Operations Management (OPMG)

Indian School of Business

Practice Problems

1. Since becoming the CEO, you have noticed that your mailbox is perennially full. On average,
about half of the emails in your inbox are “unimportant” emails (those from professional
industry associations, company-wide emails etc.) that are copied to you but don’t need your
response. The rest are “important” and need your personal response. You believe in clutter
free inbox, so you process every email that comes, either by simply moving it to a separate
folder or by responding to it and then moving it to a separate folder. Naturally you prioritize
processing “important” emails and as a result the time spent by “unimportant” emails in your
inbox is four times that spent by “important” emails. If you receive 15 total emails every hour
on average, how many of these are “important”?
A. 3
B. 5
C. 12
D. Cannot be determined

Solution: Let ri be the flow rate of important emails and ru of unimportant emails. Let
t be the flow time of important emails so that 4t is the flow time for unimportant emails.
We know that ri ∗ t = 60 and ru ∗ 4t = 60. Therefore, ri = 4 ∗ ru . But ri + ru = 15. So
ri = 12.

2. When a homeowner wants to sell his or her home, the property is typically posted on a regional
listing service. In a particular region, 60% of the new properties added to the listing service in
an average week are “houses”. The remaining 40% of new listings added are “condominiums”.
The average house listed on the service stays there for 4.5 weeks, before it is either sold or
the listing taken out for other reasons. The average condominium stays listed for 8.25 weeks.
What percent of active property listings on the service (i.e., properties posted on the listing
service but not yet removed) are condominiums?
A. 40
B. 45
C. 50
D. 55
 Solution: Looking at the process of posting new properties on the listing service, we can
identify two different flow units: houses and condominiums. Let R be the total flow rate
of the overall listing process, that is the number of properties posted per week. Then, Rh ,
the flow rate of the houses is equal 0.6R, while the flow rate of condominiums, Rc = 0.4R.
From the problem, we know Th = 4.5 weeks, while Tc = 8.25 weeks. Using Little’s law on
each of these flow units, we have Ih = Rh × Th = 0.6R × 4.5 = 2.7R and Ic = Rc × Tc =
0.4R × 8.25 = 3.3R. The total number of active property listings, I = Ih + Ic = 2.7R +
3.3R = 6R, and the proportion of active property listings that are condominiums is equal
to Ic /I = 3.3R/6R = 0.55, or 55%.

(Q3–Q4) Mt. Kinley is a strategy consulting firm that divides its consultants into three classes:
associates, managers, and partners. The firm has been stable in size for the last 20 years with
about 200 associates, 60 managers, and 20 partners on average. The work environment at Mt.
Kinley is quite competitive with an “up or out” culture. After four years, an associate either
becomes a manager or is asked to leave the firm. Similarly, after six years, a manager either
becomes a partner or leaves the firm. Partners stay with the firm for another 10 years on
average before leaving to pursue other “life goals”. The firm recruits MBAs from top B-schools
as associates. No hires are made at the manager or partner level.

3. How many new MBA graduates does Mt. Kinley have to hire every year?
A. 200 graduates per year
B. 100 graduates per year
C. 50 graduates per year
D. 10 graduates per year

Solution: This is an application of Little’s law to the associate process. The first step is
to figure out what is inventory (I), throughput rate (R) and flow time (T ) for this process.
We are told that the flow time is T = 4 years, and the inventory of associates is I = 200.
Hence, by Little’s law, to maintain this inventory we need that the number of graduates
hired every year (the throughput R) is given by R = I/T = 200/4 = 50.

4. What % of MBA graduates hired by Mt. Kinley become partners?

A. 100%
B. 20%
C. 10%
D. 4%
 Solution: The input to the partner process is managers, and the input to the manager
process is associates. However, not all MBA graduates hired as associates get promoted
and not all associates get promoted. The fraction of MBA graduates who get promoted to
be a manager is the ratio of the flow rates of the manager and associate process.
As the process is stable, the input rate to the manager process is equal to the flow rate
of the manager process. For the manager process, we are told that the inventory I = 60
managers and flow time T = 6 years. Applying Little’s law to the manager process, the
number of associates promoted to be managers each year (the flow rate R) is given by
R = I/T = 60/6 = 10. Similarly, applying Little’s Law to the partner process, we have
R = I/T = 20/10 = 2 per year for the partner process.
From the previous question we know that 50 MBA graduates are hired each year as asso-
ciates. Thus, the % of MBA graduates who get promoted to become partner is equal to
2/50 = 0.04, or 4%.

5. After going through tough times, Rent-A-Dress has successfully pivoted its business model and
has scaled up to a demand of 150 rentals per day. They only serve the 10-day rental segment
and continue sending two dresses per rental to provide flexibility to their customers. They have
improved their internal processes so that it only takes 5 days to complete ALL steps needed
for a returned dress to be with a new customer again. They have also made changes to the
dresses they procure and the cleaning process so that the dresses now have a useful life of 150
days. What is the average number of new dresses that RAD needs to purchase each month?
(If none of the choices match your exact answer, pick the closest value.)
A. 9000
B. 900
C. 300
D. 30

Solution: First, consider RAD process consisting of “customer” step and “internal” step.
Applying Little’s Law to each of these Icustomer = 10 ∗ (150 ∗ 2) = 3000 and Iinternal =
5 ∗ (150 ∗ 2) = 1500. Thus total inventory of dresses is Itotal = 4500. Now, if we consider
RAD as a process with new dresses as the input and discarded dresses as the output,
then we have I = 4500 and T = 150 days for this process. Using Little’s Law, we get
R = I/T = 4500/150 = 30 dresses / day. Assuming a month of approximately 30 days,
this translates to 900 dresses per month.

6. (Q6–Q7) The X-ray scanner used at the airport security check can screen carry-on bags at
the rate of 12 per minute. Around 5% of the bags set off an alarm during this scan and have to
undergo an additional thorough check by a security officer, which takes 90 seconds on average.
 What is the maximum possible rate at which bags can go through the security check? Assume
that the security officer’s time required in all other steps is negligible.
A. 600 per hour
B. 40 per hour
C. 720 per hour
D. Cannot be determined

Solution: We need to find the capacity of the security check process. The X-ray scanner
has a capacity of 12 bags per minute, which is equivalent to 720 bags per hour. The
security officer’s capacity is 60/90 = 2/3 bags per minute, or 40 bags per hour. This is
lower than the X-ray scanner’s capacity. However, only 5% of the bags screened by the X-
ray scanner require thorough screening. When the X-ray scanner is working at capacity, the
demand placed on the security officer is 0.05 × 720 = 36 bags per hour and her utilization is
36/40 = 0.9. Thus, the X-ray scanner is the bottleneck resource and the maximum possible
flow rate for the security check process is equal to 720 per hour.

7. During heightened security conditions, close to 6.67% of the bags screened by the X-ray scanner
are selected for additional screening by the security officer. What is the maximum possible rate
at which bags can go through the security check during heightened security conditions?
A. 600 per hour
B. 40 per hour
C. 720 per hour
D. Cannot be determined

Solution: If the X-ray scanner is working at capacity, then the demand placed on the
security officer is equal to 0.067 × 720 = 48 bags per hour, which is greater than the
security officer’s capacity. Therefore the maximum flow rate through the X-ray scanner,
and hence the overall security check process, has to be less than 720 per hour. Let r be the
maximum possible flow rate in bags per hour. We require r × 0.067 = 40, or r = 597 ≈ 600
bags per hour.
Notice that when a larger percentage of bags require additional screening, the demand
placed on the security inspector increases and she becomes the bottleneck resource.

8. (Q8–Q9) Albiote Laboratories has invested in a high-performance centrifuge to treat blood

samples received from clinics. In the first step of chemical purification, a technician places
test tubes containing the sample in a purifier and adds necessary chemicals; this step takes 3
minutes per sample. In the second step, the same technician loads the sample test tubes in
the centrifuge (which can hold up to 5 samples) and adjusts its settings and switches it on.
This loading and setting step takes 5 minutes, irrespective of the number of samples placed
 in the centrifuge. The actual centrifuging step itself takes an additional 25 minutes and does
not require the technician to supervise. Assume that the unloading step is instantaneous and
clinics send their samples in groups of 5 to the lab. What is the capacity of the process in
terms of the number of treated samples per hour?
A. 5 samples per hour
B. 10 samples per hour
C. 15 samples per hour
D. 20 samples per hour

Solution: Using a group (5 samples) as a flow unit, the time contribution and capacity of
the different resources used in the process are as shown in the table below.

Resource Time contribution (min.) Capacity (flow units / hr)

Tech. 15 + 5 = 20 60/20 = 3
Purifier 15 60/15 = 4
Centrifuge 30 60/30 = 2

The centrifuge is the bottleneck resource with a capacity of 2 flows units per hour, which
is equal to 2 × 5 = 10 samples per hour.

9. It has been observed that the minimum time taken to finish treating a set of 5 samples, starting
with the time it is received from the clinic to when it is finished, is 45 minutes. In steady state,
the process operates at 75% capacity and there are 15 samples in process on average. What is
the average waiting time experienced by samples?
A. 75 minutes
B. 65 minutes
C. 45 minutes
D. 35 minutes

Solution: Let r be the actual flow rate of the process. We have r = 0.75 × 10 = 7.5
samples per hour. We are given that the inventory I = 15 and using Little’s Law we have
the actual flow time, t, for the process as t = I/r = 15/7.5 = 2 hours, i.e., 120 minutes.
This flow time consists of both value added activity times as well as non-value added
waiting time. The value-added activity time is nothing but the theoretical flow time, which
is given as 45 minutes. Thus, the waiting time experienced in the process is equal to
t − 45 = 120 − 45 = 75 minutes.

10. In a simple process, work flows from activity A to activity B to activity C. Activity A takes
10 minutes and is performed by Abhilash. Activity B takes 15 minutes and is performed by
Bharat and Chetna together, while activity C takes 10 minutes and is performed by Chetna
and Abhilash. What is the bottleneck in this process?
 A. Activity A
B. Activity C
C. Activity B
D. Cannot be determined

Solution: Only a resource, not an activity, can be a bottleneck. But the time spent by
each resource on each activity is not given and hence the bottleneck cannot be determined.

11. Alarmed by complaints about poor customer service, Cheap Loans Inc. have decided to replace
human agents with a chatbot. This has reduced the standard deviation in the service times
by half. The overall capacity and all other aspects of the service process remain unchanged. If
the average waiting time with human agents was 100 seconds, which of the following is NOT
likely to be the new waiting time?
A. 20
B. 30
C. 40
D. 50

Solution: The new waiting time cannot be less than a fourth of the old waiting time
because the variability in the inter-arrival times and all other parameters are unchanged.
Hence, the only option that is NOT likely is 20 seconds.

12. Snippy, a new startup by an ISB alumnus, provides EV bikes on rent from a single location in
Bandra Kurla Complex, Mumbai. The average rental duration is around 30 minutes whereas
the standard deviation is around 15 minutes. The time between two successive arrivals to the
station has a mean of 5 minutes and a standard deviation of 10 minutes. Snippy wants the
average waiting time for a customer to be less than 15 minutes. The minimum number of bikes
that are required to be kept at the station are:
A. 6
B. 7
C. 8
D. 9

Solution: Ri = 12 per hour. Rp = 2 per hour. Also, CVi = 2 whereas CVp = 0.5.
Substituting in the formula for different values of c > 6 shows that average waiting time is
more than 40 minutes at c = 7 and around 13 minutes at c = 8.
13. Vroom runs a hyperlocal model of grocery delivery for busy professionals. A survey of the target
segment found that the customers want the delivery time to be between 150 min and 180 min.
However historical data shows that the actual delivery time follows a Beta distribution with
mean 165 and standard deviation 10. Vroom’s delivery manager checks for abnormal shocks
to the process by collecting a random sample of 25 orders every day and checking the sample
average against the control limits of 161 and 169. The cost of each investigation is INR 20,000.
What is the approximate average cost of investigation per day (over the long run)?
A. 20,000
B. 5,000
C. 1,000
D. 500

Solution: The sample average follows a normal distribution with mean 165 and standard
√
deviation 10/ 25 = 2. The control limits are two-standard deviations away from the mean
of the sample average’s distribution.
An investigation is performed whenever the sample average value lies outside the control
limits. Thus, in the long-run the probability of carrying out an investigation is equal to
P rob(X̄ ≤ 161) + P rob(X̄ ≥ 169) = 2 × P rob(Z ≤ −2) = 0.0455, and the approximate
average cost of investigation per day (in the long run) is equal to 20, 000 × 0.0455 ≈ 1, 000.

14. The specification limits for the coil temperature in an Espresso machine are 185 Fahrenheit
and 175 Fahrenheit. However, there is variability in the process with mean of 180 Fahrenheit
and standard deviation of 5 Fahrenheit. Suppose the machine has a problem and the standard
deviation increases to 10 Fahrenheit. Which of the following is LIKELY to happen?
A. The width of the USL and LSL band will increase.
B. More defects will be generated until the problem is detected and resolved.
C. Sample mean is less likely to be out of control limits until the problem is detected
and resolved.
D. None of the above

Solution: The USL and LSL will not change because of a change in the process. As
σ of the process increases, the sigma–capability of the process decreases and the process
produces more “defective” cups of coffee. Further, more sample means are likely to fall
outside the control limits when the standard deviation increases.

15. C & A’s potato chip filling process has a lower specification limit of 9.5 oz. and an upper
specification limit of 10.5 oz. What is the maximum allowable standard deviation that achieves
the sigma-capability of 3?
A. 0.5
 B. 0.25
C. 0.167
D. Not enough information

Solution: For standard deviation to have the maximum possible value to achieve a given
sigma-capability, the mean should be equi-distant from the two specification limits. Thus,
we need µ = 10 oz, and the maximum possible standard deviation is equal to (U SL−µ)/3 =
0.5/3 = 0.167.

16. You are planning to carry a prepaid travel card during a month long international vacation (er..
international immersion program!!). Your incidental expenditure during the trip is likely to be
normally distributed with a mean of $900 and standard deviation of $100. If the balance on
the travel card is exhausted, you will have to use your international credit card, which charges
1.2% for every dollar you spend. Any unspent amount can be transferred back to your bank
account after the trip. Assume that the savings bank interest is around 4.8% per annum. How
much amount should you load on the card (rounded up to the full dollar)?
A. $833
B. $968
C. $978
D. $985

0.012
Solution: Co ≈ 0.048/12 = 0.004. Cu = 0.012. Thus, the target service level is 0.004+0.012 =
0.75. The z-value corresponding to this is approximately 0.675. Hence, the amount that
should be loaded on the card is approximately $900 + $100*0.675 = $967.5.

17. You have started a bakery that is committed to selling fresh, artisan, organic products. Your
two main products are bread and cake. Assume that demand for both have the same mean
but bread has lower variability than cake. Your partner, who is an industry old-hand suggests
that the optimal service level for cakes should be higher. Do you agree?
A. Yes, because more variability in demand requires a greater service level.
B. Yes, because more stock should be kept to protect against higher variability.
C. No, because the optimal service level does not depend on demand vari-
ability.
D. No, because the optimal service level depends only on the mean demand.
 Solution: The optimal service level, Cu /(Cu + Co ), does not depend on the demand vari-
ability. The optimal stocking quantity, Q, to achieve the optimal service level depends on
the demand variability and this stocking quantity can increase or decrease, depending on
the optimal service level.

18. Exquisite Wine Wholesalers (EWW) imports fine European wines into the country four times
a year to meet its stable demand of around 1600 barrels per annum. The cost of placing and
receiving each order (including freight, portage etc.) is $5,000 whereas the opportunity cost
for EWW is 10% per annum and the cost of each barrel is $4,000. The annual cost of ordering
and inventory holding for EWW’s operation is:
A. $200,000
B. $100,000
C. $80,000
D. $20,000

Solution: As EWW imports four times a year, the total ordering cost per year is 4×5000 =
$20, 000 and the average order size is 1600/4 = 400 barrels. The inventory holding cost
per barrel per year is 0.1 × 4000 = $400. With an order size of 400, the average inventory
is equal to Q/2 = 400/2 = 200 and the total inventory holding cost per year is equal
to 200 × 400 = $80, 000. The total cost of ordering and inventory holding is therefore
20, 000 + 80, 000 = $100, 000.

19. (Questions 19–20) Inspired by Kristen’s success story, Krish from the Inspirational School
of Business decided to launch a cookie business. The business receives 12 orders per hour from
6 pm until 9 pm. Krish starts processing the orders from 6 pm itself (without postponing
homework!!) at the rate of 6 orders per hour. What is the average wait time for an order
during a typical night?
A. 0.5 hours
B. 1 hour
C. 1.5 hours
D. 2 hours

Solution: The number of orders waiting increases at a rate of (12 − 6) = 6 orders / hr

from 6 to 9 PM, and reduces at the rate of (6 − 0) = 6 orders / hr from 9 PM onwards. As
there are 18 orders waiting at 9 PM, it takes 18/6 = 3 hours after 9 PM to finish processing
all orders.
The average number of orders waiting is equal to 18/2 = 9 and the average flow rate of the
process is equal to 36/6 = 6 orders / hr. Applying Little’s law, we get the average wait
time as 9/6 = 1.5 hours.
20. Krish can rent additional equipment to increase the process capacity to 9 orders per hour. How
much maximum rental per day should he be willing to pay for such equipment? Assume that
Krish’s business incurs a cost of $2 for every hour that an order waits and that Krish values
his own time at $10 per hour.
A. $128
B. $108
C. $92
D. $72

Solution: With the additional equipment, the number of orders waiting increases at a rate
of (12 − 9) = 3 orders / hr from 6 to 9 PM, and reduces at the rate of (9 − 0) = 9 orders
/ hr from 9 PM onwards. As there are 9 orders waiting at 9 PM, it takes 9/9 = 1 hours
after 9 PM to finish processing all orders.
The total time spent per day decreases by 2 hours (see answer above) because of the
additional equipment. In addition, the total order hours waiting reduces by 18 × 6/2 − 9 ×
4/2 = 36. Thus, the total cost saved because of additional equipment is 2hours × $10 = 20
of labor cost plus 2 × $36 of waiting cost = $ 92.