UNIT FOUR

4. DECISION THEORY/ANALYSIS
Unit objective:
After completing this unit, the learner should be able to:

 Describe the basic characteristics of decision theory problems

 Differentiate between decision analysis under certainty and uncertainty.
 Describe the different approaches (criteria) to decision making under complete
uncertainty.
1. Maximamax
2. Maximix
3. Minimax regret
4. Hurwicz
 Use decision tree as decision making tools.

? Dear learner, from your previous courses, can you define decision making? What steps are
involved in it? Can you mention conditions in which decision at different levels are made?

_____________________________________________________________________________
_____________________________________________________________________________
________________________________________________________

Dear learner, in the previous units dealing with LP, models were formulated and solved in order
to aid the manager in making decision. The solutions to the models were represented by values
for the decision variables. However, these LP models are formulated under the assumption that
certainty existed. In actual practice, however, many decision making situations occur under
conditions of uncertainty. For example, the demand for a product may be not 100 units next
week, but 50 or 200 units, depending on the market (which is uncertain).

5.1. Characteristics of Decision Theory

Decision theory problems are characterized by the following:
1. List of alternatives: are a set of mutually exclusive and collectively exhaustive decisions
that are available to the decision maker (some times, not always, one of these alternatives
will be to “do nothing”.)
2. States of nature: - the set of possible future conditions, or events, beyond the control of
the decision maker, that will be the primary determinants of the eventual consequence of
the decision. The states of nature, like the list of alternatives, must be mutually exclusive
and collectively exhaustive.
3. Payoffs: - the payoffs might be profits, revenues, costs, or other measures of value.
Usually the measures are financial. Usually payoffs are estimated values. The more
accurate these estimates, the more useful they will be for decision making purposes and
the more likely, it is that the decision maker will choose an appropriate alternative. The
number of payoffs depends on the number of alternative/state of nature combination.
 4. Degree of certainty: - the approach often used by a decision maker depends on the
degree of certainty that exists. There can be different degrees of certainty. One extreme is
complete certainty and the other is complete uncertainty. The later exists when the
likelihood of the various states of nature are unknown. Between these two extremes is
risk (probabilities are unknown for the states of nature). Knowledge of the likelihood of
each of the states of nature can play an important role in selecting a course of active.
5. Decision criteria: - the decision maker’s attitudes toward the decision as well as the
degree of certainty that surrounds a decision. Example; maximize the expected payoffs.

5.2. THE PAYOFF TABLE

A payoff table is a device a decision maker can use to summarize and organize information
relevant to a particular decision. It includes a list of alternatives, the possible future states of
nature, and the payoffs associated with each of the alternative/state of nature combinations. If
probabilities for the states of nature are available, these can also be listed. The general format
of the table is illustrated below:

States of nature
S1 S2 S3
A1 V11 V12 V13
Alternatives A2 V21 V22 V23
A3 V31 V32 V33
where:
Ai = the ith alternative
Sj = the jth states of nature
Vij = the value or payoff that will be realized if alternative i is chosen and event j
occurs.
Decision situations can be categorized in to three classes: Situation of certainty, Situations
where probabilities can not be assigned to future occurrences and Situations where
probabilities can be assigned to future occurrences. In this chapter we will discuss each of
these classes of decision situations separately.

5.3. DECISION MAKING UNDER CERTAINTY

The simplest of all circumstances occurs when decision making takes place in an
environment of complete certainty. When a decision is made under conditions of complete
certainty, the attention of the decision maker is focused on the column in the payoff table that
corresponds to the state of nature that will occur. The decision maker then selects the
alternative that would yield the best payoff, given that state of nature.
EXAMPLE
The following payoff table provides data about profits of the various states of
nature/alternative combination.
S1 S2 S3
A1 4 16 12
A2 5 6 10
A3 -1 4 15
If we know that S2 will occur, the decision maker then can focus on the first raw of the
payoff table. Because alternative A1 has the largest profit (16), it would be selected.

5.4. DECISION MAKING UNDER COMPLETE UNCERTAINTY (With out

probabilities)
Under complete uncertainty, the decision maker either is unable to estimate the probabilities
for the occurrence of the different state of nature, or else he or she lacks confidence in
available estimates of probabilities, and for that reason, probabilities are not included in the
analysis.
A decision making situation includes several components- the decision themselves and the
actual event that may occur future, known as state of nature. At the time the decision is made,
the decision maker is uncertain which state of nature will occur in the future, and has no
control over them.

Decisions made under these circumstances are at the opposite end of the spectrum from the
certainty case just mentioned. Once the decision has been organized in to a payoff table,
several criteria are available making the actual decision. There are several approaches
(criteria) to decision making under complete uncertainty. Some of these discussed in this
section include: maximax, maximin,minimax regret, Hurwicz, and equal likelihood.

5.4.1. MAXIMAX
With the maiximax criterion, the decision maker selects the decision that will result in the
maximum of the maximum payoffs.( In fact this is how this criterion derives its name-
maximum of maximum). Tha maximax is very optimistic. The decision maker assumes that
the most favorable state of nature for each decision alternative will occur. For example, the
investor would optimistically assume that good economic conditions will prevail in the
future. The best payoff for each alternative is identified, and the alternative with the
maximum of these is the designated decision.
For the previous problem:
S1 S2 S3 Row Maximum
A1 4 16 12 16*maximum
A2 5 6 10 10
A3 -1 4 15 15
Decision: A1 will be chosen.
Note: If the pay off table consists of costs instead of profits, the opposite selection would
be indicated: The minimum of minimum costs. For the subsequent decision criteria we
encounter, the same logic in the case of costs can be used.

? Dear learner, how would the decision in the above example change if the values in
the table stand for costs instead of profit?

_________________________________________________
_________________________________________________
_________________________________________________
 5.4.2. Maximin Criteria
This approach is the opposite of the previous one, i.e. it is pessimistic. This strategy is a
conservative one; it consists of identifying the worst (minimum) payoff for each alternative,
and, then, selecting the alternative that has the best (maximum) of the worst payoffs. In
effect, the decision maker is setting a floor on the potential payoff by selecting maximum of
the minimum; the actual payoff can not be less than this amount. It involves selecting best of
the worst.

For the previous problem:

S1 S2 S3 Row minimum
A1 4 16 12 4
A2 5 6 10 5*maximum
A3 -1 4 15 -1
Decision: A2 will be chosen.
Note: If it were cost, the conservative approach would be to select the maximum cost for
each decision and select the minimum of these costs.

? Dear learner, how would the decision in the above example change if the values in
the table stand for costs instead of profit?

________________________________________________
________________________________________________
______________________________________________
5.4.3. MINIMAX REGRET
Both the maximax and maximin strategies can be criticized because they focus only on a
single, extreme payoff and exclude the other payoffs. Thus, the maximax strategy ignores the
possibility that an alternative with a slightly smaller payoff might offer a better overall
choice. For example, consider this payoff table:
S1 S2 S3 Row Max.
A1 -5 16 -10 16*max
A2 15 15 15 15
A3 15 15 15 15
A similar example could be constructed to demonstrate comparable weaknesses of the
maximin criterion, which is also due to the failure to consider all payoffs.
An approach that does take all payoffs in to consideration is Minimax regret. In order to use
this approach, it is necessary to develop an opportunity loss table. The opportunity loss
reflects the difference between each payoff and the best possible payoff in a column (i.e.,
given a state of nature). Hence, opportunity loss amounts are found by identifying the best
payoff in a column and, then, subtracting each of the other values in the column from that
payoff. Therefore, this decision avoids the greatest regret by selecting the decision alternative
that minimizes the maximum regret.
 EXAMPLE:
S1 S2 S3
A1 4 16 12
A2 5 6 10
A3 -1 4 15
opportunity loss table:
S1 S2 S3
A1 5-4=1 16-16=0 15-12=3
A2 5-5=0 16-6=10 15-10=5
A3 5-(-1)=6 16-4=12 15-15=0
The values in an opportunity loss table can be viewed as potential “regrets” that might be
suffered as the result of choosing various alternatives. A decision maker could select an
alternative in such a way as to minimize the maximum possible regret. This requires
identifying the maximum opportunity loss in each row and, then, choosing the alternative
that would yield the best (minimum) of those regrets.
S1 S2 S3 Max. Loss
A1 5-4=1 16-16=0 15-12=3 3*minimum
A2 5-5=0 16-6=10 15-10=5 10
A3 5-(-1)=6 16-4=12 15-15=0 12

Decision: A1 will be chosen.

? Dear learner, how would the decision in the above example change if the values in
the table stand for costs instead of profit?

_______________________________________________
_______________________________________________
_______________________________________________
Although this approach makes use of more information than either Maximin or Maximax, it
still ignores some information, and, therefore, can lead to a poor decision.
EXAMPLE:
Opportunity loss table
S1 S2 S3 S4 Max. Loss
A1 0 0 0 24 24
A2 15 15 15 0 15*minimum
A3 15 15 15 0 15*minimum
5.4.4. PRINCIPLE OF INSUFFICIENT REASON/ Equal likelihood/ Laplace
The Minimax regret criterion’s weakness is the inability to factor row differences. Hence,
sometimes the minimax regret strategy will lead to a poor decision because it ignores certain
information.
The principle of insufficient reason offers a method that incorporates more of the information. It
treats the states of nature as if each were equally likely, and it focuses on the average payoff for
each row, selecting the alternative that has the highest row average.

EXAMPLE
S1 S2 S3 S4 S5 Row Average
A1 28 28 28 28 4 23.2*maximum
A2 5 5 5 5 28 9.6
A3 5 5 5 5 28 9.6
Decision: A1 is selected
The basis for the criterion of insufficient reason is that under complete uncertainty, the decision
maker should not focus on either high or low payoffs, but should treat all payoffs (actually, all
states of nature), as if they were equally likely. Averaging row payoffs accomplishes this.

? Dear learner, how would the decision in the above example change if the values in the table
stand for costs instead of profit?

_________________________________________________________________________________
_________________________________________________________________________________
_____________________________________________________________

5.4.5.The Hurwitz Criterion

The Hurwitz criterion strikes a compromise between the maximax and maximin criterion. The
principle underlying this decision criterion is that the decision maker is neither totally optimistic,
nor totally pessimistic. With Hurwitz criterion, the decision payoffs are weighted by a coefficient
of optimism, a measure of a decision maker’s optimism. The coefficient of optimism, which is
defined as, is between zero and one (0< <1). If  = 1, then the decision maker is said to be
completely optimistic, if = 0, then the decision maker is completely pessimistic. Given this
definition, if  is coefficient of optimism, 1- is coefficient of pessimism.

The Hurwitz criterion requires that for each alternative, the maximum payoff is multiplied by 
and the minimum payoff be multiplied by 1-.
Example: If  = 0.4 for the above example,

A1 = (0.4x16) + (0.6x4)
= 8.8
A2 = (0.4x10) + (0.6x5)
 =7
A3 = (0.4x15) – (0.6x1)
= 5.4
Decision: A1 is selected

A limitation of Hurwicz criterion is the fact that  must be determined by the decision maker.
Regardless of how the decision maker determines, it is still a completely a subjective measure
of the decision maker’s degree of optimism. Therefore, Hurwicz criterion is a completely
subjective decision making criterion.

? Dear learner, can you mention conditions under which Hurwicz criterion criteria can be
considered as maximin or minimax criterion

________________________________________________________________________________
________________________________________________________________________________
_____________________________________________________________
5.5. DECISION MAKING UNDER RISK (WITH PROBABILITIES)

Dear learner, the decision making criteria just presented were based on the assumption that no
information regarding the likelihood of the states of the nature was available. Thus, no
probabilities of occurrence were assigned to the states of nature, except in the case of the equal
likely hood criterion.

It is often possible for the decision maker to know enough about the future state of nature to
assign probabilities to their occurrences. The term risk is often used in conjunction with partial
uncertainty, presence of probabilities for the occurrence of various states of nature. The
probabilities may be subjective estimates from managers or from experts in a particular field, or
they may reflect historical frequencies. If they are reasonably correct, they provide the decision
maker with additional information that can dramatically improve the decision making process.
Given that probabilities can be assigned, several decision criteria are available to aid the decision
maker. Some of these are discussed below.

5.5.1. EXPECTED MONETARY VALUE (EMV)

The EMV approach provides the decision maker with a value which represents an average payoff
for each alternative. The best alternative is, then, the one that has the highest EMV. The average
or expected payoff of each alternative is a weighted average:
k
EMVi = Σ Pj.Vij
i=1
Where:
 EMVi = the EMV for the ith alternative
Pi = the probability of the ith state of nature
Vij = the estimated payoff for alternative i under state of nature j.
Note: the sum of the probabilities for all states of nature must be 1.

EXAMPLE:

Probability 0.20 0.50 0.30

S1 S2 S3 Expected payoff
A1 4 16 12 12.40*maximum
A2 5 6 10 7
A3 -1 4 15 6.30
Decision: A1 will be chosen.

Note that it does not necessarily follow that the decision maker will receive a payoff
equal to the expected monetary value of a chosen alternative. Similarly, the expected payoffs
for either of the other alternatives do not equal any payoffs in those rows. What, then, is the
interpretation of the expected payoff? Simply a long-run average amount; the approximate
average amount one could reasonably anticipate for a large number of identical situations.

5.5.2. Expected Opportunity Loss (EOL)

The table of opportunity loss is used rather than a table of payoffs. Hence, the opportunity
losses for each alternative are weighted by the probabilities of their respective state of nature
to compute a long run average opportunity loss, and the alternative with the smallest
expected loss is selected as the best choice.
EOL (A1) = 0.20(1) + 0.50(0) + 0.30(3) = 1.10 *minimum
EOL (A2) = 0.20(0) + 0.50(10) + 0.30(5) = 6.50
EOL (A3) = 0.20(6) + 0.50(12) + 0.30(0) = 7.20

Note: The EOL approach resulted in the same alternative as the EMV approach
(Maximizing the payoffs is equivalent to minimizing the opportunity losses).

5.5.3. Expected Value of Perfect Information (EVPI)

It can some times be useful for a decision maker to determine the potential benefit of knowing
for certain which state of nature is going to prevail. The EVPI is the measure of the difference
between the certain payoffs that could be realized under a condition involving risk.
If the decision maker knows that S1 will occur, A2 would be chosen with a payoff of $5.
Similarly for S2 $16 (for A1) and for S3, $15 (with A3) would be chosen.
Hence, the expected payoff under certainty (EPC) would be:
EPC = 0.20(5) + 0.50(16) + 0.30(15) = 13.50
The difference between this figure and the expected payoff under risk (i.e., the EMV) is the
expected value of perfect information. Thus:
EVPI = EPC – EMV
= 13.50 – 12.40 = 1.10
Note: The EVPI is exactly equal to the EOL. The EOL indicates the expected opportunity
loss due to imperfect information, which is another way of saying the expected payoff that could
be achieved by having perfect information.

Note: The expected value approach is particularly useful for decision making when a number
of similar decisions must be made; it is a long-run approach. For one-shot decisions, especially
major ones, other methods (perhaps, maximax or maximin) may be preferable. In addition, non
monetary factors, although not included in a payoff table, may be of considerable importance.
Unfortunately, there is no convenient way to include them in an expected value analysis.

??? Dear learner, can you make differences between decision making situations under uncertainty and
risk?

__________________________________________________________________________________
__________________________________________________________________________________
________________________________________________________________

5.6. DECISION TREES

Decision trees some times are used by decision makers to obtain a visual portrayal of decision
alternatives and their possible consequences. The term gets its name from the tree-like
appearance of the diagram.
Decision tree format:

Decision tree, like probably tree is composed of squares, circles, and lines:
 The squares indicate decision points
 Circles represent chance events( circles and squares are called nodes)
o The lines (branches) emanating from squares represent alternatives.
o The lines from circles represent states of nature
 The tree is read from right to left.
It should be noted that although decision trees represent an alternative approach to payoff tables,
they are not commonly used for problems that involve a single decision. Rather, their greatest
benefit lies in portraying sequential decisions (i.e., a series of chronological decisions). In the
case of a single decision, constructing a decision tree can be cumbersome and time consuming.

Example
Pay off table for Real Estate investment
State of Nature
Good economic Good economic
Decision conditions conditions
(Purchase) 0.6 0.4

Apartment building 50,000 30,000

Office building 100,000 -40,000
Warehouse 30,000 10,000

The decision tree for the above example will be:

$30,000
Good economic $50,000
conditions (0.6

Apartment 2
building Poor economic
conditions (0.4)

Purchase Good economic $100,000

conditions (0.60
Office building
1 3
Poor economic -$40,000

conditions (0.4)
Good economic
Warehouse conditions (0.60 $30,000
4

Poor economic $10,000

conditions

The circles ( ) and squares (  ) in the above figure are referred to as nodes. The
squares are decision nodes, and the branches emanating from a decision node reflect the
alternative decisions possible at that point. For example, in the above figure, node 1
signifies a decision to purchase an apartment building. Office building, or ware house. The
circles are probability nodes, and the branches emanating from them indicate the state of
 nature that can occur: good economic conditions or poor economic conditions. The
decision tree represents the sequence of events in a decision situation. First, one of the
three decision choices is selected at node 1. Depending on the branch selected, the decision
maker arrives at probability node 2, 3, or 4, where one of the states of nature will prevail,
resulting in one of six possible payoffs.

Determining the best decision using a decision tree involves computing the expected value
at each probability node. This is accomplished by starting with the final outcomes
(payoffs) and working backward through the decision tree toward node 1. First, the
expected value of the payoffs is computed at each probability node.
EV(node 2) = .60($ 50,000) + .40($ 30,000) = $42,000
EV(node 3) = .60($100,000) + .40($-40,000) = $44,000
EV(node 4) = .60($ 30,000) + .40($ 10,000) = $22,000
These values are now shown as the expected payoffs from each of the three branches
emanating from node 1 in figure below. Each of these three expected values at nodes 2, 3,
and 4 is the outcome of a possible decision that can occur at node 1. Moving toward node
1, we select the branch that comes from the probability node with the highest expected
payoff. In figure below, the branch corresponding to the highest payoff, $44,000 is from
node 1 to node 3. This branch represents the decision to purchase the office building. The
decision to purchase the office building, with an expected payoff of $44,000, is the same
result we achieved earlier using the expected value criterion. In fact, when only one
decision is to be made (i.e., there is not a series of decisions), the decision tree will always
yield the same decision and expected payoff as the expected value criterion. As a result, in
these decision situations the decision tree is not very useful. However, when a sequence or
series of decisions is required, the decision tree can be very useful.

Good economic
$42,000 Conditions (0.6)

$50,000

Apartment 2
building
Poor economic
Conditions(0.4) $30,000
$44,000
Good economic
Office building 3 Conditions (0.6)
1 $100,000
Purchase
Poor economic
Conditions -$40,000

Good economic
$22,000
Warehouse Conditions (0.6)
$30,000
4
Poor economic
Conditions(0.4) $10,000
5.6.1. Sequential Decision Trees
As noted above, when a decision situation requires only a single decision, an expected value
payoff table will yield the same result as a decision tree. However, a payoff table is usually
limited to a single decision situation like our real estate investment example. If a decision
situation requires a series of decisions, then a payoff table cannot be created and a decision
tree becomes the best method for decision analysis.

In order to demonstrate the use of a decision tree for a sequence of decisions, we will alter
our real estate investment example to encompass a ten-year period during which several
decisions must be made. In this new example, the first decision facing the investor is whether
to purchase an apartment building or land. If the investor purchases the apartment building,
two states of nature are possible. Either the population of the town will grow (with a
probability of 0.60), or the population will not grow (with a probability of 0.40). Either state
of nature will result in a payoff. On the other hand, if the investor chooses to purchase land,
three years in the future another decision will have to be made regarding the development of
the land. The decision tree for this example, shown in figure below, contains all the
pertinent data, including decisions, states of nature, probabilities, and payoffs.

At decision node 1 in figure below, the decision choices are to purchase an apartment
building and to purchase land. Notice that the cost of each venture ($800,000 and $200,000,
respectively) is shown in parentheses. If the apartment building is purchased,

two states of nature are possible at probability node 2. The town may exhibit population
growth, with a probability of .60, or there may be no population growth or a decline, with a
probability of .40. If the population grows, the investor will achieve a payoff of $2,000,000
over a ten-year period. (Note that this whole decision situation encompasses a ten-year time
span.) However, if no population growth occurs, a payoff of only $225,000 will result.

If the decision is to purchase land, two states of nature are possible at probability node 3.
These two states of nature and their probabilities are identical to those at node 2; however,
the payoffs are different. If population growth occurs for a three-year period, no payoff will occur,
but the investor will make another decision at node 4 regarding development of the land. At that
point either apartment will be built at a cost of $800,000 or the land will be sold with a payoff of
$450,000. Notice that the decision situation at node 4 can occur only if population growth occurs
first. If no population growth occurs at node 3, there is no payoff and another decision situation
becomes necessary at node 5: the land can be developed commercially at a cost of $600,000 or
the land can be sold for $210,000. (Notice that the sale of the land results in less profit if there is
no population growth than if there is population growth.)

If the decision at decision node 4 is to build apartments, two states of nature are possible. The
population may grow, with a conditional probability of .80, or there may be no population
growth, with a conditional probability of .20. The probability of population growth is higher (and
the probability of no growth is lower) than before because there has already been population
growth for the first three years, as shown by the branch from node 3 to node 4. The payoffs for
these two states of nature at the end of the ten-year period are $3,000,000 and $700,000,
respectively, as shown in figure below.
 If the investor decides to develop the land commercially at node 0. 5, then two states of nature
can occur. Population growth can occur, with a probability of .30 and an eventual payoff of
$2,300,000, or no population growth can occur, with a probability of .70 and a payoff of
$1,000,000. The probability of population growth is low (i.e., .30) because there has already
been no population growth, as shown by the branch from node 3 to node 5.

Sequential Decision Tree

$ 2,000,000

.60
2 $ 225,000
.40 $ 3,000,000
.80
6
6
.2
0,
1
00 $ 700,000
4 0, $ 450,000
Sell land
00
0
.3
3 0, $ 2,000,000
7
00
0,
 .60
.40

This decision situation encompasses several sequential decisions that can be analyzed using the
decision tree approach outlined in our earlier (simpler) example. As before, we start at the end of
the decision tree and work backward toward a decision at node 1.
First we must compute the expected values at nodes 6 and 7.
EV (node 6) = .80($3,000,000) + .20($ 700,000) = $2,540,000
EV (node 7) = .30($2,300,000) + .70($1,000,000) = $1,390,000
Both of these expected values (as well as all other nodal values) are shown in boxes in Figure 4.4.
At decision nodes 4 and 5, we must make a decision. As with a normal payoff table, we make
the decision that results in the greatest expected value. At node 4 we have a choice between two
values: $1,740,000, the value derived by subtracting the cost of building an apartment building
($800,000) from the expected payoff of $2,540,000, or $450,000, the expected value of selling
the land computed with a probability of 1.0. The decision is to build the apartment building, and
the value at node 4 is $1,740,000.

This same process is repeated at node 5. The decisions at node 5 result in payoffs of $790,000
(i.e., $1,390,000 - 600,000 = $790,000) and $210,000. Since the value $790,000 is higher, the

decision is to develop the land commercially.

Sequential Decision Tree with Nodal Expected Values

$ 2,000,000

$1,290,000

.60
Purchase 2 $ 225,000 $ 7540,000
$ 3,000,000
Apartment .40 .80
Building 6
6
.2 .20
$1,740,000 0,
1
$1,160,000 00 $ 700,000
4 0, $ 450,000
Sell land
00
$1,390,000 0
.60 .3
3 0, $ 2,300,000
.40 7
00
$1,360,000 $790,000 0,
.7
00
0,
5 0
00
 $ 1,000,000

Next we must compute the expected values at nodes 2 and 3.

EV (node 2) = .60($2,000,000) + .40($225,000) = $1,290,000

EV (node 3) = .60($1,740,000) + .40($790,000) = $1,360,000
(Note that the expected value for node 3 is computed from the decision values previously
determined at nodes 4 and 5.)
Now we must make the final decision for node 1. As before, we select the decision with
the greatest expected value after the cost of each decision is subtracted.

Apartment building: $1,290,000 - 800,000 = $ 490,000

Land: $1,360,000 - 200,000 = $1,160,000
Since the highest net expected value is $1,160,000, the decision is to purchase land and the
payoff of the decision is $1,160,000.
This example demonstrates the usefulness of decision trees for decision analysis. The
decision tree allows the decision maker to see the logic of decision making, because it
provides a picture of the decision process. Decision trees can be used for decision problems
more complex than the example above without too much difficulty.

5.6.2. DECISION ANALYSIS WITH ADDITIONAL INFORMATION

Earlier in this unit we discussed the concept of the expected value of perfect information. We
noted that if perfect information could be obtained regarding which states of nature would
occur in the future the decision maker could obviously make better decisions. Although
perfect information about the future is rare, it is often possible to gain some amount of
additional (imperfect) information that will improve decisions.

In this section we will present a process for using additional information m the decision-
making process by applying Bayesian analysis, a probabilistic technique. This process will
be demonstrated using the real estate investment example employed throughout this
chapter. To
briefly review this example, a real estate investor is considering three alternative in-
vestments, which will occur under one of two possible economic conditions (states of
nature) shown in table above.
 Payoff Table for the Real Estate Investment Example
States of Nature
Decision Good Economic Conditions Poor Economic Conditions
(purchase) .60 .40

Apartment building $ 50,000 $ 30,000

Office building 100,000 - 40,000
Warehouse 30,000 10,000
Recall that using the expected value criterion, we found the best decision to be the purchase of
the office building, with an expected value of $44,000. We also computed the expected value of
perfect information to be $28,000. Therefore, the investor would be willing to pay up to $28,000
for information about the states of nature, depending on how close to perfect the information
was.

Now suppose that the investor has decided to hire a professional economic analyst who will
provide additional information about future economic conditions. The analyst is constantly
researching the economy, and the results of this research are what the investor will be
purchasing.
The economic analyst will provide the investor with a report predicting one of two outcomes.
The report will be either positive, indicating that good economic condition are most likely to
prevail in the future, or negative, indicating that poor economic conditions will probably occur.
Based on the analyst's past record in forecasting future economic conditions, the investor has
determined conditional probabilities of the different report outcomes given the occurrence of each
state of nature in the future. We will use the following notations to express these conditional
probabilities:
g = good economic conditions
p = poor economic conditions
P = positive economic report
N = negative economic report

The conditional probability of each report outcome given the occurrence of each state of nature
is shown below.
P(P/g) = .80
P(N/g) = .20
P(P/p) = .10
P(N/p) = .90

For example, if future economic conditions are in fact good (g), the probability that a positive
report (P) will have been given by the analyst, P(P/g), is .80. The other three conditional
probabilities can be interpreted similarly. Notice that these probabilities indicate that the analyst
is a relatively accurate forecaster of future economic conditions.
The investor now has quite a bit of probabilistic information available - not only the conditional
probabilities of the report, but also the prior probabilities that each state of nature will occur.
These prior probabilities that good or poor economic conditions will occur in the future are
P(g) = .60
P(p) = .40
Given the conditional probabilities, the prior probabilities can be revised to form posterior
probabilities by means of Bayes's rule. If we know the conditional probability that a positive
report was presented given that good economic conditions prevail, P(P / g), the posterior
probability of good economic conditions given a positive report, P(g/P), can be determined using
Bayes's rule, as follows.
P(g /P) = P(P / g)P(g)
P(P/g)P(g) + P(P/p)P(p)
(.80) (.60)
(.80) (.60) + (.10) (.40)
= .923
The prior probability that good economic conditions will occur in the future is .60. However,
by obtaining the additional information of a positive report from the analyst, the investor can
revise the prior probability of good conditions to a .923 probability that good economic
conditions will occur. The remaining posterior (revised) probabilities are
P(g/N) = .250
P(p /P) = .077
P(p/N) = .750

5.6.3. Decision Trees with Posterior Probabilities

The original decision tree analysis of the real estate investment example is shown in Figures 4.1
and 4.2. Using these decision trees, we determined that the appropriate decision was the purchase
of an office building, with an expected value of $44OOO. However, if the investor hires an
economic analyst, the decision regarding which piece of real estate to invest in will not be made
until after the analyst presents the report. This creates an additional stage in the decision-making
process, which is shown in the decision tree in Figure 4.5.
The decision tree shown in Figure 4.5 differs in two respects from the decision trees in Figures
4.1 and 4.2. The first difference is that there are two new branches at the beginning of the
decision tree that represent the two report outcomes. Notice, however, that given either report
outcome, the decision alternatives, the possible states of nature, and the payoffs are the same as
those in the first two figures.

The second difference is that the probabilities of each state of nature are no longer the prior
probabilities given in Figure 4.1; instead they are the revised posterior probabilities computed in
the previous section using Bayes's rule. If the economic analyst issues a positive report, then the
upper branch in figure below (from node 1 to node 2) will be taken.

If an apartment building is purchased (the branch from node 2 to node 4), the probability of
good economic conditions is .923, whereas the probability of poor conditions is .077. These are
the revised posterior probabilities of the economic conditions given a positive report. However,
before we can perform an expected value analysis using this decision tree, one more piece of
probabilistic information must be determined-the initial branch probabilities of a positive and a
negative economic report.

The probability of a positive report, P(P), and of a negative report, P(N), can be determined
according to the following logic. Recall from Chapter 10 that the probability that two dependent
events, A and B, will both occur is

P(AB) = P(A/B) P(B)

Decision Tree with Posterior Probabilities

$50,000
Pig p1=.923
Apartment
building 4
P/P/p) = 077 $30,000
Office building PigP)= 923
2
$100,00
Positive 5
report P(P/P)=077 -$ 40,000
Warehouse P(g/p)= 923
$ 30,000
6

1 P(p/p)= 077 $ 10,00

P(g/N)= 250
$ 50,00
Apartment
Building 7
Negative $30,000
report Office building P(p/N) = 750
8 P(g/N) +250 $100,000
3 -$40,000
P(g/N)= . 750
Warehouse P(g/N)= .250 $ 30,000
9
$ 10,000
P(p/N)= 750

If event A is a positive report and event B is good economic conditions, then according to the
above formula,
P(Pg) = pcp / g)P(g)

We can also determine the probability of a positive report and poor economic conditions the
same way.
P(Pp) = P(P/p)P(p)

Next consider the two probabilities P(Pg) and P(Pp). These are, respectively, the probability of a
positive report and good economic conditions and the probability of a positive report and poor
economic conditions. These two sets of occurrences are mutually exclusive, since both good and
poor economic conditions cannot occur simultaneously in the immediate future. Conditions will
be either good or poor, but not both. To determine the probability of a positive report, we add the
mutually exclusive probabilities of a positive report with good economic conditions and a
positive report with poor economic conditions, as follows.
P(P) = P(Pg) + P(Pp)
Now, if we substitute into this formula the relationships for P(Pg) and P(Pp) determined
earlier, we have
P(P) = P(P/g)P(g) + P(P/p)P(p)
You might notice that the right-hand side of this equation is the denominator of the Bayesian
formula we used to compute P(g/P) in the previous section. Using the conditional and prior
probabilities that have already been established, we can determine that the probability of a
positive report is
P(P) = P(P/g)P(g) + P(P/p)P(p) = (.80) (.60) + (.10) (.40)
= .52
Similarly, the probability of a negative report is
P(N) = P(N/g)P(g) + P(N/p)P(p)
= (.20) (.60) + (.90) (.40)
= .48
Now we have all the information needed to perform a decision tree analysis. The decision tree
analysis for our example is shown in Figure 11.6. To see how the decision tree analysis is
conducted, consider the result at node 4 first. The value $48,460 is the expected value of the
purchase of an apartment building given both states of nature. This expected value is computed
as follows.
EV (Apartment building) = $50,000(.923) + 30,000(.077) = $48,460

The expected values at nodes 5, 6, 7, 8, and 9 are computed similarly.

The investor will actually make the decision as to which investment to make at nodes 2 and 3. It
is assumed that the investor will make the best decision in each case. Thus, the decision at node
2 will be to purchase an office building, with an expected value of $89,212; the decision at node
3 will be to purchase an apartment building, with an expected value of $35,000. These two
results at nodes 2 and 3 are referred to as decision strategies. They represent a plan of decisions to
be made given either a positive or a negative report from the economic analyst.
The final step in the decision tree analysis is to compute the expected value of the decision
strategy given that an economic analysis is performed. This expected value, shown as $63,190 at
node 1 in Figure 11.6, is computed as follows.
EV (strategy) = $89,212(.52) + 35,000(.48) = $63,190
 Decision Tree Analysis

$50,000

$30,000

$100,000

-$40,000

$30,000

$10,000

$50,000

$30,000

$100,000

-$40,000

$30,000

$10,000
$15,000
P (p/N) =.750

This amount, $63,190, is the expected value of the investor's decision strategy given that a report
forecasting future economic condition is generated by the economic analyst.

5.6.4. Computing Posterior Probabilities with Tables

One of the difficulties that can occur with this type of decision analysis with additional
information is that as the size of the problem increases (i.e., as we add more decision
alternatives and states of nature) the application of Bayes's rule to compute the posterior
probabilities becomes more complex. In such cases, the posterior probabilities can be computed
using tables. This tabular approach will be demonstrated with our real estate investment
example. The table for computing posterior probabilities for a positive report and P(P) is initially
set up as shown in table below.
The posterior probabilities for either state of nature (good or poor economic conditions) given a
negative report are computed similarly.
 (1) (2) (3) (4) (4)
States Prior Conditional Prior Posterior
Of Nature Probabilities Probabilities Probability x Probabilities:
Conditional (4)   (4)
Probability:
(2) x (3)
Good P(g) = .6 P(P/g) = 80 P(Pg) = .48 .48
Conditions P(g/P) =  .923
.52
Proof P(P) = .4 P(P/g) = 40 P( Pp )  .04 .04
Pp/P)= .  .077
Conditions   P( P)  .52 52

No matter how large the decision analysis, the steps of this tabular approach can be
followed the same way as in this relatively small problem. This approach is more
systematic than the direct application of Bayes's rule, making it easier to compute the
posterior probabilities for larger problems.

5.6.5. The Expected Value of Sample Information

Recall that we computed the expected value of our real estate investment example to be
$44,000 when we did not have any additional information. After obtaining the additional
information provided by the economic analyst, we computed an expected value of
$63,190 using the decision tree in Figure 4.6. The difference between these two expected
values is called the expected value of sample information (EVSI), and it is computed as follows.
EVSI = EV with information - EV with out information
For our example, the expected value of sample information is
EVSI = $63,190 - 44,000
= $19,190
This means that the real estate investor would be willing to pay the economic analyst up
to $19,190 for an economic report that forecasted future economic conditions.
After we computed the expected value of the investment without additional information,
we computed the expected value of perfect information, which equaled $28,000. However, the
expected value of the sample information was only $19,190. This is a logical result, since
it is rare than absolutely perfect information can be determined. Since the additional
information that is obtained is less than perfect, it will be worth less to the decision maker.
We can determine how close to perfect our sample information is by computing the
efficiency of sample information as follows.

EVSI
Efficiency = EVPI

= $ 19,190
28,000
=.68
Thus, the analyst's economic report is viewed by the investor to be 68% as efficient as
perfect information. In general, a high efficiency rating indicates that the information is
very good, or close to being perfect information, and a low rating indicates that the addi-
tional information is not very good. For our example, the efficiency of .68 is relatively
high; thus it is doubtful that the investor would seek additional information from an
alternative source. (However, this is usually dependent on how much money the decision
maker has available to purchase information.) If the efficiency had been lower, however,
the investor might seek additional information elsewhere.

Summary
 Decision theory problems are characterized by, list of alternatives, states of nature,
payoffs, degree of certainty, decision criteria.

 Decision situations can be categorized in to three classes: Situation of certainty,

Situations where probabilities can not be assigned to future occurrences and Situations
where probabilities can be assigned to future occurrences.

 There are several approaches (criteria) to decision making under complete uncertainty.
Some of these discussed in this section include: maximax, maximin,minimax regret,
Hurwitz, and equal likelihood.

 Decision making under risk (with probabilities) involves several decision criteria
including Expected Monetary Value (EMV), Expected Opportunity Loss (EOL),
Expected Value of Perfect Information (EVPI).

 Decision trees represent an alternative approach to payoff tables; which are used for
problems that involve a series of chronological decisions by portraying sequential
decisions graphically.

 The circles ( ) and squares (  ) in the above figure are referred to as nodes. The
squares are decision nodes, and the branches emanating from a decision node reflect
the alternative decisions possible at that point.
Activity
1. The owner of the Burger Doodle Restaurant is considering two ways to expand operations:
opening a drive-up window or serving breakfast. The increase in profits resulting from these
proposed expansions depends on whether a competitor opens a franchise down the street. The
possible Profits from each expansion in operations given both future competitive situations are
shown in the following payoff table.

Competitor
Decision open Not Open
Drive-up window $-6,000 $20,000
Breakfast 4,000 8,000
Select the best decision using the following decision criteria.
a) Maximax
b) Maximin
c) Equally likely
d) Minimax regret
2. Consider the following payoff table for three alternatives, A,B, and C, under two future states
of the economy, good and bad.

Economic Conditions
Investment Good Bad
A $ 70,000 $ 25,000
B 120,000 -60,000
C 40, 000 40,000
Determine the decision using the following decision criteria.
a) Maximax
b) Maximin
 c) Minimax regret
d) Hurwicz (α = 0.3)
e) Equal likelihood

3. An investor is considering investing in stock, real estate, or bonds under uncertain economic
conditions. The payoff table of returns for the investor’s decision situation is shown below
Economic Conditions
Investment Good Stable Poor
Stocks $ 5,000 $ 7,000 $ 3,000
Real estate -2,000 10,000 6,000
Bonds 4,000 4,000 4,000
Determine the best investment using the following decision criteria.
a) Equal likelihood
b) Maximin
c) Maximax
d) Hurwicz ( α = 0.3)
e) Minimax regret