Chapter Four

Decision Theories
 Learning Objectives

In this chapter, we will be able to:

1. Understand the steps of the decision-making process.

2. Describe the types of decision-making environments.
3. Make decisions under uncertainty.
4. Use probability values to make decisions under risk.
5. Develop accurate and useful decision trees.
6. Compute expected value and expected opportunity loss.
7. Assess the value of information.
 Introduction

• Decision theory is a systematic and scientific

approach to the study of decision making.
• What is involved in making a good decision?
• A good decision is one that is based on logic,
considers all available data and possible alternatives,
and identify the best alternative by applying the
quantitative approach.
19-2

Introduction

• Classical statistics focuses on estimating a parameter,

such as the population mean, constructing confidence
intervals, or hypothesis testing.
• Decision Theory is concerned with determining
which decision, from a set of possible decisions, is
optimal for a particular situation.
 Types of Decision-Making Environments
Type 1: Decision making under certainty
· The decision maker knows with certainty the
consequences of every alternative or decision
choice.
Type 2: Decision making under uncertainty
· The decision maker does not know the
probabilities of the various outcomes.
Type 3: Decision making under risk
· The decision maker knows the probabilities of
the various outcomes.
 The Steps in Decision Theory
1. Clearly define the problem at hand.
2. List the possible alternatives.
3. Identify the possible outcomes or states of nature.
4. List the payoff (typically profit) of each combination of
alternatives and outcomes.
5. Select one of the mathematical decision theory models.
6. Apply the model and make your decision.
19-3

Elements of a Decision

• There are three components to any decision-making

situation:
· The available choices (alternatives or acts)
· The states of nature, which are not under the control
of the decision maker - uncontrollable future events
· The payoffs – Benefits accrued from each
combination of decision alternative and state of
nature
19-4

Payoff Table and Expected Payoff

• A Payoff Table is a listing of all possible

combinations of decision alternatives and states of
nature.
• The Expected Payoff or the Expected Monetary
Value (EMV) is the expected value of each
decision criterion.
19-5
Calculating the EMV
• Let Ai be the ith decision alternative.
• Let P(Sj) be the probability of the jth state of nature.
• Let V(Ai, Sj) be the value of the payoff for the combination of
decision alternative Ai and state of nature Sj.
• Let EMV (Ai) be the expected monetary value for the decision
alternative Ai.
EMV ( Ai ) = [ P( S j ) V ( Ai , S j ) ]

EMV (alternative i) = (payoff of first state of nature)

x (probability of first state of nature)
+ (payoff of second state of nature)
x (probability of second state of nature)
+ … + (payoff of last state of nature)
x (probability of last state of nature)
Decision Making Under Risk

• This is decision making when there are several possible

states of nature, and the probabilities associated with each
possible state are known.
• The most popular method is to choose the alternative with
the highest expected monetary value (EMV).

EMV ( Ai ) = [ P( S j ) V ( Ai , S j ) ]
Decision Making under Risk: Example
The Addis Ababa city administration is attempting to
determine the size of condominium houses to be built for
medium and low class income group of its residents. Three
sizes of developments considered by project planning team
are small, medium, and large houses. There are three future
conditions with regard to the level of demand for the houses
as a result of fluctuating household income. The demand for
the condominium houses can be low, medium, or high. The
project team has estimated probabilities for the occurrences
of high and medium demand as 0.5 and 0.3, respectively.
The project team has also prepared the payoff table,
expressed in revenue per 000' birr, as given below.
19-6

EXAMPLE
Alternatives, Ai States of nature, Sj

High (S1) Moderate (S2) Low (S3)

Large(A1) 50 70 100

Medium(A2) 40 80 90

Small(A3) 90 70 60

• What decision would you recommend using:

a) EMV decision criterion
b) EOL decision criterion
19-7

EXAMPLE continued

EMV decision criterion

• EMV of Large (A1) =(.5)(50)+(.3)(70)+(.2)(100)=66
• EMV Medium (A2) =(.5)(40)+(.3)(80)+(.2)(90)=62
• EMV Small (A3) =(.5)(90)+(.3)(70)+(.2)(60)=78 Maximum
EV
• Choose alternative A3 since it gives the largest expected
monetary value or expected payoff.
• Recommended decision is to build small size house with
highest revenue of 78000 Birr.
19-8
Expected Opportunity Loss
• Opportunity Loss or Regret is the loss due to the fact
that the exact state of nature is not known at the time a
decision is made.
• The opportunity loss is computed by taking the
difference between the optimal decision for each state
of nature and the other decision alternatives.
• What decision would you make based on the expected
opportunity loss criterion?
19-9

EXAMPLE 1 continued

• OPPORTUNITY LOSS TABLE

Alternative S1 S2 S3

A1 40 10 0

A2 50 0 10

A3 0 10 40
19-10

Expected Opportunity Loss

• Let Ai be the ith decision alternative.

• Let P(Sj) be the probability of the jth state of
nature.
• Let R(Ai,Sj) be the value of the regret for the
combination of decision alternative Ai and state
of nature Sj.
• Let EOL(Ai) be the expected opportunity loss for
the decision alternative Ai.
EOL( Ai ) = P( S j ) R( Ai , S j )
19-11

EXAMPLE 1 continued

Expected Opportunity losses

• EOL (A1) =(.5)(40)+(.3)(10)+(.2)(0)=23
• EOL (A2) =(.5)(50)+(.3)(0)+(.2)(10)=27
• EOL (A3) =(.5)(0)+(.3)(10)+(.2)(40)=11 Minimum
EOL
• Choose alternative A3/ build small size since it gives
the smallest/minimum expected opportunity loss
• Note: This decision is the same when using the highest
expected payoff (EMV). These two approaches will
always lead to the same decision.
 Expected Value of Perfect Information (EVPI)
•EVPI places an upper bound on what you should pay for additional
information.
EVPI = EVwPI – Maximum EMV
•EVwPI is the long run average return if we have perfect information
before a decision is made.
What is the worth of information known in advance before a strategy
is employed?
Optimal Decision Strategy
When High occurs, select Small
When Moderate occurs, select Medium
When Low occurs, select Large
EMV of optimal decision strategy (EPPI) is then multiplying the
occurrences by their probability values.
(.5)(90) + (.3)(80) + (.2)(100) = 89
Expected Value of Perfect Information (EVPI)

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)
Expected Value of Perfect Information (EVPI)

Suppose Scientific Marketing, Inc. offers analysis

that will provide certainty about demand
conditions (favorable). Additional information
will cost $15,000. Should the administration
purchase the information?
 EVPI cont’d) Decision Table with Perfect Information
Alternative High Moderate Low EMV
(S1) (S2) (S3)
Large(A1) 50 70 100 66
Medium(A2) 40 80 90 62
Small(A3) 90 70 60 78
With PI 90 80 100 89EVwPI
Probabilities .5 .3 .2
The maximum EMV without additional information is $78,000.
EVPI = EVwPI – Maximum EMV
= $89,000 - $78,000
= $11,000

So the maximum the administration should pay for the additional

information is $11,000. Therefore, it should not pay $15,000 for this
information.
19-14
Value of Perfect Information
• What is the worth of information known in advance before a
strategy is employed?
• Optimal Decision Strategy
· When High occurs, select Small
· When Moderate occurs, select Medium
· When Low occurs, select Large
EMV of optimal decision strategy (EPPI)
(.5)(90) + (.3)(80) + (.2)(100) = 89
• Expected Value of Perfect Information (EVPI) is the difference
between the expected payoff if the state of nature was known
and the optimal decision under the conditions of risk.
• EVPI= [(.5)(90) + (.3)(80) + (.2)(100)]-78 = 11.
 Decision Trees
• Any problem that can be presented in a decision table
can also be graphically represented in a decision tree.
• Decision Trees are useful for structuring the various
alternatives. They present a picture of the various
courses of action and the possible states of nature..
• All decision trees contain decision points or nodes, from
which one of several alternatives may be chosen.
• All decision trees contain state-of-nature points or nodes,
out of which one state of nature will occur.
 Steps of
Decision Tree Analysis
1. Define the problem.
2. Structure or draw the decision tree.
3. Assign probabilities to the states of nature.
4. Estimate payoffs for each possible combination of
alternatives and states of nature.
5. Solve the problem by computing expected
monetary values (EMVs) for each state of nature
node.
 Structure of Decision Trees
• Trees start from left to right.
• Trees represent decisions and outcomes in sequential
order.
· Squares represent decision nodes.
· Circles represent states of nature nodes.
· Lines or branches connect the decisions nodes and the
states of nature.
 High, S1 = .5
50
Moderate, S2 =.3
Large, A1 70 EMV= 66
2
Low S3 =.2
100
High, S1 = .5
40
Medium, A2 Moderate S3 =.3
1 3 80 EMV = 62
Low S3 =.2
90
High, S1 = .5
90
Small, A3 Moderate S3 =.3
4 70 EMV = 78
Low S3 =.2
60
Decision Making Under Uncertainty

There are several criteria for making decisions under

uncertainty:

1. Maximax (optimistic)
2. Maximin (pessimistic)
3. Criterion of realism (Hurwicz)
4. Equally likely (Laplace)
5. Minimax regret
19-12
Decision Strategies/ criteria
• Maximax Strategy - maximizes the maximum gain (optimistic
strategy)
• Maximin Strategy - maximizes the minimum gain (pessimistic
strategy)
• Minimax Regret Strategy - minimizes the maximum opportunity
loss
• Hurwicz Strategy – Highest weighted average value
• Laplace Strategy – Highest average value
• Under the maximax strategy, what is the optimal act and revenue
you are expecting?
• Under the maximin strategy, what is the optimal act and revenue
you are expecting?
• Under the minimax regret, Hurwicz and Laplace Criteria, what will
be your strategy?.
Maximax Decision Strategy: EXAMPLE continued
Used to find the alternative that maximizes the maximum
payoff.
◼ Locate the maximum payoff for each alternative.
◼ Select the alternative with the maximum number
Alternative High Moderate Low Maximum
(S1) (S2) (S3) in a row
Large(A1) 50 70 100 100
Medium(A2) 40 80 90 90
Small(A3) 90 70 60 90

The recommended maximax decision is to build a large house

yielding highest return of 100,000 Birr.
Maximin Strategy : EXAMPLE continued
Used to find the alternative that maximizes the minimum payoff.
◼ Locate the minimum payoff for each alternative.
◼ Select the alternative with the maximum number

AlternativeHigh Moderate Low Minimum in

(S1) (S2) (S3) a row
Large(A1) 50 70 100 50
Medium(A2) 40 80 90 40
Small(A3) 90 70 60 60

Under the maximin strategy, the recommended maximin decision is

to build a small house with a minimum return of 60,000 Birr.
Minimax Regret Strategy

This is based on opportunity loss or regret, which is the

difference between the optimal profit and actual payoff for a
decision.
· Create an opportunity loss table by determining the opportunity
loss from not choosing the best alternative.
· Opportunity loss is calculated by subtracting each payoff in the
column from the best payoff in the column.
· Find the maximum opportunity loss for each alternative and
pick the alternative with the minimum number.
19-13

EXAMPLE continued
• Under the minimax regret strategy, what will be your
strategy?
Expected Opportunity Loss Table
Alternative S1 S2 S3 Maximum
regret
A1 40 10 0 40

A2 50 0 10 50

A3 0 10 40 40

From the opportunity loss table, the strategy would be to build a

large or small house since these minimize the maximum regret.
 Criterion of Realism (Hurwicz, Middle Ground)

This is a weighted average compromise between

optimism and pessimism.
◼ Select a coefficient of realism , with 0≤α≤1.
◼ A value of 1 is perfectly optimistic, while a value of 0 is
perfectly pessimistic.
◼ Compute the weighted averages for each alternative.
◼ Select the alternative with the highest value.

Weighted average =(maximum in row) + (1 – )(minimum in row)

 Criterion of Realism (Hurwicz)
◼ Assume a criterion of realism,  = 0.8
◼ For the large house alternative using  = 0.8:
(0.8)(100,000) + (1 – 0.8)(50,000) = 90,000
◼ For the medium house alternative using  = 0.8:
(0.8)(90,000) + (1 – 0.8)(40,000) = 80,000
◼ For the small house alternative using  = 0.8:
(0.8)(90,000) + (1 – 0.8)(60,000) = 84,000
STATE OF NATURE
CRITERION OF
ALTERNATIVE S1 S2 S3 REALISM ( = 0.8) $
Construct a large
50,000 70,000 100,000 90,000
house
Realism
Construct a medium
40,000 80,000 90,000 80,000
house
Construct a small
90,000 70,000 60,000 84,000
house
Equally Likely (Laplace, criterion of rationality, principle of
insufficient reason)

Considers all the payoffs for each alternative

◼ Find the average payoff for each alternative.

◼ Select the alternative with the highest average.

STATE OF NATURE

ALTERNATIVE S1 S2 S3 Row Average $

Construct a large
50,000 70,000 100,000 73,333
house
Construct a medium Equall
40,000 80,000 90,000 70,000
house
likely
Construct a small
90,000 70,000 60,000 73,333
house
 Take home test
Record and Tape manufacturing company is considering several alternative methods of
expanding its production to accommodate increasing demand for its products. Based on this, the
company planners indicate that only three variable options are available to the company. These
are:
i. Expand the present plant
ii. Build a new plant
iii. Subcontract out extra production to other similar manufacturers
Having identified all the available alternatives, the decision makers identified and listed the
future events that may occur. These refer to demand for the products as high demand,
moderate demand, low demand and failure. Using the best information available, management
has also estimated the payoff in profits for its decisions as shown slide

Payoff table expressed in (000) profits over the next 5 years

Decision States of nature (Demand) Sj
alternatives (di) High, S1 Medium, S2 Low, S3 Failure, S4

Expand, d1 500 250 -250 -450

Build, d2 700 300 -400 -800
Subcontract, d3 300 150 -10 -100

Cont’d
Take home test
1. What would have been your decision using each of these decision criteria.?
a) Maximax decision criterion
b) Maximin decision criterion
c) Minimax regret criterion
d) Middle ground, Hurwicz decision criterion
e) Laplace decision criterion
2. Assume that management has given further details on the probabilities of the occurrences of the
states of nature as follows: P(high demand) = 0.2, P(moderate demand) = 0.3, and the P(low
demand) = 0.3 Given this information, what is the recommended decision using:
a) EMV criterion
b) EOL criterion
c) What is the optimal decision strategy with perfect information; calculate the EPPI, and EVPI?
d) Draw a decision tree for 2 (a)
3. If management wants to hire a consulting firm so as to obtain additional information the would
enable to better estimate the likelihoods for the occurrences of the states of nature, how much
should it be willing to pay for the additional information or consulting firm.