2010

Quantitative
Decision Making
Decision Analysis
This document will tell you the decision analysis. That how and in
which situations we can make decisions and what are the
methods by which we can take decision based on.

Document Outline:
Introduction
Six Steps in Decision Making
Types Of Decision Making Environments
Decision Making Under Uncertainty
Decision Making Under Risk
Sensitivity Analysis

Khawaja Naveed
Haider
University Of Management and Technology, Lahore Pakistan
Sunday, February 07, 2010
 QUANTITATIVE DECISION
MAKING

Introduction To Decision Analysis:

Decision making is very important part in any kind of business. To take
decision isn’t a big deal but to take the right decision on particular time with
complete alternatives and facts and figures is important. So here I would like
to discuss the “Decision Analysis” in the context of Quantitative Decision
Making, a part of statistical technique in the business.

We have the six steps in decision making.

1. Clearly define the problem

2. List all the possible alternatives

3. Identify all the possible outcomes or states of nature

4. List the payoffs or profits in front of each alternatives

5. Select one mathematical decision theory model

6. Apply the model

Example Question:

Ali wants to establish the factory in Pakistan. He has two outcomes or states
of nature which are favorable market and unfavorable market. He can either
establish last factory, small factory or he won’t go for the factory that means
do nothing. If he establishes a large factory, he could earn profit of PKR
200,000 or loose PKR 180,000 in unfavorable market. If he establishes the
small factory, he could earn PKR 100,000 in favorable market. He could lose
PKR 20,000 in unfavorable market. Now the purpose of establishing the
factory is to earn maximum profit.

naveedtaji@gmail.com Page Number: 2

Now if we shape the above example in a tabular form, we will get this kind of
table as given below.

STATES OF NATURE

FAVORABLE MARKET UNFAVORABLE

ALTERNATIVES
(PKR) MARKET (PKR)

ESTABLISH LARGE 200,000 -180,000

FACTORY

ESTABLISH SMALL 100,000 -20,000

FACTORY

DO NOTHING 0 0

Types of decision making environment:

We can make decisions in three environments.

1. Decision Making Under Certainty (We already know)

2. Decision Making Under Uncertainty (Probability is not given)

3. Decision Making Under Risk (Probability is given)

naveedtaji@gmail.com Page Number: 3

Decision Making Under Certainty:
In this environment, we already know what is best for us. We know the
alternatives and results and we choose the best naturally. Like if I have
100,000 Rupees and I want to invest, then I would like to open a bank’s
saving account which will give me more interest. Like I have the options of
5% annually, 10% annually and 12% annually. So I would automatically and
naturally choose 12%.

Decision Making Under Uncertainty:

In this environment, probability is not given. So we have the five situations
under which we can make decisions.

1. Maximax

2. Maximin

3. Criterion of Realism

4. Equally Likely

5. Minimax Regret

Let’s discuss the Maximax first.

Maximax:
We usually use the optimistic or positive thinking approach here. We will
choose which will give us maximum payoff.

Let’s have a look to the table that we have constructed before discussing the
environments.

States Of Nature
Alternatives Favorable Unfavorable Maximum in a
Market Market Row
Establish Large 200,000 -180,000 200,000
Factory
Establish Small 100,000 -20,000 100,000
Factory
Do Nothing 0 0 0

naveedtaji@gmail.com Page Number: 4

 If we are taking decision under Maximax, we will go for establishing a large
factory because the payoff of the large factory is higher among all the
payoffs of alternatives. So the shaded part is the answer.

Maximin:
Maximin is the totally opposite to the Maximax. In this environment, we take
the pessimistic approach. In this scenario, we will take the lowest payoff
which will be like decreasing the risk factor. We will see the value which will
be minimum in a row. Now take the above table.

States Of Nature
Alternatives Favorable Unfavorable Minimum in a
Market Market Row
Establish Large 200,000 -180,000 200,000
Factory
Establish Small 100,000 -20,000 100,000
Factory
Do Nothing 0 0 0
If we look at the above table, we will come to know that we will not establish
the factory or you can say do nothing. The lowest value in a row is 0 that is
kind of minimum risk.

Criterion Of Realism:
In this situation, we are given the percentage of either pessimistic or
optimistic. Like a question may contain that the person is 80% optimistic. So
the probability will be 0.80 and 0.20. 0.80 of optimistic and 0.20 of
pessimistic. We have to complete the 1 or 100% like 0.80+0.20=1

States Of Nature
Alternatives Favorable Unfavorable Criterion Of
Market Market Realism
Establish Large 200,000 -180,000 124,000
Factory
Establish Small 100,000 -20,000 76,000
Factory
Do Nothing 0 0 0
We can also call criterion of realism as Weighted Average. The formula of
calculating weighted average is given below.

Weighted Average = α (Maximum in a row) + (1-α) (Minimum in a row)

naveedtaji@gmail.com Page Number: 5

= 0.80(200,000) + (1-0.20)(-180000)

= 124,000

Note: the value of α will be given to you.

In this pattern, we will take the higher realism value among all the
alternatives. In this table we will take 124,000. That means we will establish
the large factory.

Equally Likely (Laplace):

Equally likely criterion uses the average outcome. One criterion that uses all
the payoffs for each alternative is equally likely. It is also called Laplace
decision criterion. This involves finding the average payoff for each
alternative and selective the alternative with the highest average. The
equally likely approach assumes that all probabilities of occurrence for the
states of nature are equal and thus each state of nature is equally likely.

States Of Nature
Alternatives Favorable Unfavorable Criterion Of
Market Market Realism
Establish Large 200,000 -180,000 10,000
Factory
Establish Small 100,000 -20,000 40,000
Factory
Do Nothing 0 0 0
In the above table, we will take the average of each alternative like for large
factory, we will add 200,000 and -180,000 and divide it on 2, we will get
10,000. Same is for small factory. After getting all the values, we will see that
the highest average is of establishing a small factory which has the value
rupees 40,000. So our decision will be establishing a small factory.

Minimax Regret:
This criterion is based on opportunity loss. We will discuss and draw the
opportunity loss table. Opportunity loss refers to the difference between the
optimal profit or payoff for a given state of nature and the actual payoff
received for a particular decision. In other words, it is the amount lost by not
picking the best alternative in a given outcome.

naveedtaji@gmail.com Page Number: 6

 Opportunity Loss Table

States Of Nature
Favorable Unfavorable
Market Market
200,000 -
200,000 0 - (-180,000)
200,000 -
100,000 0 - (-20,000)
200,000 - 0 0-0

States Of Nature
Favorable Unfavorable Maximum in a
Alternatives Market Market Row
Establish Large
Factory 0 180,000 180,000
Establish Small
Factory 100,000 20,000 100,000
Do Nothing 200,000 0 200,000

Here the minimum value of “Maximum in a Row” is 100,000 so our minimax

is 100,000. That means we will go for establishing a small factory. It is
because of minimizing the maximum opportunity loss.

Now we have completed the entire decision making criterion for decision
making under uncertainty.

naveedtaji@gmail.com Page Number: 7

Decision Making Under Risk:
In decision making under risk, I have already mentioned that we are given
the probabilities. That means we are aware of success or failure of decision to
some extent. We select the alternative with the highest expected value. We
also use the probabilities with the opportunity loss table to minimize the
expected opportunity loss. In decision making under risk, we will discuss
about Expected Monetary Value, Expected Value Of Perfect Information, and
Expected Opportunity Loss.

Expected Monetary Value:

The EMV means the long run average value of that decision. The EMV for an
alternative is just the sum of possible payoffs of the alternatives, each
weighted by the probability of that payoff occurring the expected value
material.

We can calculate EMV with the following formula.

EMV (alternative i) = (Payoff of first state of nature) X (probability of first

state of nature) + (Payoff of second state of nature) X (probability of second
state of nature) + …..(Payoff of i state of nature) X (probability of i state of
nature)

Now let’s calculate the EMV of every alternative. The probability is given that
is 0.5

EMV (Large Factory) = (0.5) (200,000) + (0.5) (-180,000 ) =

10,000

EMV (Small Factory) = (0.5) (100,000) + (0.5) (-20,000)

=40,000

EMV (Do Nothing) = (0.5) (0) + (0.5) (0) = 0

Now we can put these values in the table.

States Of Nature
Alternatives Favorable Unfavorable EMV
Market Market
Establish Large 200,000 -180,000 10,000
Factory
Establish Small 100,000 -20,000 40,000
Factory
Do Nothing 0 0 0
naveedtaji@gmail.com Page Number: 8
Probabilities 0.5 0.5
The largest value of EMV is of small factory. So Ali needs to go and establish
a small factory. The long run average of small factory is high.

Expected Value Of Perfect Information (EVPI):

Ali wants to produce another product and he is not aware of real market
situation. Ali was suggested by the scientific marketing company to avail
their services which will tell the whole scenario after conducting the survey
that whether is it good for Ali to produce another product or not according to
market situation? The firm is charging 65,000 rupees. The term EVPI is used
to check the worth of the information that the firm is giving in a particular
amount.

The formula for EVPI:

EVPI = expected value with perfect information – maximum EMV

Now here comes another question that what the hell Expected Value With
Perfect Information is? So here is the formula for EVwPI.

EVwPI = (best payoff for first state of nature) X (probability of first state of
nature) + (best payoff for second state of nature) X (probability of second
state of nature) ……. (best payoff for i state of nature) X (probability of i state
of nature)

States Of Nature
Alternatives Favorable Unfavorable EMV
Market Market
Establish Large 200,000 -180,000 10,000
Factory
Establish Small 100,000 -20,000 40,000
Factory
Do Nothing 0 0 0
Probabilities 0.5 0.5

Look at the table given above. We have two states of nature. One is favorable
market and the other one is unfavorable market. The best alternative for
favorable market is Establishing a large factory with the payoff of 200,000.
The best alternative for unfavorable market is Do Nothing with the payoff of 0
only. By using this information, we can put the values in our EVwPI formula.
naveedtaji@gmail.com Page Number: 9
 EVwPI = (200000)(0.5) + (0)(0.5) = 100,000
We have the EVwPI 100,000 rupees. Now we have also the formula of EVPI
where the value of EVwPI is used.

EVPI = EVwPI – Maximum EMV

= 100000 – 40000

= 60,000
Thus, Ali would be willing to pay for perfect information is 60,000 rupees and
the firm is charging 65,000 rupees. So Ali won’t go for the scientific
marketing firm because the value of information is not 65,000.

Expected Opportunity Loss:

The alternative approach to maximizing the EMV is to minimize expected
opportunity loss. We will construct opportunity loss table first. Then the EOL
is computed for each alternative by multiplying the opportunity loss by the
probability and adding these together.

States Of Nature
Alternatives Favorable Market Unfavorable Market
Establish Large Factory 0 180,000
Establish Small Factory 100,000 20,000
Do Nothing 200,000 0
Probabilities 0.5 0.5
The above table is describing the Opportunity Loss Table. Now we will
calculate the Expected Opportunity Loss for every alternative.

EOL For Large Factory = (0.5) (0) + (0.5) (180000) = 90000

EOL For Small Factory = (0.5) (100,000) + (0.5) (20,000) = 60000

EOL For Do Nothing = (0.5) (200,000) + (0.5) (0) = 100000

naveedtaji@gmail.com Page Number: 10

Now we will again construct the table of opportunity loss and will put the
values of EOL in front of each alternative.

States Of Nature
Alternatives Favorable Unfavorable EOL
Market Market
Establish Large 0 180,000 90,000
Factory
Establish Small 100,000 20,000 60,000
Factory
Do Nothing 200,000 0 100,000

Here we have the results in a tabular form. I have already told you that our
decision will be based on the minimum EOL. So the choice that Ali will select
is to establish a small factory as the Expected Opportunity Loss is less than
the EOL of Large factory and if he don’t construct a factory. One thing to
remember is that minimum EOL will always result in the same decision as
maximum EMV and that the EVPI will always equal to minim EOL. You can
compare the EOL of small factory with the EVPI that we have calculated. The
values of both the things are 60,000 rupees.

naveedtaji@gmail.com Page Number: 11

Sensitivity Analysis:
In previous pages, we have discussed and demonstrated the best decision
with known probabilities for Ali was to establish a small factory with an
expected value of 40,000 rupees. This conclusion depends on the values of
economic consequences and the two probability values of a favorable and an
unfavorable market. Sensitivity analysis investigates how our decision might
change given a change in the problem data. In this part, we investigate the
impact that a change in the probability value would have on the decision
facing Ali. We first would like to define the “P” which is probability of
favorable market and “1-P” is the probability of unfavorable market. Because
we have only two outcomes or you can say states of nature.

We can express the EMV in terms of P as shown in the following equations. I

will also show it on the graph.

EMV for large factory = 200000P – 180000(1-P)

= 380000P – 180000

EMV for small factory = 100000P – 20000(1-P)

= 120000P – 20000

EMV for do nothing = 0p + 0(1-P)

=0

Now we will define two points. Point 1 will be EMV do nothing = EMV small
factory, Point 2 will be EMV small plant = EMV large plant.

Point 1: 0 = 120000P – 20000

P = 20000/120000 = 0.167

Point 2: 120000P – 20000 = 380000P – 180000

260000P = 160000, P = 160000/260000 = 0.615

Now the next work is to put all the data in a tabular form.
naveedtaji@gmail.com Page Number: 12
 Best Range Of P
Alternative Value
Do Nothing Less than 0.167
Small Factory 0.167 - 0.615
Greater than
Large Factory 0.615

Here we have the tabular form of our analysis. If the value of P is less than
0.167 we will go for the option of Do Nothing.

Now it could become easier if we take this data on the graph.

As you can see in the graph, the best decision is to do nothing as los as P is
between 0 and the probability associated with point 1, where the EMV for
doing nothing is equal to the EMV for the small factory. When P is between
the probabilities for point 1 and 2, the best decision is to establish small
factory. Point 2 where EMV of small factory is equal to EMV of large factory.
When P is greater than probability of point 2, the decision will be establishing
a large factory.

naveedtaji@gmail.com Page Number: 13

Read Me:
You have read the whole document. I have taken
help from the notes given by my teacher. The topic
is very important and needs to be very easy so that
students could understand about the decision. So
by the grace of Allah and my teacher, I tried my
best in making you understand about the topic. At
least I am sure that I have written the whole
document in such a manner that everyone who has
a little sense about the statistical topics can
understand this whole document.
If you like this document, kindly refers to other
students, friends and to whom it may concern. Do
comment on this document if I am wrong anywhere
in this document. Do tell me. My email address is
written in the end of every page.

naveedtaji@gmail.com Page Number: 14