 0.98 0.98
z 2.05 2.05 2.05
Is 1232 1643 2054
S 45000 45000 45000
H 83.3 83.3 83.3
EOQ 1039 1247 1623
C_Is 1.0 1.4 1.7
C_EOQ 0.9 1.0 1.4
C_Ic 0.4 0.5 0.7
C_Is + C_Ic 1.5 1.9 2.4

13
Q25. What is the total monthly inventory holding cost for processors incurred by Smartcom
(in lacs)?
a. 1.7
b. 2.2
c. 2.4
d. 2.6

Solution: See the table above. Note that we are calculating only the invenoty holding cost
and not the cost of ordering. So, this is equal to H*(Is+Ic), where Ic = Q_EOQ/2. Again, note
that calculating the inventory and costs separately for two models and then adding up will
give incorrect answer.

14