Decision Making Under Uncertainty (2014-15)

Max. Marks: 100

Time: 2 hours 30 minutes

--------------------------------------------------------------------------------------------------------------------Note:
I.

There are two sections in the question paper. Answers to each section should be written
on separate answer-books.

II.

Questions 1 to 4 form constitute Section-I (65 Marks).

a. Questions 1 and 2 are compulsory (30+20 = 50 Marks)
b. Candidates should answer any one among Questions 3 and 4. (15 Marks)

III.

All questions in Section-II (35 Marks) are compulsory.

Section I Decision Making

Q1.

[30 Marks] The FurnishHome Company has decided to buy a plot of land with a good
timber forest on it. The price of the land is Rs.220 lakhs now, in January. The company
does not need to harvest the timber immediately and can wait until March to buy the plot
of land. They estimate the land including the timber forest on it is worth Rs.300 lakhs to
them. FurnishHome is aware of a competitor who is also planning to buy the plot.
The price of the plot is likely to change from now till March. The following table shows
the three possible changes in price of the plot from January to February if the plot is not
sold in January.
Price change

Probability

Increase by Rs.30 lakhs

0.1

Remain the same

0.6

Decrease by Rs.30 lakhs

0.3

Changes in the price of the plot from February to March depend on the price changes that
take place in February. The price in March can either remain the same as that in February
or decrease by a further Rs.30 lakhs. In March, the company estimates that the price of
the plot will remain the same as in January with a probability of 0.54, reduce by Rs.30
lakhs with a probability of 0.18, and reduce by Rs.60 lakhs with a probability of 0.24.
Page 1 of 5

(These values are under the assumption that the plot remains unsold in January and
February.) FurnishHome will know the price of the plot in a particular month before they
decide whether or not to buy the plot during that month.
FurnishHome estimates that if they do not buy the plot in January, the probability that the
plot will not be available to FurnishHome in February (i.e., be bought up by a competitor)
is 0.05; and if the plot is still for sale after February, the probability that the plot will not
be available to FurnishHome in March is 0.1.
Note: Ignore the time value of money. Also ignore the possibility of alternate uses of the
money that FurnishHome should set aside for the purchase decision.
a) [2 Marks] When would FurnishHome plan to buy the plot of land if they use maximax
criterion to take decisions? What is the value of maximax payoff?
b) [4 Marks] If FurnishHome decides not to buy the plot in January or February, what would
their expected payoff from the plot be?
c) [12 Marks] When should FurnishHouse buy the plot of land if they use expected value
criterion to take decisions? What would be their expected payoff?
d) [12 Marks] Suppose that FurnishHome has perfect information about the change in price
from January to February. In each of the three scenarios possible, what would be
FurnishHomes expected payoff if they do not buy the plot of land in January? What is
the expected value of perfect information?

Q2.

[20 Marks] The chief executive officer of a firm in a highly competitive industry
believes that one of her key employees is providing confidential information to the
competition. She is 90% certain that this informer is the vice president of finance, whose
contacts have been extremely valuable in obtaining financing for the company. If she
decides to fire this vice president and he is the informer, she estimates that the company
will gain $500,000. If she decides to fire this vice president but he is not the informer, the
company will lose his expertise and still have an informer within the staff; the CEO
estimates that this outcome would cost her company about $2.5 million. If she decides
not to fire this vice president, she estimates that the firm will lose $1.5 million regardless
of whether he actually is the informer (because in either case the informer is still with the
company). Before deciding whether to fire the vice president for finance, the CEO could
Page 2 of 5

order lie detector tests. To avoid possible lawsuits, the lie detector tests would have to be
administered to all company employees, at a total cost of $150,000.
Another problem she must consider is that the available lie detector tests are not perfectly
reliable. In particular, if a person is lying, the test will reveal that the person is lying 95%
of the time. Furthermore, if a person is not lying, the test will indicate that the person is
not lying 85% of the time.
a) [4 Marks] To minimize the expected total cost of managing this difficult situation, what
strategy should the CEO adopt?
b) [8 Marks] Should the CEO order the lie detector tests for all of her employees? Why or
why not?
c) [8 Marks] Determine the maximum amount of money that the CEO should be willing to
pay to administer lie detector tests.

Q3

[15 Marks] If an athlete is tested for a certain type of drug use (steroids etc.), the test
result will be either positive or negative. However, these tests are never perfect. Some
drug-free athletes test positive, and some drug users test negative. The former are called
false positives whereas the latter are called false negatives. Lets assume that 5% of
all athletes use drugs, 3% of all tests on drug-free athletes yield false positives, and 7%
of all tests on drug users yield false negatives. Suppose a typical athlete is tested.
a) [2.5 Marks] If the athlete tests positive, how sure can you be that he is a drug user?
b) [2.5 Marks] If the athlete tests negative, how sure can you be that he does not use drugs?
Now assume that there is a second test, independent of the first, that can be used as a
follow-up. Assume that its false positive and false negative rates are 0.02 and 0.06.
c) [5 Marks] If the athlete also tests positive in this second test, what is the posterior
probability that he is a drug user?
d) [5 Marks] We assumed that the two tests are independent. Why might this not be
realistic? If they are not independent, what kind of additional information would you
need about the likelihoods of the test results?
Page 3 of 5

Q4.

[15 Marks] A company is considering whether to market a new product. Assume, for
simplicity, that if this product is marketed, there are only two possible outcomes: success
or failure. The company assesses that the probabilities of these two outcomes are p and 1p, respectively. If the product is marketed and it proves to be a failure, the company will
have a net loss of $450,000. If the product is marketed and it proves to be a success, the
company will have a net gain of $750,000. If the company decides not to market the
product, there is no gain or loss. The company is also considering whether to survey
prospective buyers of this new product. The results of the consumer survey can be
classified as favorable, neutral, or unfavorable. In similar cases where proposed products
were eventually market successes, the fractions of cases where the survey results were
favorable, neutral, or unfavorable were 0.6, 0.3, and 0.1, respectively. In similar cases
where proposed products were eventually market failures, the fractions of cases where
the survey results were favorable, neutral, or unfavorable were 0.1, 0.2, and 0.7,
respectively. The total cost of administering this survey is C dollars.
a) [2 Marks] Let p = 0.4. For which values of C, if any, would this company choose to
conduct the consumer survey?
b) [5 Marks] Let p = 0.4. What is the largest amount that the company will be willing to pay
for perfect information about the potential successes or failure of the new product?
c) [8 Marks] Let p = 0.4 and C = $15000. Find the strategy that maximizes the companys
expected earnings in this situation. Does this optimal strategy involve conducting the
consumer survey? Why or why not?
Section II Simulation

Q5

[4 Marks] Show how a random sample can be obtained from a distribution with the

following probability density function: f (x ) x 1e (x / ) ; x, , 0

Use the following random numbers: 0.7736, 0.4475 and 0.5818.

Q6

[10 Marks] A company trading motor car spares wishes to determine the level of stock it
should carry for the items in its range. Demand is not certain and there is lead time for
stock replenishment. For one item, the following demand information is obtained:
Page 4 of 5

Demand (units/day)
Probability

3
0.1

4
0.2

5
0.3

6
0.3

7
0.1

For the said item, the carrying cost is 20 paisa per unit for a day. The ordering cost is Rs.
5 and lead time is 3 days.
Stock in hand at the beginning of the simulation process is 20 units. You are required to
carry out a simulation run over a period of 10 days if the ordering policy is to order 15
units when the stock in hand falls to 15 units. Calculate the total cost of operating this
inventory for 10 days using the following sequence of random numbers:
0, 90, 10, 10, 50, 10, 80, 60, 30, 50, 70, 10, 20 and 90
Q7

[21 Marks] Determine the optimum number of mechanics for 100 semi-automatic
machine tools. The operation of the machine tools is automatic and warrants attention of
the mechanics only when there is breakdown. The breakdowns have been seen to occur at
the following times
Breakdown 1
2
3
4
5
6
7
8 9
Time (hour) 0 1.2 2.1 2.4 2.6 3.8 4.3 5.1 6

The repair time for the machine tools have been observed to be according to the
following distribution:
Time of repair (hr) 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Frequency
50 110 210 350 105 70 50 45 10

The wage of a mechanic is Rs. 18 per hour. Down time cost of the machine is Rs. 5 per
hour. Calculate whether two or three mechanics should be employed.
Use the random numbers : 105, 159, 885, 989, 657, 888, 729, 285, 530

END

Page 5 of 5