DECISION

ANALYSIS

MANSCI
 LESSON OBJECTIVES

01 Explain the concept of Decision Analysis

and the Decision-making Environments

02 Identify and construct a basic

decision model.

03 Compare and evaluate various

decision-making approaches.
 Ronald A. Howard, a professor
Prof. Ronald A. Howard of Management Science and
Engineering at Stanford
Stanford University University, is credited with
originating the term in 1964.
 Decision analysis (DA) is a

SISYLANA
NOISICED
systematic, quantitative, and
visual approach to addressing and
evaluating the important choices
that businesses sometimes face.

The idea is used by large and small

NOISICED

corporations alike when making various

SISYLANA

types of decisions, including

management, operations, marketing,
capital investments, or strategic
choices.
 The goal of decision
analysis is to ensure that
decisions are made with all
the relevant information
and options available.

For example, a corporation

may use it to make million-
dollar investment decisions,
INFORMATION or an individual can use it to

IS KING! decide on their retirement

savings.
THE
ASPECTS

Decision analysis uses a variety of

tools to evaluate all relevant
information to aid in the decision- Decision analysis allows
making process and incorporates corporations to evaluate and
aspects of psychology, management model the potential outcomes of
techniques, training, and economics. various decisions to determine the
correct course of action.
 Decision Models
Decision Models come in the form
of a Decision Tree
or an Influence
Diagram.
The Decision Tree
The Decision Tree

One of the most common tools for decision

analysis, a decision tree is a flowchart that shows

your options' potential outcomes. To create a

decision tree, start by drawing a square, which

represents your question or problem, then you

create branches that represent your choices and

their outcomes.
The Influence Diagram
The Influence Diagram

An influence diagram displays a summary of the

information contained in a decision tree.

It involves four variable types for notation: a decision (a

rectangle), chance (an oval), objective (a hexagon), and

function (a rounded rectangle). Influence diagrams also

use solid lines to denote influence.

Their appearance is very similar to a flowchart.

 What's the Difference Between

Decision Trees and Influence Diagrams?

Influence diagrams show the

dependencies among variables. This is an

important distinction between influence

diagrams and decision trees.

Decision trees offer much more detail

about each possible decision.

The Decision Table

A decision table is a tabular method for representing

and analyzing decision logic.

STNEMNORIVNE GNIKAM-NOISICED

C U R .

Such type of In the decision making Under the condition of

environment is very environment of uncertainty, risk, there is the
sure and certain by its the information available to possibility of more than
ytniatrecnU
nature. This means that the manager is incomplete, one event taking place.
insufficient and often Which means the
ytniatreC

all the information is unreliable. Everything is in a manager has to first

available and at hand. state of flux. Several ascertain the possibility
Such data is also easy external and random forces and probability of the
to attain and not very mean that the environment occurrence or non-

ksiR
expensive to gather. is most unpredictable. occurrence of the event
STNEMNORIVNE GNIKAM-NOISICED

C U R .

Such type of In the decision making Under the condition of

environment is very environment of uncertainty, risk, there is the
sure and certain by its the information available to possibility of more than
ytniatrecnU
nature. This means that the manager is incomplete, one event taking place.
insufficient and often Which means the
ytniatreC

all the information is unreliable. Everything is in a manager has to first

available and at hand. state of flux. Several ascertain the possibility
Such data is also easy external and random forces and probability of the
to attain and not very mean that the environment occurrence or non-

ksiR
expensive to gather. is most unpredictable. occurrence of the event
Sample Problem (Certainty)

A bakery wants to decide whether to bake 100

or 200 cupcakes for a school event.
The school guarantees to buy exactly 100
cupcakes at ₱40 each.
The cost to make each cupcake: is ₱25.
So, the profit per cupcake = ₱40 – ₱25 = ₱15
And the school will buy only 100, not more.
COMPUTATION

Decision:
If the bakery bakes 100 cupcakes → all will be sold →
Profit = 100 pcs × ₱15 profit = ₱1,500

If the bakery bakes 200 cupcakes → still, only 100 will

be sold →
Profit = (100 × ₱15) – (100 × ₱25)
Profit = ₱1,500 – ₱2,500 = –₱1,000 loss
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA
Outcomes are
known
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are

known
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
STNEMNORIVNE GNIKAM-NOISICED

C U R .

Such type of In the decision making Under the condition of

environment is very environment of uncertainty, risk, there is the
sure and certain by its the information available to possibility of more than
ytniatrecnU
nature. This means that the manager is incomplete, one event taking place.
insufficient and often Which means the
ytniatreC

all the information is unreliable. Everything is in a manager has to first

available and at hand. state of flux. Several ascertain the possibility
Such data is also easy external and random forces and probability of the
to attain and not very mean that the environment occurrence or non-

ksiR
expensive to gather. is most unpredictable. occurrence of the event
Sample Problem (RISK)

A software company recently designed and developed a new

service for its customers. It needs to decide whether to launch
the service next month or wait for six months. The company
performs research and discovers the probability of success for
both options, along with their potential revenue. It also learns the
probability of failure and corresponding losses for each.
Option A: Launching the service next month has a 55% probability
of success with potential revenue of $250,000. It has a 45%
probability of failure with a potential loss of $125,000.
Option B: Launching the service in six months has a 65%
probability of success with potential revenue of $400,000. It has
a 35% probability of failure with a potential loss of $200,000.
 Formula time
To calculate the expected value of each option,
use the following formula:
EV =
(probability of success x potential revenue)
+
(probability of its failure x potential loss)
ALTERNATIVES SUCCESS Probability FAILURE Probability

Option A $250,000 55% $125,000 45%

(Next month)
Option B $400,000 65% $200,000 35%
(Six months)
Option A
EV =
(probability of success x potential revenue)
+
(probability of its failure x potential loss)

EV = (55% x 250,000) + (45% x -$125,000)

EV = $81,250
Option B
EV =
(probability of success x potential revenue)
+
(probability of its failure x potential loss)

EV = (65% x 400,000) + (35% x -$200,000)

EV = $190,000
Decision

Based on these calculations, the software

company decides to wait six months to
launch its new service because its
expected value is about double the value
they could earn if they launch it next
month.
Sample Problem (RISK)

DECISION TREE
Sample Problem (RISK)

DECISION TREE
Sample Problem (RISK)

DECISION TREE

DECISION NODE
Sample Problem (RISK)

DECISION TREE

DECISION NODE

CHANCE NODE
OUTCOME NODE
EVENT NODE
DECISION TREE
DECISION TREE
DECISION TREE

nth
mo
ext
N
DECISION TREE

nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

Success (0.55)
Failure (0.45)
nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

Payoff
Success (0.55)
Failure (0.45)
nth
mo
ex t
N
Six
mo
nth
s
DECISION TREE

Payoff
Success (0.55) $250,000
Failure (0.45)
o nth -$125,000
x t m
Ne
Six
mo
nth
s
DECISION TREE

Payoff
Success (0.55) $250,000
Failure (0.45)
o nth -$125,000
x t m
Ne
Six
mo
nth
s Success (0.65) $400,000
Failure (0.35)
-$200,000
DECISION TREE

Payoff
Success (0.55) $250,000 (0.55)(250k)
+
Failure (0.45) (0.45)(-125k)
o nth -$125,000
x t m
Ne
Six
mo
nth
s Success (0.65) $400,000
Failure (0.35)
-$200,000
DECISION TREE

Payoff
Success (0.55) $250,000 (0.55)(250k)
+
Failure (0.45) (0.45)(-125k)
o nth -$125,000
x t m
Ne
Six
mo
nth
s Success (0.65) $400,000 (0.65)(400k)
+
Failure (0.35) (0.35)(-200k)
-$200,000
DECISION TREE

Payoff
81,250 Success (0.55) $250,000 (0.55)(250k)
+
Failure (0.45) (0.45)(-125k)
o nth -$125,000
x t m
Ne
Six
mo
nth
s Success (0.65) $400,000 (0.65)(400k)
+
Failure (0.35) (0.35)(-200k)
-$200,000
DECISION TREE

Payoff
81,250 Success (0.55) $250,000 (0.55)(250k)
+
Failure (0.45) (0.45)(-125k)
o nth -$125,000
x t m
Ne
Six
mo
nth 190,000
s Success (0.65) $400,000 (0.65)(400k)
+
Failure (0.35) (0.35)(-200k)
-$200,000
DECISION TREE

Payoff
81,250 Success (0.55) $250,000 (0.55)(250k)
+
Failure (0.45) (0.45)(-125k)
o nth -$125,000
x t m
Ne
190,000 Six
mo
nth 190,000
s Success (0.65) $400,000 (0.65)(400k)
+
Failure (0.35) (0.35)(-200k)
-$200,000
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes
uncertain, but
probabilities known
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes
RISK uncertain, but
probabilities known
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes Expected Value, Expected
RISK uncertain, but Utility / Decision Tree
(Folding Back)
probabilities known
Sample Problem (UNCERTAINTY)

Decision-making Approaches
Sample Problem (UNCERTAINTY)

Decision-making Approaches
Maximax (Optimistic)
Maximin (Conservative or Pessimistic)
Minimax Regret
Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Decision Alternatives
Sample Problem (UNCERTAINTY)

Outcomes
Sample Problem (UNCERTAINTY)

Payoffs
Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Best
Sample Problem (UNCERTAINTY)

Best
Sample Problem (UNCERTAINTY)

Best
45
Sample Problem (UNCERTAINTY)

Best
45
70
Sample Problem (UNCERTAINTY)

Best
45
70
53
Sample Problem (UNCERTAINTY)

Best
45
70
53
Next, choose the Best of Best (Highest/Optimistic)
Sample Problem (UNCERTAINTY)

Best
45
70
53
Sample Problem (UNCERTAINTY)

Best
45
70
53
Decision: Invest in Stocks
Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Worst
Sample Problem (UNCERTAINTY)

Worst
Sample Problem (UNCERTAINTY)

Worst
5
Sample Problem (UNCERTAINTY)

Worst
5
-13
Sample Problem (UNCERTAINTY)

Worst
5
-13
-5
Sample Problem (UNCERTAINTY)

Worst
5
-13
-5
Sample Problem (UNCERTAINTY)

Worst
5
-13
-5
Decision: Invest in Bonds
Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Regret = Opportunity Loss

Sample Problem (UNCERTAINTY)

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Look for the highest value

Sample Problem (UNCERTAINTY)

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

70 - 40 = 30

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

70 - 40 = 30

70 - 70 = 0

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

70 - 40 = 30

70 - 70 = 0

70 - 53 = 17

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

45 - 45 = 0

45 - 30 = 15

45 - 45 = 0

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

5 - 5 = 0

5 - (-13) = 18

5 - (-5) = 10

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)

70 - 40 = 30 45 - 45 = 0 5 - 5 = 0

70 - 70 = 0 45 - 30 = 15 5 - (-13) = 18

70 - 53 = 17 45 - 45 = 0 5 - (-5) = 10

Regret = Best Payoff - Payoff Received

Sample Problem (UNCERTAINTY)
Sample Problem (UNCERTAINTY)

Maximum
Sample Problem (UNCERTAINTY)

Maximum
Sample Problem (UNCERTAINTY)

Maximum
30
Sample Problem (UNCERTAINTY)

Maximum
30
18
Sample Problem (UNCERTAINTY)

Maximum
30
18
17
Sample Problem (UNCERTAINTY)

Maximum
30
18
17
“Minimize the maximum regret” / Choose lowest
Sample Problem (UNCERTAINTY)

Maximum
30
18
17
Decision: Invest in Mutual Funds
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes Expected Value, Expected
RISK uncertain, but Utility / Decision Tree
(Folding Back)
probabilities known
Outcomes unknown,
no probabilities
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes Expected Value, Expected
RISK uncertain, but Utility / Decision Tree
(Folding Back)
probabilities known
Outcomes unknown,
UNCERTAINTY no probabilities
 GUIDE
ENVIRONMENT CHARACTERTISTICS APPLICABLE
DECISION CRITERIA

CERTAINTY Outcomes are Choose best

known outcome directly
Outcomes Expected Value, Expected
RISK uncertain, but Utility / Decision Tree
(Folding Back)
probabilities known
Outcomes unknown, Maximin, Minimin,
UNCERTAINTY no probabilities Minimax Regret
GUIDE
 GUIDE
When the decision-maker is risk-averse or conservative.
 GUIDE
When the decision-maker is risk-averse or conservative.
“I don’t know what will happen, so I’ll prepare for the worst.”
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or

optimistic.
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or

optimistic.
“Even if I don’t know the future, I want the best
possible outcome.”
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or Maximax

optimistic. (Optimistic
“Even if I don’t know the future, I want the best Approach)
possible outcome.”
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or Maximax

optimistic. (Optimistic
“Even if I don’t know the future, I want the best Approach)
possible outcome.”

When the decision-maker wants to avoid

future regret.
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or Maximax

optimistic. (Optimistic
“Even if I don’t know the future, I want the best Approach)
possible outcome.”

When the decision-maker wants to avoid

future regret.
“I don’t want to look back and regret not
choosing something better.”
 GUIDE
Maximin
When the decision-maker is risk-averse or conservative. (Pessimistic
“I don’t know what will happen, so I’ll prepare for the worst.” Approach)

When the decision-maker is risk-seeking or Maximax

optimistic. (Optimistic
“Even if I don’t know the future, I want the best Approach)
possible outcome.”

When the decision-maker wants to avoid

future regret. Minimax Regret
(Regret-Averse
“I don’t want to look back and regret not Approach)
choosing something better.”
THANK
YOU!
Seatwork (Risk)

A consulting business wants to build a new office and has narrowed its
list to three choices: Austin, Boston and Chicago. After completing
research on each option, it can define the probability of success for each
location and the potential revenue each can earn. The business has also
determined the options' probabilities of failure and the corresponding
potential losses.
Austin: Opening the Austin office represents a 35% probability of
success with expected potential revenue of $12 million. It has a 65%
probability of failure with an expected potential loss of $2 million.
Boston: Opening the Boston office represents a 40% probability of
success with expected potential revenue of $16 million. It has a 60%
probability of failure with an expected potential loss of $8 million.
Chicago: Opening the Chicago office represents a 45% probability of
success with expected potential revenue of $20 million. It has a 55%
probability of failure with an expected potential loss of $10 million.