Lecture 4

Decision Analysis
Chapter 13 in the textbook: INTRODUCTION TO MANAGEMENT SCIENCE

DR. NOR KHAIRUSSHIMA

 Problem Formulation TYPE 1 :
Decision Making without Probabilities
Influence Diagram Optimistic Approach
Payoff Table Conservative Approach
Decision Tree Minimax Regret Approach

TYPE 2 : TYPE 3 :
Decision Making with Probabilities Decision Making with Perfect Information
Expected Value (EV) Approach EVPI
Risk Analysis
Sensitivity Analysis
TYPE 4 :
Decision Making with Sample Information
Decision Strategy
EVSI

DR. NOR KHAIRUSSHIMA 2

 INTRODUCTION
Definition of Decision-making according to
Trewatha & Newport,

"Decision-making involves the selection of a

course of action from among two or more
possible alternatives in order to arrive at a
solution for a given problem".

DR. NOR KHAIRUSSHIMA

 DECISION PROBLEM

A decision problem is characterized by decision

alternatives, states of nature, and resulting payoffs.

The decision alternatives are the different possible

strategies the decision maker can employ.

The states of nature refer to future events, not under

the control of the decision maker, which may occur.

DR. NOR KHAIRUSSHIMA 4

PROBLEM FORMULATION IN
DECISION MAKING
1 Clearly define the problem at hand

2 List the possible alternatives

3
Identify the possible outcomes or states of
nature
List the payoff or profit of each combination of
4
alternatives and outcomes

5
Select one of the mathematical decision theory
models
6 Apply the model and make your decision
5
DR. NOR KHAIRUSSHIMA
 CASE STUDY

 PDC purchased land that will be the site of a new luxury

condominium complex. PDC plans to price individual
condominium units between $300,000 and $1,400,000.

 PDC commissioned preliminary architectural drawings for 3

different projects, 30 condominium, 60 condominium and 90
condominium.

 The financial success of the project depends upon the size of

the condominium complex and the chance event concerning the
6
demand for the condominium.
DR. NOR KHAIRUSSHIMA
 Step 1 – Define the problem

 select the size of the new luxury condominium that will lead to the largest
profit given the uncertainty concerning the demand for the condominium

Step 2 – List decision alternatives

 d1=small complex with 30 condo

 d2 = medium complex with 60 condo
 d3 = large complex with 90 condo

DR. NOR KHAIRUSSHIMA

Step 3 – Identify possible outcomes/ states of nature

 s1 = strong demand for the condo

 s2 = weak demand for the condo

DR. NOR KHAIRUSSHIMA

 INFLUENCE DIAGRAMS
 An influence diagram is a graphical device showing the
relationships among the decisions, the chance events, and the
consequences.
 Squares or rectangles depict decision nodes.
 Circles or ovals depict chance nodes.

 Diamonds depict consequence nodes.

 Lines or arcs connecting the nodes show the
direction of influence.

DR. NOR KHAIRUSSHIMA

 CASE STUDY: INFLUENCE
DIAGRAM
States of nature:
Demand Strong (s1)
Weak (s2)

Complex Size Consequence:

Profit
Profit
Decision Alternatives
Small complex (d1)
10
Medium complex (d2)
Large complex (d3)
DR. NOR KHAIRUSSHIMA
 PAYOFF TABLES
 The consequence resulting from a specific combination of a
decision alternative and a state of nature is a payoff.

 A table showing payoffs for all combinations of decision

alternatives and states of nature is a payoff table.

 Payoffs can be expressed in terms of profit, cost, time,

distance or any other appropriate measure.

11

DR. NOR KHAIRUSSHIMA

 CASE STUDY
 For example, if a small complex (d1) is built and demand
turn out to be strong (s1), a profit of $8m will be realised. If
the demand turns out to be weak (s2) then the profit will be
$7m.

 If a medium sized complex (d2) is built and the demand turns

out to be strong (s1) then a profit of $14m is anticipated.
Instead if the demand is weak, a profit of $5m is achieved.

 If PDC choose to build a large complex (d3) and the demand

is strong (s1), the company will profited $20m. Instead if the
demand is weak (s2) the company is expected to lose $9m. 12

DR. NOR KHAIRUSSHIMA

 Step 4 – List the payoffs

Identify conditional values for the profits for large, medium

ad small condo for the two possible market demand
conditions

13

DR. NOR KHAIRUSSHIMA

 CASE STUDY: PAYOFF TABLE

State of Nature
Decision Alternatives
Strong Demand, s1 Weak Demand, s2

Small complex, d1

Medium complex, d2

Large complex, d3

14
DR. NOR KHAIRUSSHIMA
 Step 5 – Select the decision model

Depends on the situation and amount of risk and uncertainty

Step 6 – Apply the model to the data

Solution and analysis used to help the decision making

15

DR. NOR KHAIRUSSHIMA

 DECISION TREES
 Represent decisions and outcomes in sequential order

 Trees start from left to right

 Squares represent decision alternatives

 Circles represent states of nature
 Lines or branches connect the decisions nodes and the states
of nature nodes

 At the end of each limb of a tree are the payoffs attained 16

from the series of branches making up that limb.

DR. NOR KHAIRUSSHIMA
CASE STUDY: DECISION TREE
Strong (s1)
Small (d1) 8
2
7
Weak (s1)

Medium (d2) Strong (s1)

14
1 3
5
Weak (s1)

Strong (s1)
Large (d3) 20
4 17

Weak (s1)
-9
DR. NOR KHAIRUSSHIMA
TYPES OF DECISION MAKING
CONDITION
Type 1: Decision making without probabilities
 The decision maker does not know the probabilities
of the various outcomes

Type 2: Decision making with probabilities

 Decision maker knows with certainty the
consequences of every alternative or decision
choice

DR. NOR KHAIRUSSHIMA 18

 TYPES OF DECISION MAKING
CONDITION
Type 3: Decision making with perfect information
 The decision maker knows exactly the
probabilities of the various outcomes

Type 4: Decision making with sample information

 The decision maker knows the probabilities of the
various outcomes through additional information

DR. NOR KHAIRUSSHIMA 19

 TYPE 1: DECISION MAKING
WITHOUT PROBABILITIES
 Three commonly used criteria for decision making when
probability information regarding the likelihood of the states
of nature is unavailable are:

a) the OPTIMISTIC approach

b) the CONSERVATIVE approach
c) the MINIMAX REGRET approach.

20

DR. NOR KHAIRUSSHIMA

 a) OPTIMISTIC APPROACH
 The decision with the largest possible payoff is chosen.

 If the payoff table was in terms of costs, the decision

with the lowest cost would be chosen.

21

DR. NOR KHAIRUSSHIMA

 Case Study: Optimistic
Approach (Maximax Approach )
State of Nature
Decision Alternatives
Strong Demand, s1 Weak Demand, s2

Small complex, d1 8 7
Medium complex, d2 14 5
Large complex, d3 20 -9

Max payoff
22
DR. NOR KHAIRUSSHIMA
 b) CONSERVATIVE APPROACH
 For each decision the minimum payoff is listed and then the
decision corresponding to the maximum of these minimum
payoffs is selected. (Hence, the minimum possible payoff is
maximized.)

 If the payoff was in terms of costs, the maximum costs would

be determined for each decision and then the decision
corresponding to the minimum of these maximum costs is
selected. (Hence, the maximum possible cost is minimized.)

23

DR. NOR KHAIRUSSHIMA

 Case Study: Conservative
Approach
State of Nature
Decision
Alternatives
Strong Weak Minimum
Demand, s1 Demand, s2 payoff
Small complex, 8 7 7
d1
Ma
Medium 14 5 5 x
complex, d2 pa of m
yo
ffs in
Large complex, 20 -9 -9
d3
24
DR. NOR KHAIRUSSHIMA
c) MINIMAX REGRET APPROACH
 The minimax regret approach requires the construction of a
regret table or an opportunity loss table.

 This is done by calculating for each state of nature the

difference between each payoff and the largest payoff for that
state of nature.

 Then, using this regret table, the maximum regret for each
possible decision is listed.

 The decision chosen is the one corresponding to the

minimum of the maximum regrets.
DR. NOR KHAIRUSSHIMA 25
 Case Study: Minimax
Approach
Regret/ Opportunity Loss Table

State of Nature
Decision Alternatives
Strong Demand, s1 Weak Demand, s2

Small complex, d1 8 7
Max
Medium complex, d2 14 Max 5

Large complex, d3 20 -9
DR. NOR KHAIRUSSHIMA 26
 State of Nature
Decision Alternatives
Strong Demand, s1 Weak Demand, s2

Small complex, d1 20 - 8 = 12 7-7=0

Medium complex, d2 20 -14 = 6 7-5=2
Large complex, d3 20 - 20 = 0 7 - (-9) = 16

DR. NOR KHAIRUSSHIMA 27

 State of Nature
Decision Alternatives
Strong Weak Maximum
Demand, s1 Demand, s2 payoff
Small complex, d1 12 0 12

Medium complex, d2 6 2 6
Mini
Large complex, d3 0 16 16 ma
regre x
t

DR. NOR KHAIRUSSHIMA 28

For minimizing cost problems:

1. Calculating opportunity loss table: Subtract the lowest

payoff or cost in a column from each payoff in that
column.

2. Once the regret table has been constructed, the

minimax approach is applied exactly as the previous
way.

29

DR. NOR KHAIRUSSHIMA

 Type 2: Decision Making
with
Probabilities
Decision making when we know the probabilities associated
with each possible state of nature P(sj).

The decision yielding the best expected return is chosen as

the best decision alternative using:

 Expected Value (EV) Approach

30
DR. NOR KHAIRUSSHIMA
 Decision Making with
Probabilities

This method was used by The Worker’s Compensation Board

of British Columbia (WCB) to determine whether a short term
disability claim should be considered high risk or low risk
claim.

31
DR. NOR KHAIRUSSHIMA
The probabilities must satisfy 2 conditions:
1. P(sj) ≥ 0
2. ∑ P(sj) = 1

32

DR. NOR KHAIRUSSHIMA

 EXPECTED VALUE OF A
DECISION ALTERNATIVE
 The expected value (EV) of decision alternative di is
defined as:
N
EV( di )   P( s j )Vij
j 1

where: Vij = the payoff corresponding to decision

alternative di and state of nature sj

 The expected value of a decision alternative is the sum33

of weighted payoffs for the decision alternative.
DR. NOR KHAIRUSSHIMA
 CASE STUDY 1:
from previous example
 For example, if a small complex (d1) is built and demand
turn out to be strong (s1), a profit of $8m will be realised. If
the demand turns out to be weak (s2) then the profit will be
$7m.

 If a medium sized complex (d2) is built and the demand turns

out to be strong (s1) then a profit of $14m is anticipated.
Instead if the demand is weak, a profit of $5m is achieved.

 If PDC choose to build a large complex (d3) and the demand

is strong (s1), the company will profited $20m. Instead if the
demand is weak (s2) the company is expected to lose $9m. 34

DR. NOR KHAIRUSSHIMA

 PDC is optimistic about the potential of high rise condo
and initial probability assessments by the company
revealed that there is 0.8 that the demand will be strong
(s1) and corresponding probability that 0.2 the demand
will be weak (s2). P(s1)=0.8
P(s2)=0.2
State of Nature
Decision Alternatives Strong Demand Weak Demand
s1 s2
Small complex d1 8 7
Medium complex d2 14 5
Large complex d3 20 -9
35
DR. NOR KHAIRUSSHIMA
 EV(d1)= 0.8 (8) + 0.2 (7) = 7.8

EV(d2)= 0.8 (14) + 0.2 (5) = 12.2

EV(d3)= 0.8 (20) + 0.2 (-9) = 14.2

State of Nature
Strong DemandThe large condominium
Weak Demand
Decision Alternatives s1 complex with an expected
s2
0.8 value of $14.2 million
0.2
is the recommended decision
Small complex d1 8 7
Medium complex d2 14 5
Large complex d3 20 -9

DR. NOR KHAIRUSSHIMA 36

CASE STUDY 1: APPLYING
EV USING DECISION TREE
Strong (s1)
8
Small (d1) P(s1) =0.8
2
Weak (s2) 7
P(s2) =0.2

Strong (s1)
14
Medium (d2) P(s1) =0.8
1 3
Weak (s2) 5
P(s2) =0.2

Strong (s1)
20
Large (d3) P(s1) =0.8
4
Weak (s2) -9
P(s2) =0.2 37
DR. NOR KHAIRUSSHIMA
 CASE STUDY 1:
APPLYING EV USING
DECISION TREE
Small (d1) EV(d1)= 0.8(8) + 0.2 (7) =$7.8
2

Medium (d2)
1 3 EV(d2)= 0.8(14) + 0.2 (5) =$12.2

Large (d3) EV(d3)= 0.8(20) + 0.2 (-9) =$14.2

4 38

DR. NOR KHAIRUSSHIMA

 CASE STUDY 2: BURGER
PRINCE
• Burger Prince Restaurant is considering opening a new
restaurant on Main Street. It has three different models, each
with a different seating capacity.
• Burger Prince estimates that the average number of
customers per hour will be 80, 100, or 120. The probability value
for each state of nature is is given as P(s1)= 0.4 , P(s2)= 0.2,
P(s3)=0.4
• The payoff table is in the following slide.

39

DR. NOR KHAIRUSSHIMA

CASE STUDY 2: PAYOFF
TABLE

Average Number of Customer per Hour

Seating s1=80 s2= 100 s3=120
Capacity 0.4 0.2 0.4

MODEL A $10,000 $12,000 $14,000

MODEL B $ 8,000 $18,000 $12,000
MODEL C $16,000 $-9,000 $21,000

40

DR. NOR KHAIRUSSHIMA

 CASE STUDY 2

• Using the information given, construct a

decision tree and use the expected value approach to
determine optimal decision.

41

DR. NOR KHAIRUSSHIMA

EXPECTED VALUE FOR EACH
DECISION
EV(Model A)
2 = 0.4(10,000) + 0.2(12,000) +
0.4(14,000)
Model A
d1 = $12,000

Model B EV(Model B)
1 3
d2 = 0.4(8,000) + 0.2(18,000) + 0.4(12,000)
= $11,600
Model C
d3 4 EV(Model C)
= 0.4(16,000) + 0.2(-9,000) + 0.4(21,000)
= $13,000

Choose the model with largest EV, Model C 42

DR. NOR KHAIRUSSHIMA
TYPE 3: EXPECTED VALUE OF
PERFECT INFORMATION
(EVPI)
The expected value of perfect information (EVPI) is the
increase in the expected profit that would result if one
knew with certainty which state of nature would occur.

DR. NOR KHAIRUSSHIMA 43

 CASE STUDY 1: EVPI
PDC is optimistic about the potential of high rise condo
and initial probability assessments by the company
revealed that there is 0.8 that the demand will be strong
(s1) and corresponding probability that 0.2 the demand
will be weak (s2). P(s1)=0.8
P(s2)=0.2
State of Nature
Decision Alternatives Strong Demand Weak Demand
s1 s2
Small complex d1 8 7
Medium complex d2 14 5
Large complex d3 20 -9
44
DR. NOR KHAIRUSSHIMA
 If PDC knew for sure s1 will occur, then the best decision
is d3.

 Similarly if PDC knew for sure that s2 will occur, then

the best decision is d1

 Therefore, PDC optimal decision strategy when perfect

info is available are:

If s1, select d3 and received payoff of $20 million

If s2, select d1 and received payoff of $7 million

DR. NOR KHAIRUSSHIMA 45

 CASE STUDY 1: EVPI

 So what is the expected value for this decision

strategy?

 There is 0.8 probability that the prefect information will

indicate state of nature s1 and the resulting decision d3
will provide $20 million profit

 Similarly, with a 0.2 probability for s2, the optimal

decision alternative d1 will provide $7 million profit

DR. NOR KHAIRUSSHIMA 46

 CASE STUDY 1: EVPI
 The expected value of the decision strategy that uses
perfect information is:
N
EV( d i )   P( s j )Vij
j 1

0.8(20)+0.2(7)=$17.4 million

tio h
PI a it
n
w rm w
EV nfo lue
t i va
In the previous example, because

ec d
rf te
pe pec
decision is obtained without the
Ex
benefit of perfect information $14.2
million is referred to as EVwoPI

DR. NOR KHAIRUSSHIMA 47

 CASE STUDY 1: EVPI

The expected value of perfect information

EVPI=$17.4 - $14.2= $3.2 million

$3.2 million represent the additional expected value that

can be obtained if perfect information is available about the
states of nature

PCD might have to seriously consider a

market survey as a way to obtain more
accurate information about the states of
nature

DR. NOR KHAIRUSSHIMA 48

 EXPECTED VALUE OF
PERFECT INFORMATION
(EVPI)

Expected Value of Perfect Information

EVPI =| EVwPI – EVwoPI |

DR. NOR KHAIRUSSHIMA 49

 END
OF CLASS

DR. NOR KHAIRUSSHIMA 50