01 Decision Analysis
01 Decision Analysis
Decision Making
1 2
What is involved in making a good decision? 1. Clearly define the problem at hand.
Decision theory is an analytic and systematic 2. List the possible alternatives.
approach to the study of decision making. 3. Identify the possible outcomes or states of
A good decision is one that is based on logic, nature.
considers all available data and possible 4. List the payoff (typically profit) of each
alternatives, and the quantitative approach combination of alternatives and outcomes.
described here.
5. Select one of the mathematical decision theory
models.
6. Apply the model and make your decision.
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-3 AS PRENTICE HALL 3-4
3 4
1
12/13/19
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-5 AS PRENTICE HALL 3-6
5 6
Decision Table with Conditional Values for Type 1: Decision making under certainty
Thompson Lumber ◦ The decision maker knows with certainty the
consequences of every alternative or decision
STATE OF NATURE choice.
FAVORABLE UNFAVORABLE Type 2: Decision making under uncertainty
ALTERNATIVE MARKET ($) MARKET ($)
◦ The decision maker does not know the
Construct a large plant 200,000 –180,000 probabilities of the various outcomes.
Construct a small plant 100,000 –20,000 Type 3: Decision making under risk
◦ The decision maker knows the probabilities of
Do nothing 0 0
the various outcomes.
Table 3.1
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-7 AS PRENTICE HALL 3-8
7 8
2
12/13/19
There are several criteria for making decisions Used to find the alternative that maximizes the
under uncertainty: maximum payoff.
1. Maximax (optimistic) n Locate the maximum payoff for each alternative.
n Select the alternative with the maximum number.
2. Maximin (pessimistic)
STATE OF NATURE
3. Criterion of realism (Hurwicz) FAVORABLE UNFAVORABLE MAXIMUM IN
ALTERNATIVE MARKET ($) MARKET ($) A ROW ($)
4. Equally likely (Laplace) Construct a large
plant 200,000 –180,000 200,000
5. Minimax regret Construct a small
Maximax
100,000 –20,000 100,000
plant
Do nothing 0 0 0
Table 3.2
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-9 AS PRENTICE HALL 3-10
9 10
Used to find the alternative that maximizes This is a weighted average compromise
the minimum payoff. between optimism and pessimism.
n Locate the minimum payoff for each alternative. n Select a coefficient of realism a, with 0≤α≤1.
n Select the alternative with the maximum n A value of 1 is perfectly optimistic, while a
number. value of 0 is perfectly pessimistic.
STATE OF NATURE
n Compute the weighted averages for each
FAVORABLE UNFAVORABLE MINIMUM IN
ALTERNATIVE MARKET ($) MARKET ($) A ROW ($) alternative.
Construct a large n Select the alternative with the highest value.
plant 200,000 –180,000 –180,000
Weighted average = a(maximum in row)
Construct a small
plant
100,000 –20,000 –20,000 + (1 – a)(minimum in row)
Do nothing 0 0 0
Table 3.3 Maximin
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-11 AS PRENTICE HALL 3-12
11 12
3
12/13/19
n For the large plant alternative using a = 0.8: Considers all the payoffs for each alternative
(0.8)(200,000) + (1 – 0.8)(–180,000) = 124,000 n Find the average payoff for each alternative.
n For the small plant alternative using a = 0.8: n Select the alternative with the highest average.
(0.8)(100,000) + (1 – 0.8)(–20,000) = 76,000
STATE OF NATURE STATE OF NATURE
CRITERION FAVORABLE UNFAVORABLE ROW
FAVORABLE UNFAVORABLE OF REALISM ALTERNATIVE MARKET ($) MARKET ($) AVERAGE ($)
ALTERNATIVE MARKET ($) MARKET ($) (a = 0.8) $ Construct a large
plant 200,000 –180,000 10,000
Construct a large
plant 200,000 –180,000 124,000
Realism Construct a small 100,000 –20,000 40,000
Construct a small plant
plant 100,000 –20,000 76,000 Equally likely
Do nothing 0 0 0
Do nothing 0 0 0
Table 3.5
Table 3.4
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-13 AS PRENTICE HALL 3-14
13 14
Table 3.6
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-15 AS PRENTICE HALL 3-16
15 16
4
12/13/19
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-17 AS PRENTICE HALL 3-18
17 18
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-19 AS PRENTICE HALL 3-20
19 20
5
12/13/19
Expected Value of Perfect Information (EVPI) Expected Value of Perfect Information (EVPI)
EVPI places an upper bound on what you should pay for additional
information. Suppose Scientific Marketing, Inc. offers analysis
that will provide certainty about market conditions
EVPI = EVwPI – Maximum EMV
(favorable).
Additional information will cost $65,000.
EVwPI is the long run average return if we have perfect
information before a decision is made. Should Thompson Lumber purchase the
information?
EVwPI = (best payoff for first state of nature)
x (probability of first state of nature)
+ (best payoff for second state of nature)
x (probability of second state of nature)
+ … + (best payoff for last state of nature)
x (probability of last state of nature)
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-21 AS PRENTICE HALL 3-22
21 22
Expected Value of Perfect Information (EVPI) Expected Value of Perfect Information (EVPI)
Decision Table with Perfect Information The maximum EMV without additional information is
$40,000.
STATE OF NATURE EVPI = EVwPI – Maximum EMV
FAVORABLE UNFAVORABLE = $100,000 - $40,000
ALTERNATIVE MARKET ($) MARKET ($) EMV ($)
Construct a large = $60,000
plant 200,000 -180,000 10,000
Construct a small
plant
100,000 -20,000 40,000 So the maximum Thompson
Do nothing 0 0 0 should pay for the additional
information is $60,000.
With perfect
200,000 0 100,000
information
EVwPI Therefore, Thompson should not
Probabilities 0.5 0.5 pay $65,000 for this information.
Table 3.10
COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING COPYRIGHT ©2012 PEARSON EDUCATION, INC. PUBLISHING
AS PRENTICE HALL 3-23 AS PRENTICE HALL 3-24
23 24
6
12/13/19
25 26
30 31
7
12/13/19
PROBLEM SOLVING:
Mickey’s Investments
32