4

Decision Analysis MULTIPLE CHOICE

1. The options from which a decision maker chooses a course of action are
a. called the decision alternatives.
b. under the control of the decision maker.
c. not the same as the states of nature.
d. All of the alternatives are true.
ANSWER: d
TOPIC: Structuring the decision problem

2. States of nature
a. can describe uncontrollable natural events such as floods or freezing temperatures.
b. can be selected by the decision maker.
c. cannot be enumerated by the decision maker.
d. All of the alternatives are true.
ANSWER: a
TOPIC: Structuring the decision problem

3. A payoff
a. is always measured in profit.
b. is always measured in cost.
c. exists for each pair of decision alternative and state of nature.
d. exists for each state of nature.
ANSWER: c
TOPIC: Payoff tables

4. Making a good decision

a. requires probabilities for all states of nature.
b. requires a clear understanding of decision alternatives, states of nature ,
and payoffs.
c. implies that a desirable outcome will occur.
d. All of the alternatives are true.
ANSWER: b
TOPIC: Decision making without probabilities

5. A decision tree
a. presents all decision alternatives first and follows them with all states of nature.
b. presents all states of nature first and follows them with all decision alternatives.
c. alternates the decision alternatives and states of nature.
d. arranges decision alternatives and states of nature in their natural chronological
order.
ANSWER: d
TOPIC: Decision trees

1
2 Chapter 4 Decision Analysis

6. Which of the methods for decision making without probabilities best protects the decision
maker from undesirable results?
a. the optimistic approach
b. the conservative approach
c. minimum regret
d. minimax regret
ANSWER: b
TOPIC: Conservative approach

7. Sensitivity analysis considers

a. how sensitive the decision maker is to risk.
b. changes in the number of states of nature.
c. changes in the values of the payoffs.
d. changes in the available alternatives.
ANSWER: c
TOPIC: Sensitivity analysis

8. To find the EVSI,

a. use the EVPI to calculate sample information probabilities.
b. use indicator probabilities to calculate prior probabilities.
c. use prior and sample information probabilities to calculate revised probabilities.
d. use sample information to revise the sample information probabilities.
ANSWER: c
TOPIC: Expected value of sample information

9. If P(high) = .3, P(low) = .7, P(favorable | high) = .9, and P(unfavorable | low) = .6, then
P(favorable) =
a. .10
b. .27
c. .30
d. .55
ANSWER: d
TOPIC: Conditional probability

10. The efficiency of sample information is

a. EVSI*(100%)
b. EVSI/EVPI*(100%)
c. EVwoSI/EVwoPI*(100%)
d. EVwSI/EVwoSI*(100%)
ANSWER: b
TOPIC: Efficiency of sample information

11. Decision tree probabilities refer to

a. the probability of finding the optimal strategy
b. the probability of the decision being made
c. the probability of overlooked choices
d. the probability of an uncertain event occurring
ANSWER: d
Chapter 4 Decision Analysis 3

TOPIC: Decision tree

12. For a maximization problem, the conservative approach is often referred to as the
a. minimax approach
b. maximin approach
c. maximax approach
d. minimin approach
ANSWER: b
TOPIC: Decision making without probabilities

13. For a minimization problem, the optimistic approach is often referred to as the
a. minimax approach
b. maximin approach
c. maximax approach
d. minimin approach
ANSWER: d
TOPIC: Decision making without probabilities

14. For a maximization problem, the optimistic approach is often referred to as the
a. minimax approach
b. maximin approach
c. maximax approach
d. minimin approach
ANSWER: c
TOPIC: Decision making without probabilities

15. For a minimization problem, the conservative approach is often referred to as the
a. minimax approach
b. maximin approach
c. maximax approach
d. minimin approach
ANSWER: a
TOPIC: Decision making without probabilities

TRUE/FALSE

1. Sample information with an efficiency rating of 100% is perfect information.

ANSWER: True
TOPIC: Efficiency of sample information

2. States of nature should be defined so that one and only one will actually occur.
ANSWER: True
TOPIC: Structuring the decision process

3. Decision alternatives are structured so that several could occur simultaneously.

4 Chapter 4 Decision Analysis

ANSWER: False
TOPIC: Structuring the decision problem

4. Square nodes in a decision tree indicate that a decision must be made.

ANSWER: True
TOPIC: Decision trees

5. Circular nodes in a decision tree indicate that it would be incorrect to choose a path from
the node.
ANSWER: True
TOPIC: Decision trees

6. Risk analysis helps the decision maker recognize the difference between the expected
value of a decision alternative and the payoff that may actually occur.
ANSWER: True
TOPIC: Risk analysis

7. The expected value of an alternative can never be negative.

ANSWER: False
TOPIC: Decision making with probabilities

8. Expected value is the sum of the weighted payoff possibilities at a circular node in a
decision tree.
ANSWER: True
TOPIC: Decision making with probabilities

9. EVPI is always greater than or equal to EVSI.

ANSWER: True
TOPIC: Expected value of sample information

10. After all probabilities and payoffs are placed on a decision tree, the decision maker
calculates expected values at state of nature nodes and makes selections at decision nodes.
ANSWER: True
TOPIC: Developing a decision strategy

11. A decision strategy is a sequence of decisions and chance outcomes, where the decisions
chosen depend on the yet to be determined outcomes of chance events.
ANSWER: True
TOPIC: Decision strategy

12. EVPI equals the expected regret associated with the minimax decision.
ANSWER: True
TOPIC: Minimax regret approach

13. The expected value approach is more appropriate for a one-time decision than a repetitive
decision.
ANSWER: False
TOPIC: Decision making with probabilities
Chapter 4 Decision Analysis 5

14. Maximizing the expected payoff and minimizing the expected opportunity loss result in the
same recommended decision.
ANSWER: True
TOPIC: Minimax regret approach

15. The expected value of sample information can never be less than the expected value of
perfect information.
ANSWER: False
TOPIC: Expected value of sample information

SHORT ANSWER

1. Explain why the decision maker might feel uncomfortable with the expected value
approach, and decide to use a non-probabilistic approach even when probabilities are
available.
TOPIC: Decision making with probabilities

2. Why perform sensitivity analysis? Of what use is sensitivity analysis where good
probability estimates are difficult to obtain?
TOPIC: Sensitivity analysis

3. How can a good decision maker improve luck?

TOPIC: Introduction

4. Use a diagram to compare EVwPI, EVwoPI, EVPI, EVwSI, EVwoSI, and EVSI.
TOPIC: Expected value of sample information.

5. Show how you would design a spreadsheet to calculate revised probabilities for two states
of nature and two indicators.
TOPIC: Decision analysis and spreadsheets

PROBLEMS

1. Jim has been employed at Gold Key Realty at a salary of $2,000 per month during the past
year. Because Jim is considered to be a top salesman, the manager of Gold Key is offering
him one of three salary plans for the next year: (1) a 25% raise to $2,500 per month; (2) a
base salary of $1,000 plus $600 per house sold; or, (3) a straight commission of $1,000
per house sold. Over the past year, Jim has sold up to 6 homes in a month.
6 Chapter 4 Decision Analysis

a. Compute the monthly salary payoff table for Jim.

b. For this payoff table find Jim's optimal decision using: (1) the conservative
approach, (2) minimax regret approach.
c. Suppose that during the past year the following is Jim's distribution of home sales.
If one assumes that this a typical distribution for Jim's monthly sales, which salary
plan should Jim select?

Home Sales Number of Months

0 1
1 2
2 1
3 2
4 1
5 3
6 2
TOPIC: Decision making with and without probabilities

2. East West Distributing is in the process of trying to determine where they should schedule
next year's production of a popular line of kitchen utensils that they distribute.
Manufacturers in four different countries have submitted bids to East West. However, a
pending trade bill in Congress will greatly affect the cost to East West due to proposed
tariffs, favorable trading status, etc.
After careful analysis, East West has determined the following cost breakdown for the four
manufacturers (in $1,000's) based on whether or not the trade bill passes:

Bill Passes Bill Fails

Country 260 210
A
Country 320 160
B
Country 240 240
C
Country 275 210
D

a. If East West estimates that there is a 40% chance of the bill passing, which country
should they choose for manufacturing?
b. Over what range of values for the "bill passing" will the solution in part (a) remain
optimal?
TOPIC: Decision making with probabilities

3. Transrail is bidding on a project that it figures will cost $400,000 to perform. Using a
25% markup, it will charge $500,000, netting a profit of $100,000. However, it has been
learned that another company, Rail Freight, is also considering bidding on the project. If
Rail Freight does submit a bid, it figures to be a bid of about $470,000. Transrail really
wants this project and is considering a bid with only a 15% markup to $460,000 to ensure
winning regardless of whether or not Rail Freight submits a bid.
Chapter 4 Decision Analysis 7

a. Prepare a profit payoff table from Transrail's point of view.

b. What decision would be made if Transrail were conservative?
c. If Rail Freight is known to submit bids on only 25% of the projects it considers,
what decision should Transrail make?
d. Given the information in (c), how much would a corporate spy be worth to
Transrail to find out if Rail Freight will bid?
TOPIC: Decision making with and without probabilities

4. The Super Cola Company must decide whether or not to introduce a new diet soft drink.
Management feels that if it does introduce the diet soda it will yield a profit of $1 million if
sales are around 100 million, a profit of $200,000 if sales are around 50 million, or it will
lose $2 million if sales are only around 1 million bottles. If Super Cola does not market
the new diet soda, it will suffer a loss of $400,000.

a. Construct a payoff table for this problem.

b. Construct a regret table for this problem.
c. Should Super Cola introduce the soda if the company: (1) is conservative; (2) is
optimistic; (3) wants to minimize its maximum disappointment?
d. An internal marketing research study has found P(100 million in sales) = 1/3;
P(50 million in sales) = 1/2; P(1 million in sales) = 1/6. Should Super Cola
introduce the new diet soda?
e. A consulting firm can perform a more thorough study for $275,000. Should
management have this study performed?
TOPIC: Decision making with and without probabilities

5. Super Cola is also considering the introduction of a root beer drink. The company feels
that the probability that the product will be a success is .6. The payoff table is as follows:

Success (s1) Failure (s2)

Produce (d1) $250,000 -$300,000
Do Not Produce -$ 50,000 -$ 20,000
(d2)

The company has a choice of two research firms to obtain information for this product.
Stanton Marketing has market indicators, I1 and I2 for which P(I1 s1) = .7 and P(I1 s2)
= .4. New World Marketing has indicators J1 and J2 for which P(J1 s1) = .6 and
P(J1 s2) = .3.

a. What is the optimal decision if neither firm is used? Over what probability of
success range is this decision optimal?
b. What is the EVPI?
c. Find the EVSIs and efficiencies for Stanton and New World.
d. If both firms charge $5,000, which firm should be hired?
e. If Stanton charges $10,000 and New World charges $4,000, which firm should
Super Cola hire? Why?
8 Chapter 4 Decision Analysis

TOPIC: Computing branch probabilities

6. Dollar Department Stores has just acquired the chain of Wenthrope and Sons Custom
Jewelers. Dollar has received an offer from Harris Diamonds to purchase the Wenthrope
store on Grove Street for $120,000. Dollar has determined probability estimates of the
store's future profitability, based on economic outcomes, as: P($80,000) = .2,
P($100,000) = .3, P($120,000) = .1, and P($140,000) = .4.

a. Should Dollar sell the store on Grove Street?

b. What is the EVPI?
c. Dollar can have an economic forecast performed, costing $10,000, that produces
indicators I1 and I2, for which P(I1 80,000) = .1; P(I1 100,000) = .2;
P(I1 120,000) = .6; P(I1 140,000) = .3. Should Dollar purchase the forecast?
TOPIC: Computing branch probabilities

7. An appliance dealer must decide how many (if any) new microwave ovens to order for
next month. The ovens cost $220 and sell for $300. Because the oven company is
coming out with a new product line in two months, any ovens not sold next month will
have to be sold at the dealer's half price clearance sale. Additionally, the appliance dealer
feels he suffers a loss of $25 for every oven demanded when he is out of stock. On the
basis of past months' sales data, the dealer estimates the probabilities of monthly demand
(D) for 0, 1, 2, or 3 ovens to be .3, .4, .2, and .1, respectively.

The dealer is considering conducting a telephone survey on the customers' attitudes

towards microwave ovens. The results of the survey will either be favorable (F),
unfavorable (U) or no opinion (N). The dealer's probability estimates for the survey
results based on the number of units demanded are:

P(F D = 0) = .1 P(F D = 2) .3 P(U D = 0) = .8 P(U D = 2) = .1

P(F D = 1) = .2 P(F D = 3) .9 P(U D = 1) = .3 P(U D = 3) = .1

a. What is the dealer's optimal decision without conducting the survey?

b. What is the EVPI?
c. Based on the survey results what is the optimal decision strategy for the dealer?
d. What is the maximum amount he should pay for this survey?
TOPIC: Computing branch probabilities

8. Lakewood Fashions must decide how many lots of assorted ski wear to order for its three
stores. Information on pricing, sales, and inventory costs has led to the following payoff
table, in thousands.

Demand
Order Size Low Medium High
1 lot 12 15 15
2 lots 9 25 35
3 lots 6 35 60
Chapter 4 Decision Analysis 9

a. What decision should be made by the optimist?

b. What decision should be made by the conservative?
c. What decision should be made using minimax regret?
TOPIC: Decision making without probabilities

9. The table shows both prospective profits and losses for a company, depending on what
decision is made and what state of nature occurs. Use the information to determine what
the company should do.

State of Nature
Decision s1 s2 s3
d1 30 80 -30
d2 100 30 -40
d3 -80 -10 120
d4 20 20 20

a. if an optimistic strategy is used.

b. if a conservative strategy is used.
c. if minimax regret is the strategy.
TOPIC: Decision making without probabilities

10. A payoff table is given as

State of Nature
Decisio s1 s2 s3
n
d1 10 8 6
d2 14 15 2
d3 7 8 9

a. What decision should be made by the optimistic decision maker?

b. What decision should be made by the conservative decision maker?
c. What decision should be made under minimax regret?
d. If the probabilities of s1, s2, and s3 are .2, .4, and .4, respectively, then what
decision should be made under expected value?
e. What is the EVPI?
TOPIC: Decision making with and without probabilities

11. A payoff table is given as

State of Nature
Decisio s1 s2 s3
n
d1 250 750 500
d2 300 -250 1200
d3 500 500 600
10 Chapter 4 Decision Analysis

a. What choice should be made by the optimistic decision maker?

b. What choice should be made by the conservative decision maker?
c. What decision should be made under minimax regret?
d. If the probabilities of d1, d2, and d3 are .2, .5, and .3, respectively, then what choice
should be made under expected value?
e. What is the EVPI?
TOPIC: Decision making with and without probabilities

12. A decision maker has developed the following decision tree. How sensitive is the choice
between N and P to the probabilities of states of nature U and V?

N
U
50
P
V
5

Topic: Sensitivity analysis

13. If p is the probability of Event 1 and (1-p) is the probability of Event 2, for what values of
p would you choose A? B? C? Values in the table are payoffs.

Choice/Event Event 1 Event 2

A 0 20
B 4 16
C 8 0
Chapter 4 Decision Analysis 11

TOPIC: Sensitivity analysis

14. Fold back the decision tree and state what strategy should be followed.

S .55
300
H

C .25 T .45
100

J
180
A D .35 U .25
20 10
K
V .75
80
E .40 M
50
W .5
-30
N

F .2 X .5
-50 120
B
Y .4
500
Q

G .8 Z .6
-200
AA .3
150
R BB .4
100
CC .3
50

TOPIC: Expected value and decision trees

15. Fold back this decision tree. Clearly state the decision strategy you determine.
12 Chapter 4 Decision Analysis

I .82
100
D .42
J .18
E .24 -40
A P .63
-50 30
K
F .34
Q .37
60
L
B 35
25
M
50
G .2 R .3
-100
N S .3
C 120
T .4
150

H .8
10
TOPIC: Expected value and decision trees

16. If sample information is obtained, the result of the sample information will be either
positive or negative. No matter which result occurs, the choice to select option A or
option B exists. And no matter which option is chosen, the eventual outcome will be good
or poor. Complete the table.

Sample States Prior Conditional Joint Posterior

Result of Probabilitie Probabilities Probabiliti Probabiliti
Nature s es es
Positive good .7 P(positive | good) =
.8
poor .3 P(positive | poor) =
.1
Negativ good .7 P(negative | good)
e =
poor .3 P(negative | poor)
=

TOPIC: Posterior probabilities

17. Use graphical sensitivity analysis to determine the range of values of the probability of
state of nature s1 over which each of the decision alternatives has its largest expected
value.

State of Nature
Decision s1 s2
d1 8 10
d2 4 16
d3 10 0
Chapter 4 Decision Analysis 13

TOPIC: Graphical sensitivity analysis

18. Dollar Department Stores has received an offer from Harris Diamonds to purchase
Dollars store on Grove Street for $120,000. Dollar has determined probability estimates
of the store's future profitability, based on economic outcomes, as: P($80,000) = .2,
P($100,000) = .3, P($120,000) = .1, and P($140,000) = .4.
a. Should Dollar sell the store on Grove Street?
b. What is the EVPI?
c. Dollar can have an economic forecast performed, costing $10,000, that produces
indicators I1 and I2, for which P(I1 80,000) = .1; P(I1 100,000) = .2;
P(I1 120,000) = .6; P(I1 140,000) = .3. Should Dollar purchase the forecast?

TOPIC: Posterior probabilities

SOLUTIONS TO PROBLEMS

1. a. There are three decision alternatives (salary plans) and seven states of nature (the
number of houses sold monthly).

PAYOFF Number of Homes Sold

TABLE
0 1 2 3 4 5 6
Salary Plan I 2500 2500 2500 2500 2500 2500 2500
Salary Plan II 1000 1600 2200 2800 3400 4000 4600
Salary Plan III 0 1000 2000 3000 4000 5000 6000

b. (1) Conservative Approach (Maximin): Plan I

REGRET Number of Homes Sold

TABLE
0 1 2 3 4 5 6
Salary Plan I 0 0 0 500 1500 2500 3500
Salary Plan II 1500 900 300 200 600 1000 1400
Salary Plan III 2500 1500 500 0 0 0 0
(2) Minimax Regret Approach: Plan II

c. Use the relative frequency method for determining the probabilities.

Using the EV approach: EV(Plan I) = 2500, EV(Plan II) = 3050, EV(Plan III) =
3417;
Choose Plan III.

2. a. Using EV approach: EV(A) = 230, EV(B) = 224, EV(C) = 240; Choose Country
B (lowest EV)
14 Chapter 4 Decision Analysis

b. As long as the probability of the bill passing is less than .455, East West should
choose Country B.

Expected
Value
300

250 C
|
B |
200 |
A |
|
150 |
|
|
100 |
|
|
|
50 |
|
|
|
0 | Probability
0 .1 .2 .3 .4 | .5 .6 .7 .8 .9 1.0
.455

3. a.
Rail Freight
Transrail Bid $470,000 Doesn't Bid
Bid $0 $100,000
$500,000
Bid $60,000 $ 60,000
$460,000

b. Bid $460,000
c. Bid $500,000
d. $15,000

4. a.
PAYOFF Sales ($millions)
TABLE
100 50 1
Introduce $1,000,000 $200,000 -$2,000,000
Do Not -$400,000 -$400,000 -$400,000
Introduce
b.
REGRET Sales ($millions)
TABLE
100 50 1
Introduce $0 $0 $1,600,000
Chapter 4 Decision Analysis 15

Do Not $1,400,000 $600,000 $0

Introduce
c. (1) do not introduce; (2) introduce; (3) do not introduce
d. Yes
e. No

5. a. Introduce root beer; p < .483

b. EVPI = $112,000
NOTE: The answers to (c)-(e) are very sensitive to roundoff error.
Figures in parentheses are for two decimal places only.
c. Stanton: EVSI = $13,200 ($11,862)
Efficiency = .118 (.106)
New World: EVSI = $6,400 ($6,424)
Efficiency = .057 (.057)
d. Hire Stanton (Stanton)
e. Hire New World (Stanton)

6. a. Yes, Dollar should sell store

b. EVPI = $8,000
c. No; survey cost exceeds EVPI

7.
Demand For Ovens
Ovens Ordered 0 1 2 3
0 0 -25 -50 -75
1 -70 80 55 30
2 -140 10 160 135
3 -210 -60 90 240

a. Order one oven: EV = $25.00

b. EVPI = $63.00
c. Favorable: order 2; Unfavorable: order 0; No opinion: order 1
d. EVSI = $9.10

8. a. 3 lots
b. 1 lot
c. 3 lots

Regret table:
Demand Maximu
Order
m
Size Low Medium High
Regret
1 lot 0 20 45 45
2 lots 3 10 25 25
3 lots 6 0 0 6
16 Chapter 4 Decision Analysis

9. a. d3
b. d1
c. d4

Regret table:
State of Nature Maximu
s1 s2 s3 m
Decision
Regret
d1 70 0 150 150
d2 0 50 160 160
d3 180 90 0 180
d4 80 60 100 100

10. a. d2
b. d3
c. a three way tie
d. EV(d1) = 7.6
EV(d2) = 9.6 (the best)
EV(d3) = 8.2
e. EVPI = 12.4 - 9.6 = 2.8

11. a. d2
b. d3
c. d1
d. EV(d1) = 695 (the best)
EV(d2) = 385
EV(d3) = 500
e. EVPI = 925 - 695 = 230

12. Choose N if p < .78.

13. Choose A if p .5, choose B is .5 p .8, and choose C if p .8.

Chapter 4 Decision Analysis 17

14.

S .55
300
H 210

C .25 210 T .45 100

J
84.5 180
A D .35 U .25
20 62.5 10
K
V .75 80
E .40 M
50
62.5 W .5
N 45 -30

F .2 X .5
70 -50 120
B Y .4
500
Q 80

G .8 Z .6 -200
AA .3
100 150
R 100 BB .4
100
CC .3 50

Strategy: Select A. If C happens, select H. If D happens, you are done. If E happens,

select K.
18 Chapter 4 Decision Analysis

15.
I .82
100
D .42 74.8
J .18
A 33.39 E .24 -40 P .63
-50 30
K 41.1
F .34 Q .37
L 60
B 35
25
M
50
G .2 R .3
-100
N 66 S .3
C 21.2 120
T .4
150

H .8
10
Choose A. If F happens, choose K.

16.
States Prior Joint
Sample Conditional Posterior
of Probabilitie Probabiliti
Result Probabilities Probabilities
Nature s es
P(positive | good) = .
Positive good .7 .56 .9492
8
P(positive | poor)
poor .3 .03 .0508
= .1
Negativ P(negative | good)
good .7 .14 .3415
e = .2
P(negative | poor) =
poor .3 .27 .6585
.9

17. EV(d1) and EV(d2) intersect at p = .6. EV(d1) and EV(d3) intersect at p = .8333.
Therefore when 0 p .6, choose d2. When .6 p .8333, choose d1. When p .8333,
choose d3.

16 16
d2
12 12
d1 d3
8 8

4 4

p=0 .6 .8333 p = 1
Chapter 4 Decision Analysis 19

18. a. Yes, Dollar should sell store.

b. EVPI = $8,000
c. No; survey cost exceeds EVPI.