Scribd
Upload
61 views25 pages

Decision Tree & Utility Theory

1) Decision trees can be used to model multi-stage decision problems that involve uncertainty. They contain decision nodes, chance nodes, and end nodes with associated payoffs. 2) The five steps of decision tree analysis are to define the problem, structure the tree, assign probabilities, estimate payoffs, and compute expected monetary values to identify the optimal decision. 3) Utility theory is a framework for understanding risk preferences where individuals' utilities of outcomes are measured on a scale from 0 to 1, allowing calculation of expected utility to compare risky and riskless options.

Uploaded by

Thien Tran
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
61 views25 pages

Decision Tree & Utility Theory

1) Decision trees can be used to model multi-stage decision problems that involve uncertainty. They contain decision nodes, chance nodes, and end nodes with associated payoffs. 2) The five steps of decision tree analysis are to define the problem, structure the tree, assign probabilities, estimate payoffs, and compute expected monetary values to identify the optimal decision. 3) Utility theory is a framework for understanding risk preferences where individuals' utilities of outcomes are measured on a scale from 0 to 1, allowing calculation of expected utility to compare risky and riskless options.

Uploaded by

Thien Tran
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 25

Chapter 4

Decision Tree & Utility Theory


Decision Trees
A decision node

A state of nature node


Five Steps of Decision Tree Analysis
• Define the problem
• Structure or draw the decision tree
• Assign probabilities to the states of nature
• Estimate payoffs for each possible combination
of alternatives and states of nature
• Compute expected monetary values (EMVs) for
each state of nature node. At decision nodes,
the alternative with the best EMV is selected.
EMV (node1)=$10,000 = (0.50)($200,000)+(0.5)(-180,000)=$10,000
Payoffs
Favor-Market (0.5)
$200,000

1 Unfavor-market (0.5)
$-180,000
Large plant
Favor-Market (0.5)
$100,000
Small plant
2 Unfavor-Market (0.5)
$-20,000
EMV (node2)=$40,000

Do nothing

$0

Build the small plant


Modeling in the real world
• Defining the problem
• Developing a model
• Acquiring input data
• Developing a solution
• Testing the solution
• Analyzing the results
• Implementing the results
A more complex decision for Thompson
Lumber
• Now, Thompson has two decisions to make,
with the second decision dependent on the
outcome of the first. In this case, he has the
option to conduct a survey at a cost of
$10,000. The information from his survey
could help to decide whether to construct a
large plant, a small plant, or do nothing.
(cont)
• EMV (conduct study) = $49,200 > EMV (not
conduct)= $40,000  best choice is to
conduct study.
– If the results are favorable  build a large plant
– If the results are unfavorable  build a small plant
Expected Value of Sample Information

• EVSI = (expected value of best decision with


sample info, assuming no cost to gather it) –
(expected value of best decision without
sample info) = (EV with sample info + cost) –
(EV without sample info)
= ($49,200 + $10,000) – $40,000 = $19,200
i.e. the cost of survey should not exceed $19,200
(in this case it is $10,000 OK)
How probability values are estimated by
Bayesian analysis
• In the Thomson Lumber case, we made the
assumption that the following four conditional
probabilities were known:
P(favor-market FM/survey positive SP) = 0.78
P(UM/SP) = 0.22 (=1-.78)
P(FM/SN) = 0.27
P(UM/SN) =0.73 (=1-.27)
How to derive these values with Bayes’ theorem?
(cont)
• In discussion with a market researcher, Thompson
knows that the survey can be positive (SP) or
negative (SN).
• The researcher told that, statistically, of all new
products with a FM, market survey were correct in
70% of the time, 30% of the time the surveys were
false.
• On the other hand, when there was an unfavorable
market for a new product, 80% of the surveys were
correctly predicted, 20% were false.
(cont)
•  P(A|B) =
Utility theory
• Whether should you keep the lottery ticket
with a chance of 50/50 winning $5 mil, or take
$2 mil for sure?
• Most people take $2 mil. What about you?
Utility theory (cont)
Measuring utility & constructing a utility
curve
• U(best outcome) =1
• U(worst outcome) = 0
(cont)
(cont)
• E.g. Jane would like to construct her utility curve to reveal
her preference for money between $0 and $10,000.
• A utility curve is a graph that plots utility value versus
monetary value.
• If the money is invested in the bank, in 3 years Jane would
receive $5,000. If she invested in the real estate, after 3 years
she could receive either nothing or $10,000. Jane thinks that
unless she has the chance of 80% getting $10,000 from the
real estate deal, she would prefer to deposit her money into
the bank. Thus, her utility of $5,000 is
U($5,000) = 0.8
P = 0.80 $10,000
U($10,000) = 1.0

$0
1-P = 0.20
Real estate U($0.00) = 0.0

Bank $5,000
U($5,000) = p = 0.8

U($5,000) = pU($10,000) + (1-p) U($0) = 0.8*1 + 0*0.0 = 0.8


Other utility values can be conducted in the same manner
Preferences for Risk

Risk seeker Risk avoider Risk indifference


Utility as a Decision Making Criterion

• E.g. Mark loves to gamble. He plays the game


tossing thumbtacks. If the point on the
thumbtack is facing up after landing, Mark
wins $10,000, if down then Mark loses
$10,000. Mark believes that 45% chance of
winning $10,000 and 55% of losing $10,000.
Alternative 2 is not to gamble. What should
Mark do? Suppose his utility curve as in the
next fig.
U(-$10,000) = 0.05
U($0) = 0.15
U($10,000) = 0.30
E(play) = 0.45*0.30 + 0.55*0.05 = 0.162
E(not play) = 0.15
End

You might also like