Documente Academic
Documente Profesional
Documente Cultură
Decision Analysis
(Chapter 3)
Learning Objectives
Students will be able to:
List the steps of the decision-making process.
Describe the types of decision-making
environments.
Make decisions under uncertainty.
Use probability values to make decisions
under risk.
Chapter Outline
3.1
3.2
3.3
3.4
3.5
3.6
3.7
Introduction
The Six Steps in Decision Theory
Types of Decision-Making Environments
Decision Making under Uncertainty
Decision Making under Risk
Decision Trees
How Probability Values Are Estimated by
Bayesian Analysis
3.8 Utility Theory
Introduction
Decision theory is an analytical and
systematic way to tackle problems.
A good decision is based on logic.
List alternatives
1.
2.
3.
Identify outcomes
List payoffs
Select a model
Favorable
Market ($)
Unfavorable
Market ($)
200,000
-180,000
100,000
-20,000
Do nothing
Optimistic (Maximax)
Pessimistic (Maximin)
Criterion of Realism (Hurwicz)
Equally likely (Laplace Criterion)
Minimax Regret (Opportunity Loss or Regret)
Alternative
Favorable
Market ($)
Unfavorable
Market ($)
200,000
-180,000
100,000
-20,000
Do nothing
Maximum in
Row
Favorable
Market ($)
Unfavorable
Market ($)
Construct a large
plant
200,000
-180,000
Construct a small
plant
100,000
-20,000
100,000
Do nothing
200,000
Optimistic
(Maximax)
Minimum in
Row
Favorable
Market ($)
Unfavorable
Market ($)
Construct a large
plant
200,000
-180,000
-180,000
Construct a small
plant
100,000
-20,000
-20,000
Do nothing
Pessimistic
(Maximin)
Unfavorable
Market ($)
Weighted
Average in
Row
( = 0.8)
Construct a large
plant
200,000
-180,000
124,000
Construct a small
plant
100,000
-20,000
Alternative
Do nothing
Criterion
of Realism
76,000 (Hurwicz)
Average in
Row
Favorable
Market ($)
Unfavorable
Market ($)
Construct a large
plant
200,000
-180,000
10,000
Construct a small
plant
100,000
-20,000
40,000
Do nothing
Equally
likely
(Laplace)
Construct a
large plant
200,000 200,000 = 0
Construct a
small plant
Do nothing
200,000 0 = 200,000
0 (180,000) = 180,000
00=0
Construct a large
plant
Favorable Market
($)
Unfavorable
Market ($)
Maximum in
Row
Minimax
180,000 Regret
(Opportunity
Loss)
180,000
Construct a
small plant
100,000
20,000
100,000
Do nothing
200,000
200,000
Favorable
Market ($)
200,000
Unfavorable
Market ($)
180,000
(200,000)(0.5) +
(180,000)(0.5) = 10,000
(100,000)(0.5) +
(20,000)(0.5) =
Construct a
small plant
100,000
20,000
Do nothing
Probabilities
0.50
0.50
40,000
(0)(0.5) + (0)(0.5) = 0
Favorable
Market, S1 ($)
Unfavorable
Market, S2 ($)
200,000
-180,000
10,000
100,000
-20,000
40,000
0.50
0.50
Do nothing
Probabilities
With Perfect
Information
Best payoff
in S1
200,000
Best payoff in
EVwPI = (200,000)(0.5)
S2
+ (0)(0.5) = 100,000
0
Thompson Lumber:
Payoff Table
State of Nature
Alternative
Favorable Market
($)
Unfavorable
Market ($)
Construct a large
plant
200,000
-180,000
Construct a small
plant
100,000
-20,000
Do nothing
Probabilities
0.50
0.50
Unfavorable Market
($)
200,000 200,000
0- (-180,000)
Construct a small
plant
200,000 - 100,000
0 (-20,000)
Do nothing
200,000 - 0
0-0
Probabilities
0.50
0.50
Thompson Lumber:
Opportunity Loss Table
State of Nature
Alternative
Favorable Market
($)
Unfavorable
Market ($)
Construct a large
plant
180,000
Construct a small
plant
100,000
20,000
Do nothing
200,000
Probabilities
0.50
0.50
EOL
(0.50)*$0 +
$90,000
(0.50)*($180,000)
(0.50)*($100,000) $60,000
+ (0.50)(*$20,000)
(0.50)*($200,000) $100,000
+ (0.50)*($0)
Note:
1. The minimum EOL and maximum EMV will suggest the same decision.
2. The value of EVPI will always equal to the minimum EOL value.
Thompson Lumber:
Sensitivity Analysis
Let P = probability of favorable market
EMV(Large Plant):
= $200,000P + (-$180,000)(1-P) = 380,000P - 180,000
EMV(Small Plant):
= $100,000P + (-$20,000)(1-P) = 120,000P - 20,000
EMV(Do Nothing):
= $0P + 0(1-P) = 0
Thompson Lumber:
Sensitivity Analysis (continued)
Five Steps to
Decision Tree Analysis
1.
2.
3.
4.
Favorable Market
Unfavorable Market
A Decision
Node
Construct
Small Plant
Favorable Market
2
Unfavorable Market
A Decision
Node
Construct
Small Plant
Outcomes
Payoffs
Favorable (0.5)
Market
$200,000
Unfavorable (0.5)
Market
-$180,000
Favorable (0.5)
Market
$100,000
Unfavorable (0.5)
Market
-$20,000
A Decision
Node
1
EMV
=$10,000
Construct
Small Plant
2
EMV
=$40,000
$200,000
Unfavorable (0.5)
Market
-$180,000
Favorable (0.5)
Market
Unfavorable (0.5)
Market
$100,000
-$20,000
0
Thompsons Decision:
A More Complex Problem
John Thompson has the opportunity of obtaining a market survey
that will give additional information on the probable state of
nature. Results of the market survey will likely indicate there is a
percent change of a favorable market. Historical data show market
surveys accurately predict favorable markets 78 % of the time.
P(Fav. Mkt | Results Favourable) = 0.78
Likewise, if the market survey predicts an unfavorable market,
there is a 73 % chance of its occurring.
P(Unfav. Mkt | Results Negative) = 0.73
Assumed the cost of conduct market survey is $10,000, and is
deducted from each payoff under Conduct Market Survey.
Now that we have redefined the problem (Step 1), lets use this
additional data and redraw Thompsons decision tree (Step 2).
EVSI ==
Sensitivity Analysis
How sensitive are the decisions to changes in the
probabilities?
e.g. If the probability of a favorable survey were
less than the current value (0.45), would the survey
still be selected? How low would this have to be to
cause a change in the decision?
p =
$37,600
$104,000
= 0.36
Prior
probabilities
Bayes Theorem
Posterior
probabilities
Bayesian Analysis
The probabilities of a favorable / unfavorable state of nature can be
obtained by analyzing the Market Survey Reliability in Predicting
Actual States of Nature.
STATE OF NATURE
RESULT OF
SURVEY
FAVORABLE MARKET
(FM)
UNFAVORABLE MARKET
(UM)
Positive (predicts
favorable market
for product)
Negative (predicts
unfavorable
market for
product)
CONDITIONAL
PROBABILITY
P(SURVEY
POSITIVE | STATE
OF NATURE)
PRIOR
PROBABILITY
FM
0.70
X 0.50
0.35
0.35/0.45 = 0.78
UM
0.20
X 0.50
0.10
0.10/0.45 = 0.22
0.45
1.00
STATE OF
NATURE
JOINT
PROBABILITY
P(STATE OF
NATURE |
SURVEY
POSITIVE)
CONDITIONAL
PROBABILITY
P(SURVEY
NEGATIVE | STATE
OF NATURE)
PRIOR
PROBABILITY
FM
0.30
X 0.50
0.15
0.15/0.55 =
0.27
UM
0.80
X 0.50
0.40
0.40/0.55 =
0.73
0.55
STATE OF
NATURE
JOINT
PROBABILITY
P(STATE OF
NATURE |
SURVEY
NEGATIVE)
1.00
1p
Best outcome
Utility = 1
Worst outcome
Utility = 0
Other outcome
Utility = ??
(1 p)= 0.20
$10,000
U($10,000) = 1.0
0
U(0) = 0
$5,000
U($5,000) = p = 0.80
Hence, Janes Utility for $5,000 = U($5,000) = p
= p*U($10,000) + (1 p)*U($0)
= (0.8)(1) + (0.2)(0) = 0.8
U ($10,000) = 1.0
U ($7,000) = 0.90
U ($5,000) = 0.80
0.7
Utility
0.6
0.5
U ($3,000) = 0.50
0.4
0.3
0.2
0.1
U ($0) = 0
$0
$1,000
$3,000
$5,000
Monetary Value
$7,000
$10,000