Chapter 5:

The Decision Theory

Prepare by:
Diana C. Magpale
Mark Joseph C. Factor
 The Decision Theory

• Introduction
• Mathematical Expectation (ME) or Expected Value (EV)
• How to compute for the ME and EV
• Value of Information
• Decision under Certainty
• Decision under Uncertainty
• Decision under Risk
• Expected Value of Perfect Information (EVPI)
• Decision Tree Analysis
 5.5

Decision under Certainty

 Decision under certainty

• When it is known of certain which of the possible future conditions will actually
happen. Choose the alternative with the highest payoff under the state of the
nature
 Decision under certainty

Example
Stereo Industries Ltd. , a company that plans to expand its product line,
must decide to build a small, medium or large facility plant to produce the products.
The payoff table below illustrates the capacity planning of the company.

Alternatives Possible Future Demand

Low Inadequate High

Small facility P350 P350 P350

Medium facility 245 420 420

Large Facility -140 70 560

*Present Value in thousand pesos

 Decision under certainty

What is the best alternative in the pay off table, if it is known with certainty that demand will be:

a. Low
b. Moderate
c. High

Solution:
Choose the alternative with the highest payoff. Thus, if we know that demand will be
low, we would choose to build a small facility with a payoff of 350, 000 pesos. If the demand will
be moderate a medium facility with the highest payoff of 429, 000 pesos. For high demand, a
large facility with the highest payoff.
 5.6

Decision under Uncertainty

 Decision under uncertainty

• There is no available information on how likely the various states of nature are.

Four possible decision criterions

a. maximin - takes into account only the worst possible outcome for each alternative.
“guarantee minimum”
b. maximax – takes into account only the best possible outcome for each alternative
c. laplace – takes into account only the best possible outcome for each alternative
d. minimax regret – determine the worst regret for each alternative, and choose the alternative
with the “best worst”
 Decision under uncertainty

Example
Referring to the preceding payoff table, determine which alternative would be
chosen under each of these strategies:
a. Maximin b. Maximax c. Laplace d. Minimax regret

Solutions :
a. Maximin
The worst payoff for the alternatives are:
Small Facility : P350T
Medium Facility : P245T
Large Facility : P140T

Since, P350T is the best; it leads to building a small facility.

 Decision under uncertainty

Solutions :
b. Maximax
the best overall payoff is the P560T, which lead to building a large
facility under maximax criterion

c. Laplace
find the row totals and divide each of the amounts by the number
of states of nature.
Row Total Row Average

Small Facility P1050T P350T

Medium Facility 1085 361.67

Large Facility 490 163.33

 Decision under uncertainty

c. Laplace
find the row totals and divide each of the amounts by the number
of states of nature.

Row Total Row Average

Small Facility P1050T P350T

Medium Facility 1085 361.67

Large Facility 490 163.33

*Since, P361.67T is the highest average, then the medium facility would be
chosen under the laplace criterion.
 Decision under uncertainty

c. Minimax regret
• Prepare a table of opportunity losses, or regrets
• Subtract every payoff in each column from the largest positive payoff in
the column. For instance, in the first column, the largest positive
payoff is P350T, so each of the three numbers in that column must be
subtracted from 350 and so on.
• Determine the worst regret for each alternative which is the highest
payoff per row, and then choose the best of these “worst” would be
chosen under minimax regret.
• The result were placed in the regret table.
 Decision under uncertainty

The Regret Table

Alternative Regrets

Low Moderate High Worst

Small Facility 0 70 210 210

Medium Facility 105 0 140 140

Large Facility 490 350 0 490

*The lowest regret is 140, which is for a medium facility. Hence, that alternative
would be chosen

13
 Decision under uncertainty

The Regret Table

Alternative Regrets

Low Moderate High Worst

Small Facility 0 70 210 210

Medium Facility 105 0 140 140

Large Facility 490 350 0 490

*The lowest regret is 140, which is for a medium facility. Hence, that alternative
would be chosen

14
 5.7

Decision under Risk

 Decision under Risk

• It is assumed that the probability distributions are known

Expected Monetary Value Criterion (EMV)

EMV is computed for each alternative, and the one with the highest expected
value is selected.

*Expected value is the sum of the payoffs for an alternative where each payoff
is weighted by the probability for the appropriate state of nature.
 Decision under Risk

Example :
Identify the best alternative for the previous payoff table using
the EMV with the following probabilities.

Low = .35
Moderate = .40
High = .25
 Decision under Risk

Solution:
1) Find the expected value of each alternative by multiplying the
probability of occurrence for each nature by the payoff
2) Sum their products
EV (small) = .35(350) + .40(350) + .25(350)
EV (medium) = .35(245) + .40(430) + .25(420)
EV (high) = .35(-140) + .40(70) + .25(560)

EV (small) = 122.5 + 140 + 87.5 = 350

EV (medium) = 85.75 + 168 + 105 = 358.5
EV (high) = -49 + 28 + 240 = 119
Hence, choose the small-sized facility because it has the highest expected
value
 5.8

Expected Value of Perfect Information

 Expected Value of Perfect Information

Perfect information depends somewhat on the nature of the decision to be

made.
For example, information about consumer preferences might come from market
research; product testing; legal experts might be called on; and so on.

Two ways to determine the EVPI:

1) By finding the difference between the expected payoff under certainty and
the expected payoff under conditions of risk.

2) Using regret table to compute for the EVPI

 Expected Value of Perfect Information

Example:
Based on the information in the preceding examples, determine the
expected value of perfect information using the first and second method.

First Method.
Compute for the expected payoff under certainty.
.35(350) + .40(420) + .25(560) = P430.5

The expected payoff under risk is P358.75, as computed earlier.

The EVPI is P430.5 – 358.75 = P71.75
 Expected Value of Perfect Information

Second Method.

Find the expected regret value for each alternative. The smallest expected
value is equal to the EVPI.

Alternatives Low Moderate High EV

Small Facility .35(0) .40(70) .25(210) = 80.5

Medium Facility .35(105) .40(0) .25(140) = 71.75

Large Facility .35(490) .40(350) .25(0) = 351.5

The lowest expected regret is 71.75, which is the same in the First method
solution. This figure indicates the upper limit on the amount the decision maker
should be willing to spend to obtain perfect information.
 5.9

Decision Tree Analysis

 Decision Tree Analysis

- A schematic representation of the alternatives available to a decision

maker and their possible consequences.
- Two types of nodes; square = decision point, circle = chance
event.
- It is read left to right and analysed from right to left (starting with the
last decision that might be made)
- For each decision, choose the alternative that will yield the greatest
return
- If chance events follow a decision, choose the alternative that has
the highest expected value (EV) or the lowest expected loss.
 Decision Tree Analysis

Example.
Mr. Ven Tibar, the ,manager of SDP corporation is
faced with deciding whether to prepare a bid or not. It costs
P5,000 to prepare the bid. If the bid is submitted, the
probability that the contract will be awarded is 65%. If the
contract will be awarded to the corporation, it may gain an
income of P80,000 if it succeeds, or pay a fine of P15,000 if
it fails. The probability of success is estimated to 60%.

Should the manager prepare a bid?

 Decision Tree Analysis

The problem is whether to prepare a bid or not.

A
Prepare

Not to Prepare
 Decision Tree Analysis

The problem is whether to prepare a bid or not.

 Decision Tree Analysis

The success branch, P80,000 is the return, but since there

is a cost of preparing the bid P5,000 this should be
subtracted.

A fine of P15,000 is imposed in case of failure and also a

cost of P5,000 for the bid preparation.
This indicate at the end of the failure branch (P15,000 –
P5,000). I
In position B, even if the contract is not awarded, there is a
cost of P5,000 for bid preparation.
 Decision Tree Analysis

To compute for position C, start from the tip of the tree.

EV = .6(75,000) + .4(-20,000)
= P45,000 – P8,000
= P37,000
For position B:
EV = .65(37,000) + P1,750
= P24,050 – P1,750
= P22,300
 BIG CONCEPT
Bring the attention of your audience over a key concept
using icons or illustrations

30
31