KARATINA UNIVERSITY

BBM 355 OR1-DECISION THEORY

Definition
Decision making is concerned with choosing a course of actions from among available
alternatives.
STEPS
1. Define the problem at hand.
2. List the possible alternative courses of action.
3. Identify the possible outcomes (states of nature/and their probability if possible).
4. Estimate the conditional pay offs for each combination of a decision alternative/state of
nature.
5. Select a suitable or appropriate mathematical tool from among the techniques.
6. Apply the model and hence recommend the best course of action.
7. Perform sensitivity or (what if? Or post optimality) analysis, set up controls and implement.

STRUCTURE/ANATOMY OF A DECISION PROBLEM

Decision
alternatives
States of nature/possible outcomes
S1 S2 Sn
a1 P11 P1 ……………... P1n

a2 P21 P22 …………….. P2n

am Pm1 Pm2 ……………. Pmm

EXPLANATIONS/DEFINITIONS
1. Alternative ai = this is a course of action or strategy that may be chosen by the decision
marker (It is controllable). We assume decision alternatives are mutually exclusive. (If one is
selected, all the others are left).
2. States of nature, sj = this represent an outcome or occurrence over which the decision-maker
has little or no control. The chances of occurrence of various states of nature may or may not
be available. (Assumption: No control).
3. Conditional pay offs (Pijs) = This is a consequence or outcome normally expressed in
monetary value that occurs as a result of a particular alternative and a state of nature.
Illustration (A farmer)

SON
DA
Little Rain Moderate Rain Much Rain
Maize -1 2 5
Beans 5 7 -3
Wheat 1 3 -4
None 0 0 0

1
DECISION MAKING ENVIRONMENTS
The study of decision making is greatly facilitated by classifying decision making scenarios or
environments depending on the level of knowledge about the possible outcomes or states of
nature and the number of participants in the decision making environment. To this extent four
decision making environments have been defined.
These are:-
 Environment of certainty
 Environment of uncertainty
 Environment of risk
 Environment of competition / Game theory

1 Environment of Certainty
This is where all the information about a state of nature is known for sure. The model used to
recommend the best course(s) of action in this environment are certainty models (deterministic
model.)
2 Environment of uncertainty
In this enviroment the probabilities for the various states of nature are not known or may be
unreliable. Uncertain events are those events that cannot be predicted with statistical confidence
i.e the decision maker does not know the relevant variables nor their probability therefore
decision making in this environment depends on the risk attitude of the decision maker i.e. risk
averse or a risk seeker or risk neutral.
A risk averse avoids risk at all cost; the individual is conservative and assumes the worst
situation will occur.
A risk seeker takes high risks in expectation of high returns. He assumes that the best outcome
will occur.
A risk neutral person is not affected by risks. Such people will make decisions based on
something else but not risk.
Some criteria used under this environment are
a) Maximax
b) Maximin
c) Minimax regret
d) Hurwicz
e) Laplace

MAXIMAX
Selects the best of the best for all possible decision alternatives. It is a criterion of extreme
optimism.
USERS
- Risk takers
- Large firms which are able to absorb huge losses in case the state of nature turns
out to be unfavourable.
Example
Alternative high medium low No Action Max. of row
Expand 50,000 25,000 -25,000 -45,000 50,000
Contract 70,000 30,000 -40,000 -80,000 70,000
Subcontract 30,000 15,000 -10,000 -10,000 30,000

2
Maxmax payoff is sh70,000 corresponding to contract alternative.

MAXIMIN
This method selects the maximum out of all minima for all decision alternatives. It is a criterion
of extreme pessimism.
USERS
- Risk averters
- Small firms which cannot absorb large losses in case things turn out to be unfavourable.
Example
Min in row (sh 000)
-45
-80
-10
Max of min is -10 hence subcontract

HURWICZ (criterion or realism)

This is a compromise between maximax and Maximin.
Argument: It is unrealistic to view decision-makers as extreme optimists or pessimists.
Solution: Obtain the degree of optimism of the decision-maker (coefficient of optimism) 
(alpha) 0 <  < 1.
When  = 0, we have an extreme pessimist (Maximin)
 = 1, we have an extreme optimist (maximax)
Thus: For each alternative we obtain the following:
Expected pay off = Best payoff) + (worst payoff) (1 -).
Example
Assuming an  of 0.8 then
EMVs (sh 000) are
Expand=50,000*0.8- 45,000*0.2=31
Contract=40
Sub contract=22
Hence contract

Laplace (Equally likely)

Also called the criterion of rationality. This method takes all states of nature as equally likely.
Thus for any states of nature, the probability of each state = 1/n
The effect is to transform the decision environment from that of under uncertainty to that of
under the environment of risk.
Example
Expand1/4(50,000+25,000 -25,000 -45,000)= 1250
= -500
= 8500
Hence subcontract

MINIMAX REGRET (Savage criterion)

This concept is based on opportunity loss. The idea is to minimize the maximum possible
opportunity loss i.e. when a decision has been taken, the decision maker may experience some

3
loss; the question is what the maximum possible loss that will be experienced is? The method
then seeks to minimize this maximum possible loss for each decision alternative.

Alternative High Moderate low No action Max

Expand 20,000 5,000 24,000 35,000 35,000
Construct 0 0 39,000 70,000 70,000
Subcontract 40,000 15,000 0 0 40,000
The company should expand because this decision will minimise its regret which is Sh. 35,000.

3 Environment of risk.
In this environment it is not known exactly which state of nature will occur. However, there is
sufficient or partial information for us to estimate the chances of occurrence of various states of
nature. Two decisions criteria may be used:
- Maximize expected monetary value (EMV) or
- Minimize expected opportunity loss (EOL)
However, both consistently rank the various alternatives.

Definition-Opportunity loss is the amount a decision maker would lose by not taking the best
alternative. Opportunity loss is also known as the amount of regret. For any states of nature this
is the difference between the consequence of any alternative and the best possible alternative.
The decision making on the environment of risk can be for single or multi stage. A payoff table
is suitable for single stage analysis while a tree is preferred for multistage analysis.

Value of information
Perfect information – this is true or flawless information. In many organizations it can be
obtained at a fee. What is the maximum amount that one should be willing to pay for such
information?
(EV of PI) = EV with PI – Best EMV with existing (imperfect) information.
Example
Metaballs Manufacturing Company Ltd. is considering introducing two new products in the
market. The company has the following options:
Option 1: Introduce both products
Option 2: Introduce either of the products.
Option 3: Introduce none of the products, depending on their performance in the
market.
An analysis of the products’ likely performance indicates the probability of a good performance
as 30%, fair performance as 50% and poor performance as 20% and the sales revenue as shown
in the table below:
State of nature
Decision S1 S2 S3
(Sh. million) (Sh. million) (Sh. million)
Neither 0 0 0
Product 1 only 10.0 10 2.4
Product 2 only 8.4 4.8 2.4
Both 17.6 8.8 3.2
Where S1 = good performance

4
 S2 = fair performance
S3 = poor performance
Required:
(i) The expected monetary value for each decision.
(ii) The decision you would recommend.
(iii) The expected value of perfect information
(iv) Would the decision in (ii) change if EOL was used?

Solution:
i)
EMV (neither) = 0
EMV (1) = 10 × 0.3+10 × 0.5+2.4 × 0.2=8.48
EMV (2) =5.4
EMV (both) =10.32
ii) Decision: both products
(iii) The value of perfect information
EV with PI=(17.6 × 0.3) + (10 × 0.5) + (3.2 × 0.2) = 10.92
Expected maximum monetary value = 10.92
∴ Expected value of perfect information = 10.92 – 10.32 =0.6
iv) Regret table
State of nature
Decision S1 S2 S3
(Sh. million) (Sh. million) (Sh. million)
Neither 17.6 10 3.2
Product 1 only 7.6 0 0.8
Product 2 only 9.2 5.2 0.8
Both 0 1.2 0
EOL (neither) =10.92
EOL (1) =2.44
EOL (2) =5.52
EOL (both) =0.6
Decision: Both products hence no change