Decision

Theory
MANAGEMENT SCIENCE
 Learning Objectives
1.Explain the Decision Theory
2.Differentiate the types of decision
environments
3.Formulate a decision under a
particular decision environment; and
4.Use the decision tree to make a
decision
 Overview
The decision theory is an analytic approach to
making an economic decision by systematically
defining and analyzing a problem using a
mathematical model. The process involves listing
possible alternatives and possible outcomes. The
given values in a problem are presented in a
decision table or payoff table. Hence, decision-
making is involved if there are different available
alternatives to choose from.
 A Quantitative Approach to Decision
Making
01 | 02 | 03 |

Define the List the Identify the

Problem alternative possible outcomes
04 | s 05 | 06 |

List the Select and Make a

payoffs or apply a decision
conditional mathematical
model
EMV – expected monetary value.

P – Probability

M - Amount of money
Formula:
EMV = P(M)
Decision Making
Environment
Decision-makers heavily consider the
information existing in a particular
environment. The following are the types of
decision environments that surround a
decision-maker:
1.Decision Making under certainty
2.Decision Making under uncertainty
3.Decision Making under risk
 1
Decision Making
Under Certainty
Decision environment
Decision maker knows certain
outcome of his decision. He/she
selects the alternative that
gives a higher payoff or
maximizes the profit.
 2
Decision Making
Under
Uncertainty
Decision environment
Decision happens when
probability of the
occurrence of any outcome
is not available or is
unknown to the decision-
There are different criteria in
making decision under uncertainty
1. Maximax
2. Maximin
3. Equally likely
4. Criterion of Realism
5. Minimax ( opportunity loss)
 Illustration 1
As it expands its operation,
Princess Bank is confronted with
the problem of whether to establish
a full-blown branch or a branch-lite
unit in the municipality it desires to
expand in or not at all.
The payoff table of Princess Bank is
as follows:
Alternatives Favorable Unfavorable
Market Market
Condition Condition
gain (in loss (in
Pesos) Pesos)
Establish a full-blown
branch 300,000 -200,000
Establish a branch-lite
unit 150,000 -40,000
Not establish a full-
blown branch nor a 0 0
 Maximax
In the maximax criterion, the decision
maker adopts an optimistic approach.
He/she selects the alternative that has the
highest possible gain or maximizes the
outcome for every alternative.
Once the highest possible gain of all
alternatives has been determined, the
alternative with the maximum value is
chosen.
 Maximax
In the payoff table of Princess Bank,
the first alternative has a maximum
outcome of P300,000 gain, the second
has P150,000 gain and the third has
zero. The highest maximum outcome
is P300,000; hence, the decision under
maximax criterion is to establish a full-
blown branch.
 Decisio
n
Theory

Alternatives Highest Possible

Gain (Maximum
Value) Alternativ
Establish full-blown
₱ 300,000 e with the
branch
maximum
Establish a branch-lite
unit ₱ 150,000 value
Not establish a branch-
lite unit ₱0
 Maximax
Decisio
n
Theory

Alternatives Highest Possible

Gain (Maximum Alternativ
Value)
e with the
Establish full-blown
branch ₱ 300,000 maximum
Establish a branch-lite value
unit ₱ 150,000
Not establish a branch-
lite unit ₱0
DECISION: Princess bank
decision is to Establish a full-
 Decisio
n

Maximin Theory

In the maximin criterion, the decision-

maker maximizes the minimum
outcome for every alternative and
selects the alternative that has the least
possible loss. The decision-maker, in this
case, adopts the pessimistic
approach.
 Decisio
n
Maximin Theory

In the payoff table of Princess Bank,

alternative 1 has a minimum outcome of
₱200,000.00 loss; alternative 2 has
₱40,000.00 loss, and alternative 3 has zero.
The alternative that has the least possible
loss is alternative 3. The decision, then,
under the maximin criterion is to not
establish a full-blown branch nor a branch-
 Decisio

Maximin
n
Theory

Alternatives Minimum Outcome

(Possible Loss)
Establish full-blown
branch ₱ - 200,000
Establish a branch-
lite unit ₱ - 40,000 Alternative
Not establish a
₱0 with the
branch-lite unit
least
possible
DECISION: Princess bank decision is loss
not establish a branch-lite nor full
 Decisio
n

Equally Likely Theory

The equally likely criterion, otherwise known as

the Laplace criterion, selects the alternative that
has the highest average value. The decision-
maker adopts an approach in which the states of
nature have equal changes. In this criterion, the
average value of every alternative is computed
and the highest average value is selected.
 The average value of the Decisio
n
payoff table of Princess Theory
Bank is determined as
follows:
Alternatives Computation of Average
value (in
Average Value (in
Pesos)
Pesos)
Establish full- 300,000+(- 50,000 Alternative
blown branch
200,000)/2 with the
Establish a
highest
150,000+(- 55,000
branch-lite average
unit 40,000)/2
value
Not establish 0
a branch-lite (0 + 0)/2
unit
 Decisio
n
Theory
Equally Likely
The best average value is
₱55,000.00; hence, the
decision under the equally
likely criterion is to establish
a branch-lite unit.
 Decisio
n
Theory
Criterion of Realism
The criterion of realism, otherwise known as the
Hurwicz criterion, adopts the weighted average
approach wherein the coefficient of realism (α) is
used. It is an approach that is a compromise
between the maximax and maximin criteria. The
coefficient of realism has a value ranging between
0 and 1. A coefficient close to 1 indicates optimism
about the future, while a coefficient value close to
zero indicates pessimism.
Criterion of Realism
The formula to compute the amount in the
coefficient value close to zero indicates pessimism.
The formula to compute the amount in the criterion
of realism is as follows:

Criterion of realism =
α (maximum value in the alternative) +
(1 – α) (minimum value in the
alternative)
 Assume that Princess Bank sets the
coefficient of realism equal to 0.80, the
criterion of realism is determined as
follows:
Alternative Computation of Average value
s Average Value
Establish a full-
(0.80x300,000) ₱200,000 Alternative
blown branch
+ [0.20x(- with the
200,000)] highest
Establish a
(0.80x150,000) ₱112,000 criterion of
branch-lite unit
+ [0.20x(- realism
40,000)]
Not establish a 0
full-blown
(0*80%)
The decision under the criterion of realism is
+(20%*0)
branch nor ato establish a full-blown branch.
 Minimax
In the minimax criterion, the payoff table is converted into an
opportunity loss table. Then, the alternative with the
minimum opportunity loss is selected. The following steps
are observed in converting a payoff table to an opportunity
loss table:

1. Find the maximum payoff per column.

2. Determine the opportunity loss by column.
3. Select the alternative with the minimum opportunity loss.

Using the payoff table of Princess Bank, the following

steps are performed to select the alternative under
Step 1
Find the maximum payoff per
column. The maximum payoff in
column 1, with the values ₱300,000,
₱150,000, and ₱0 is ₱300,000. In
column 2, the values are ₱-200,000,
₱-40,000, and ₱0; hence, the
maximum payoff is ₱0.
Step 2
Determine the opportunity loss by column. The
opportunity loss by column is computed as
follows:
Alternatives Favorable Unfavorable Opportuni
market Market ty Loss
Establish a full- 300,000- 0-(-200,000) 200,000
blown branch
300,000=0 =200,000
Establish a 300,000- 0-(-40,000) 190,000
branch-lite unit
150,000=15 =40,000
0,000
Not to establish 300,000- 0-0=0 300,000
full blown branch
0=300,000
nor a branch-lite
unit
 Opportunity loss table appears as
follows.
Alternatives Favorable Unfavorable Opportuni
market Market ty Loss
Establish a full- 0 200,000 200,000
blown branch Minimum
of the
Establish a 150,000 40,000 190,000 maximu
branch-lite unit m
opportuni
Not to establish 300,000 0 300,000 ty loss
full blown branch
nor a branch-lite
unit
Step 3
Select the alternative with the
minimum opportunity loss. The
minimum value in the last column is
₱190,000, hence, the decision is to
construct or establish a branch lite
unit.
 3
Decision Making
Under Risk
Decision environment
Under risk, decision maker
use of probability. The two
methods primarily used when
making a decision under risk
1.expected
are: monetary value approach
2.expected opportunity loss
approach.
Probability is a statistical measure of
certainty that an event will happen. Its value
ranges from 0 to 1.
If the probability is equal to zero, the
event will never happen. However, if the
probability is equal to 1, there is 100%
probability that an event will occur.

The probability that an event will happen,

plus the probability that an event will not
happen, is always equal to one.
EXPECTED MONETARY
VALUE APPROACH
In the expected monetary value
(EMV) approach, the alternative
that has the highest expected
monetary value is selected. The
two variations of the EMV approach
are EMV without perfect
information and EMV with perfect
information.
EMV Approach without
Perfect Information
EXPECTED MONETARY VALUE
The expected monetary value of each
alternative or the expected monetary
value without perfect information is
computed as follows:

EMV (alternative i) = (first state of nature

payoff) × p1 + (second state of nature
payoff) × p2+...+ (last state of nature
payoff) × p
where:
P1 - probability of the first state of nature
P2 - probability of the second state of
nature
Pi - amount of
Illustration 2. money
EMV without Perfect Information

The payoff table of Princess Bank is as

follows:
 Alternatives Favorable Unfavorable
Market Market
Condition Condition
gain (in loss (in
Pesos) Pesos)
Establish a full-blown
branch 300,000 -200,000
Establish a branch-lite
unit 150,000 -40,000
Not establish a full-
blown branch nor a 0 0
branch-life unit
Assume the following probabilities:
Case 1: Favorable market condition, 60%;
unfavorable market condition, 40%

Case 2: Favorable market condition, 50%;

unfavorable market condition, 50%

**Required. Determine the decision to be

made for two cases.
Answer - Case 1. The EMV of each
alterative as follows
EMV Information EMVwoPI
Alternatives
EMV (full- = (300,000*60%) + 100,000
blown branch) (-200,000*40%)
EMV (branch- = (150,000*60%) + 74,000
lite unit) (-40,000*40%)

EMV (not at = (0*60%) + 0

all) (0*40%)
The highest EMV without perfect
information is P100,000, which comes
from the first alternative; hence,
Princess Bank should establish a full-
blown branch.
Answer - Case 2. The EMV of each
alterative as follows
EMV Information EMVwoPI
Alternatives
EMV (full- = (300,000*50%) + 50,000
blown branch) (-200,000*50%)
EMV (branch- = (150,000*50%) + 55,000
lite unit) (-40,000*50%)

EMV (not at = (0*50%) + 0

all) (0*50%)
The highest EMV without perfect
information is P55,000, which comes
from the second alternative.
Therefore, Princess Bank should
establish a branch-lite unit.
EMV Approach with Perfect
Information
EXPECTED MONETARY VALUE
Businesses sometimes seek the assistance of
consultancy firms in making decisions under risky
environments. In this case, there is a need to determine
the EMV with perfect information. The procedures to
determine the EMV with perfect information are as
follows:

1. Find the maximum payoff per state of nature

(column).
2. Multiply the determined maximum payoff per column
by the probability.
3. Add all the values in step 2 to get the EMV with
perfect information.
Hence, the formula to compute the EMV with
perfect information is as follows:

EMV with perfect information =(maximum

payoff of state of nature 1 × probability of
state of nature 1) + (maximum payoff of
state of nature 2 × probability of state of
nature 2) + . . . + (maximum payoff of state
of nature i × probability of state of nature i)
The formula to compute the
EMV of perfect information is as
follows:

EMV of perfect information =

EMV with perfect information -
EMV without perfect information
EMV Approach with Perfect Information
Use the payoff tables of Princess Bank in
Illustration 2. The probabilities are as follows:
1. Favorable market condition (state of nature 1)
- 60%
2. Unfavorable market condition (state of nature
2) - 40%
Required. Determine the following:
1. EMV with perfect information
2. EMV of perfect information
3. The amount of gain if Princess Bank buys the
perfect information for P50,000.00
Answer 1. The EMV with perfect information
is determined as follows:
Step 1
Find the maximum payoff per state of nature
(column). The maximum payoff per column
is determined as follows:
State of nature 1 (P300,000,150,000, 0)
=P300,000
State of nature 2 (-P200,000, -P40,000, 0) =
P0
 Step 2
Multiply the determined maximum payoff
per column by the probability. The value is
determined as follows:

State of nature 1 (300,000 × 60%) =

P180,000
State of nature 2 (0 × 40%) = P0
 Step 3

Add all the values in step 2 to get the

EMV with perfect information.

EMV with perfect information

P180,000.00 +0 = P180,000.00
Answer 2. The EMV without perfect
information has been determined in
Illustration 2, Case 1, with an amount of
P100,000.00. The EMV of perfect
information is computed as follows:
EMV of perfect information
= EMV with perfect information - EMV
without perfect information
= P180,000.00 -
P100,000.00
Answer 3. The cost of perfect information
is P50,000.00. The amount of gain is
computed as follows:

Amount of gain = P80,000.00 - P50,000.00

= P30,000.00
EXPECTED
OPPORTUNITY LOSS
APPROACH
Another approach to making a decision under
risk is the expected opportunity loss (EOL)
approach. The EMV approach maximizes the
expected monetary value, while the EOL
approach minimizes the expected opportunity
loss.
The following are the steps in the EOL
approach:
1. Construct an opportunity loss table.
2. Determine the EOL of every alternative.
3. Select the alternative with the lowest EOL
Answer. The following steps are performed
to determine which alternative should be
selected:
Step 1
Construct an opportunity loss table. To
construct an opportunity loss table, the
maximum payoff per column is
determined-column 1 is P300,000 and
column 2 is P0.
 Alternatives Favorable Unfavorable
market Market
Establish a full- 300,000- 0-(-200,000)
blown branch 300,000=0 =200,000
Establish a 300,000- 0-(-40,000)
branch-lite unit 150,000=150,0 =40,000
00
Not to 300,000- 0-0=0
establish full 0=300,000
blown branch
nor a branch-
lite unit
 Opportunity loss table appears as
follows.
Alternatives Favorable Unfavorable
market Market
Establish a full- 0 200,000
blown branch
Establish a 150,000 40,000
branch-lite unit
Not to establish 300,000 0
full blown branch
nor a branch-lite
unit
 Step 2
Determine the EOL of every alternative. Multiply the
probabilities of the state of nature by the opportunity
loss. The computation appears as follows:
Alternatives Favorable Unfavorable Opportuni
market Market ty Loss Minimum
of the
Establish a full- 0*60% 200,000*40 80000 maximu
blown branch
% m
opportuni
Establish a 150,000*60 40,000*40% 106000
branch-lite unit ty loss
%
Not to establish 300,000*60 0*40% 180000
full blown branch
%
nor a branch-lite
unit
 Step 3
Select the alternative with the lowest EOL value.
The lowest EOL value is P80,000.
The decision, therefore, is to establish a full-
blown branch.
It can be observed that the EMV approach
obtained the same result as the EOL approach,
which is to establish a full-blown branch. The
lowest amount of the EOL is the EMV of perfect
information.
DECISION TREE
The decision tree is the
graphical approach to
presenting a payoff table or
decision table. It contains the
decision points, state of
nature points, and
The decision point is represented by a rectangle, the state of
nature point is represented by a circle, and the alternatives are
represented by an arrow. The expected monetary value is placed
in a circle. In computing the EMV of an alternative or event, the
procedure is to work backwards, that is, from right to left of the
The EMV amounting to 100,000 is placed in circle 1.
The values of circles 1 and 2 are compared as the basis
of the decision. As the EMV of circle 1 is higher than
that of circle 2, the decision is to establish a full-blown
branch.
The complete decision tree appears as follows: