Prepared by Eng. Denisse Zeledón Torres.

I UNIT THEORY OF DECISION

1. 1 Presentation

As we take on more responsibilities, complexity increases.

of our decisions and the way they are made. During we are in a
decision-making process shifts from using instincts, which must be guided by a
rational thinking. During this course we will study the decision-making processes from
a rational perspective. Human beings have always been faced with making
decisions. It is part of our daily tasks.
In the business world, decision-making requires greater analysis as there is much at stake.
what is at stake.
The application of simple mathematics, because what we will develop is
that through the language of rational thought, we will solve various problems
finding optimal solutions to achieve maximum profits for the business or
minimum costs.
The Decision problem, motivated by the existence of certain states of ambiguity that
consists of true propositions (known or unknown), is as old as life itself.
We can affirm that all living beings, even the simplest ones, face problems of
decision. Thus, a unicellular organism assimilates particles from its environment, some nutritious and others
harmful to him. The biological composition of the organism and the physical and chemical laws determine what
particles will be assimilated and which will be rejected.

As the complexity of the living being increases, so does the complexity of its
decisions and the way they are made. Thus, we move from a guided decision-making process
instinctively, to decision-making processes that must be guided by a thought
rationality in human beings. The Decision Theory will therefore address the study of decision-making processes.
of decisions from a rational perspective.

1.1.2 CHARACTERISTICS AND PHASES OF THE DECISION-MAKING PROCESS

A decision-making process has the following main characteristics:

There are at least two possible ways to act, which we will call alternatives or actions, that are mutually exclusive.
among themselves, in such a way that acting according to one of them makes any of the remaining ones impossible.

Through a decision-making process, one alternative is chosen, which is the one that is carried out.
The choice of an alternative must be made in such a way that it fulfills a specific purpose.
The decision-making process consists of the following fundamental phases:

Prediction of the consequences of each action. This prediction must be based on the
experience is obtained through induction on a set of data. The collection of this
Data sets and their use fall within the field of Statistics.

Assessment of the consequences according to a scale of goodness or desirability. This

value scale will lead to a system of preferences.

1
Prepared by Eng. Denisse Zeledón Torres.

Choosing the alternative using an appropriate decision criterion. This point also leads to
associated with the problem of choosing the most appropriate criterion for our decision, issue
which is not always easy to resolve in a completely satisfactory way.

1.1.3 CLASSIFICATION OF DECISION PROCESSES

Decision-making processes are classified according to the level of knowledge one has about the
set of uncontrolled factors or variables by the decision-maker that can influence it
final result (this is what is known as environment or context).

Based on the above, we can state that decisions can be made under conditions of:

Certainty A single future P = 1

Facing
states
Risk Several possible futures whose probabilities are known
natural

Uncertainty Several possible futures, the probabilities are not known

Conflict Conflict Facing rational opponents

The environment is one of certainty when its state is known with certainty, that is, each action leads
invariably to a well-defined result.
The risk environment when every decision can lead to a series of consequences to which
a known probability distribution can be assigned.
The atmosphere is one of uncertainty when every decision can lead to a series of consequences.
those that cannot be assigned a probability distribution, either because it is unknown or because it is not
It makes sense to talk about her. Depending on the context, we will say that the decision-making process (or the taking of
decisions) are made under certainty, under risk, or under uncertainty, respectively.
The environment is one of conflict when competition takes actions to counteract the effect of
In our decisions, we face a rational opponent, this type of problem is analyzed.
through game theory which is a branch of statistics

1.1.4 ELEMENTS OF A DECISION PROBLEM

In every decision problem, a series of characteristic elements can be distinguished:
The decision-maker, responsible for choosing the best way to act according to their interests.
The alternatives or actions, which are the different possible ways of acting, among which one can choose
it will select one. They must be mutually exclusive.
The possible states of nature, a term used to refer to all those events
futures that escape the control of the decision-maker and influence the process.
The consequences or results obtained by selecting the different alternatives under each one
of the possible states of nature.
The decision rule or criterion, which is the specification of a procedure to identify the best
alternative in a decision problem.

1.2 DECISION TABLES

2
 Prepared by Eng. Denisse Zeledón Torres.

Many decision-making processes can be handled through decision tables, in the

that represent the characteristic elements of these problems:
The different states that nature can present:1,e2.
The actions or alternatives from which the decision maker will select:1,a2I am.
The consequences or results xijfrom the choice of the alternativeiwhen nature
present the statej.
The odds:
It is assumed, for simplicity, the existence of a finite number of states and alternatives. The format
Generally, a decision table is as follows:

General form of a decision table

States of Nature

P1 P2 …. Pn

e1 e2 ... en

a1 x11 x12 ... x1n

a2 x21 x22 ... x2n

Alternatives

... ... ... ... ...

am xm1 xm2 ... xmn

Prototype example (decision table) No. 1

The Gofer Broker Company owns land where there may be oil. A geologist
The consultant has informed the administration that they believe there is a 1 in 4 chance of finding.
oil.
Due to this possibility, another company has offered to buy the land for $90,000. However, the
Goferbroke considers keeping it to drill it herself. The cost of the drilling is 100 thousand.
dollars. If they find oil, the expected income will be 800 thousand dollars; thus the expected profit
for the company (after deducting the cost of drilling) will be 700 thousand dollars. Costs will be incurred in
a loss of 100 thousand dollars (the cost of drilling) if oil is not found.
Build the decision table
States of Nature
Alternatives Oil Dry
To perforate searching
oil $ 700,000 $100,000

Sell the land $90,000 $90,000

Possibilities del
state 0.25 0.75

3
Prepared by Eng. Denisse Zeledón Torres.

1.3 Decision criterion under risk conditions

Decision-making processes in a risk environment are characterized because they can be associated with
probability of occurrence for each state of nature, probabilities that are known or
They can be estimated by the decision-maker before the decision-making process.

1.3.1 Expected Monetary Value Criterion

A good criterion for deciding between several alternatives under risk conditions must reflect all the
possible consequences of each act and the relative probability of those consequences.
The criterion of expected monetary value (EMV) meets these conditions and is also easy
application.
The expected monetary value criterion states that a rational decision will lean towards the alternative.
that provides the highest expected profit.
VME(X) = XiP(Xi) = X1P(X1) + X2P(X2)+…..+XnP(Xn)

Example No.2
The following payoff matrix (in thousands of dollars) is provided for a decision analysis.

States of Nature

Alternatives E1 E2 E3

A1 4 0 0

A2 0 2 0

A3 3 0 1

Probabilities 0.2 0.5 0.3

1.3.2 Applying the VME criterion

1.- Determine the VME of the different possible monetary results Xij of each alternative Ai
2.- Select the option that has the best expected monetary value (the maximum in a matrix of
profit and the minimum in a cost matrix.
4 (0.2) + 0 (0.5) + 0 (0.3) = 0.8
0 (0.2) + 2 (0.5) + 0 (0.3) = 1.0
VME (A3) = 3 (0.2) + 0 (0.5) + 1 (0.3) = 0.9

The expected profit of option A2 is greater than that of options A1 and A3, so it should
choose option A2.

Example No. 3

4
Prepared by Eng. Denisse Zeledón Torres.

You are the owner of a baseball goods store and you must decide how many gloves to order.
for the summer league season.
For a particular type of gloves, you must order in batches of 100 gloves.
If you order 100 gloves, the cost is $10 per unit.
If you order 200, the cost is $9 per unit and
If you order 300 or more, the cost is $8.50.
The selling price of each glove is $12, but if some are left unsold at the end of the
season, these must be sold at half price.
For simplicity, do you think the demand for this type of gloves is 100, 150, or 200 units?
which have been assigned based on experience from previous years the probabilities of 0.3, 0.4, and 0.3
respectively.
It is clear that you cannot sell more than you store. However, if you fall short in the
For units acquired, there is an estimated loss of good name of $0.50 for each glove that a
the person wants to buy, but cannot do so because it is out of stock.
In addition, you must place the order now, for the upcoming summer season and cannot wait until
observe how the demand for this type of item varies before ordering, I couldn't place several orders.
It is requested:

Develop a results matrix and select the alternative that provides the highest expected gain.
using the VME criterion.

States of nature
E1 E2 E3
Alternatives D = 100 D= 150 D= 200

A1 (S= 100) 200 175 150

A2 (S= 200) 0 300 600

A3 ( S= 300 ) 150 150 450

VME (A1) = 200(0.3) + 175 (0.4) + 150(0.3) = $ 175

$ 300
$150

Since Action A2 is the one that provides us with the highest VME at $300, it will be the selected alternative, that is
request 200 gloves.

1.4 DECISION-MAKING PROBLEM UNDER UNCERTAINTY

In decision processes under uncertainty, the decision-maker knows what the possible states are.
nature, although it has no information about which of them will happen. It is not only incapable of
to predict the actual state that will occur, but also cannot quantify this in any way
uncertainty. In particular, this excludes knowledge of probabilistic information about
the probabilities of occurrence of each state.

Business problems of this type arise when there is no past experience to

determine the probabilities of occurrence for the various states of nature.

The selection of a specific criterion is determined by the size of the company, the objectives and the
policies of the same, the feelings of the person in charge of making the decision or another logical basis.

5
Prepared by Eng. Denisse Zeledón Torres.

1.4.1 DECISION RULES

.
The problems of uncertainty applied to business arise when there is no experience.
past to determine the probabilities of occurrence for the various states of nature.
This type of decision is more complex since we do not have probabilities for the different states of
nature.
Problems can be solved by selecting specific criteria, which are selected
taking into account the size of the company, its objectives and policies, as well as by the
feelings of the one who makes the decision.

[Link] MAXIMAX CRITERION

This criterion adopts an optimistic mindset, as it is assumed that nature will always react.
in our favor for what will always present the most favorable state.
The Maximum criterion chooses the alternative that provides the highest level of optimism possible, which is
directly related to the highest payment that can be obtained by choosing one of the alternatives.

[Link] MAXIMIN CRITERION

This criterion takes a stance opposite to maximax. The decision-maker is a pessimistic being, so ...
therefore suggests that the decision-maker should choose the alternative that provides the highest level of
possible security.
During the application of this criterion, reasoning is given about the worst that can happen to the decision-maker when
choose an alternative. Generally, a small business should adopt that approach.
conservative, since he has invested all his assets or most of them in one place, he must
to take care not to lose them. But the medium and large company to adopt this to a certain degree
method.

[Link] MINIMAX CRITERION

This criterion is carried out by applying opportunity cost analysis associated with consequences.
economicXijThe opportunity cost is developed for each state of nature as the
difference between the maximum gain value for Ej and the considered value Xij.

The opportunity cost (OC)

Every decision has a certain cost, which is the opportunity cost or quasi-cost. The fact that
Doing a certain thing implies an opportunity cost in relation to not having done something different.
When the Stock value corresponding to the selected alternative matches the value of the
demand, the CO is zero because it is the maximum profit that can be obtained, given that value of the
demand.
It is called opportunity cost:
Associated with each result Xij (the result corresponding to alternative Ai for future Ej), to the
difference between the maximum gain value for Ej and the considered value Xij.
CO (Ai, Ej.) = Maximum Xij-Xij (in profit matrix)

6
 Prepared by Eng. Denisse Zeledón Torres.

CO (Ai, Fj) = Xij - Minimum Xij (in cost matrix)

In other words, the opportunity cost is determined by what was foregone because of having...
chosen the AI alternative instead of having chosen the optimal alternative for the future E.g.

[Link] LAPLACE CRITERION

This criterion assumes that all the different states of nature have an equal possibility of
occurrence.
To find the best alternative, the expected amount for each payment is calculated and the best one is chosen.
strategy that has the highest expected payoff (if the matrix is of utility). If the matrix is of costs, it
will choose the quantity with the lowest value found.
The absence of knowledge about the state of nature amounts to claiming that all states
they are equally probable. Thus, for a decision problem with possible states of nature,
we will assign probability 1/na to each of them.

The Laplace rule selects as the optimal alternative the one that provides the highest result.
expected

1.2 TABLAS DE DECISIÓN

Many decision-making processes can be handled through decision tables, in the

that represent the characteristic elements of these problems:
The different states that nature can present:1,e2,...,en.
The actions or alternatives from which the decision-maker will select:1,a2I am.
The consequences or results xijfrom the choice of the alternativeIwhen nature
present the statej.
The odds:
It is assumed, for simplicity, the existence of a finite number of states and alternatives. The format
Generally, a decision table is the following:

General form of a decision table

States of Nature

P1 P2 …. Pn

Alternatives e1 e2 ... en

7
Prepared by Eng. Denisse Zeledón Torres.

a1 x11 x12 ... x1n

a2 x21 x22 ... x2n

... ... ... ... ...

am xm1 xm2 ... xmn

EXAMPLE # 1
In a certain city, an airport is going to be built in one of two possible locations A and B, which will be
chosen next year. A hotel chain is interested in opening a hotel near the new
airport, for which it has to decide which lands to buy. The following table shows the price of
the land, the estimated benefit that the hotel will obtain in each possible location if the airport is
locate there, and the sale value of each plot of land if the airport is not ultimately built in that location
(the amounts are expressed in pesetas x 107What is the most appropriate decision?

Parcel in A Plot in B

Price of the land 18 12

Estimated benefit of the hotel 31 23
Sale value of the land 6 4

The possible alternatives available to the decision-maker are as follows:

Buy the plot in A.
Buy the plot in B.
Buy both plots.
Do not buy any plot.
On the other hand, the possible states of nature are:
The airport is being built in A.
The airport is being built in B.
Thus, if the hotel chain buys the land in A and the airport is finally built there, it will obtain
as final performance corresponding to the operation of the hotel, 31, minus the investment made
in the purchase of the land, 18, that is, 31-18 = 13. On the contrary, if the airport is built in B, the
The land acquired in A must be sold, which will yield a profit of 6, to which will have to
Subtract the initial investment in the purchase, 18. This results in a final return of 6-18 = -12.
In a similar way, the results of the remaining alternatives are determined for each of the
possible states of nature, leading to the following decision table:

8
Prepared by Eng. Denisse Zeledón Torres.

States of Nature

Alternatives
Airport in A Airport in B
Land purchased

A 13 12

B 8 11

A and B 5 1

None 0 0

Starting fromthe example of hotel construction,the following table shows the level of optimism of
maximax criterion

States of Nature Criterion

Alternatives
Airport in A Airport in B
Land purchased Maximax

A 13 12 1 Mayor

B 8 11 11

A and B 5 1 5

None 0 0 0

Alternative Buy
selected land in A
The optimal alternative according to the maximax criterion would be to buy the plot in location A, because
provide the highest level of optimism.
Critique. When using the maximax criterion, losses can be high if the state of the
suitable nature.

Starting fromthe example of hotel construction,the following table shows the level of pessimism of
maximin criterion.

States of Nature Criterion

Alternatives
Airport in A Airport in B
Land purchased Maximin

9
Prepared by Eng. Denisse Zeledón Torres.

A 13 12 -12

B 8 11 -8

A and B 5 1 -1

None 0 0 0

Alternative Buy none

Selected land

Starting from the hotel construction example. We will apply the Minimax criterion (CO)
Opportunity Cost Table

States of Nature Minimax

Alternatives Airport in
Airport in A
Land purchased B
CO
A 0 23 23

B 21 0 21

A and B 8 12
1
None 13 11 13

Alternative
selected Will buy at A and B

Applying the opportunity cost calculation by placing the highest value in the Airport column in A.
It is 13; by subtracting each of the values in that column from this amount, the losses are obtained.
relative to the state of nature airport in A. In the same way, the maximum of the column
Airport in B is 11; subtracting from this amount each of the values in that column yields the
elements corresponding to the state of nature at Airport in B.
Subsequently, we select the highest of the opportunity costs for each alternative and finally
We select the smallest of all opportunity costs. The smallest is obtained for the third.
alternative, so the optimal decision according to the Minimax criterion would be to buy both plots.
Starting fromexample of hotel construction,the following table shows the expected results
for each of the alternatives.

States of Nature VME

Alternatives Airport in A Airport in B Result
Land purchased P=½ P=½ expected

10
Prepared by Eng. Denisse Zeledón Torres.

A 13 -12 0.5

B -8 11 1.5

A and B 5 -1 2

None 0 0 0
Buy
Alternative
land in
selected
A and B
In this case, each state of nature would have a probability of occurrence of 1/2. The expected result
the maximum is obtained for the third alternative, so the optimal decision according to Laplace's criterion.
it would be to buy both plots.

Self-assessment activities.

A company plans to introduce a revolutionary product with a completely new packaging.

new to replace another already existing product at a higher price (A1) or a moderate change in
the ingredients of the existing product with new packaging and a slight price increase (A2)
or a small change in the ingredients of the existing product and a single change in the packaging
which consists of the word "new" with a minimal price increase (A3)
The three possible states of nature are:
Expansion of demand (N1)
No demand expansion (N2)
Demand contraction (N3)

The marketing department has calculated the expected payments in terms of annual profit.
States of nature
Strategies N1 N2 N3

A1 500,000 100,000 -50.000

A2 250,000 300,000 0

A3 100,000 100,000 100,000

When applying the different decision criteria to the example

.3 Decision criteria under risk conditions

Decision-making processes in a risky environment are characterized because they can be associated with
probability of occurrence for each state of nature, probabilities that are known or
They can be estimated by the decision-maker before the decision-making process.

1.3.1 Expected Monetary Value Criterion

A good criterion for deciding among several alternatives under conditions of risk must reflect all the
possible consequences of each act and the relative probability of those consequences.
The expected monetary value (EMV) criterion meets these conditions and is also easy to use.
application.

11
Prepared by Eng. Denisse Zeledón Torres.

The expected monetary value criterion states that a rational decision will lean towards the alternative
that provides the highest expected profit.
VME(X) = XiP(Xi) = X1P(X1) + X2P(X2)+…..+XnP(X)

Example No.2
The following payment matrix (in thousands of dollars) is provided for decision analysis.

States of nature

Alternatives E1 E2 E3

A1 4 0 0

A2 0 2 0

A3 3 0 1

Probabilities 0.2 0.5 0.3

1.3.2 Applying the VME criterion

1.- Determine the VME of the different possible monetary outcomes Xij of each alternative Ai
2.- Select the alternative that has the best expected monetary value (the maximum in a matrix of
profit and the minimum in a cost matrix.
4 (0.2) + 0 (0.5) + 0 (0.3) = 0.8
0 (0.2) + 2 (0.5) + 0 (0.3) = 1.0
VME (A3) = 3 (0.2) + 0 (0.5) + 1 (0.3) = 0.9

The expected profit of alternative A2 is greater than that of alternatives A1 and A3, so it should
choose option A2.

Example No. 3
You are the owner of a baseball goods store and you must decide how many gloves to order.
for the summer league season.
For a particular type of gloves, you must order in batches of 100 gloves.
If you order 100 gloves, the cost is $10 per unit.
If you order 200, the cost is $9 per unit and
If you order 300 or more, the cost is $8.50.
The selling price of each glove is $12, but if some remain unsold at the end of the
season, these must be sold at half price.
For simplicity, do you think the demand for this type of gloves is 100, 150, or 200 units?
which has assigned based on experience from previous years the probabilities of 0.3, 0.4, and 0.3
respectively.
It is clear that you cannot sell more than you store. However, if you fall short in the
For acquired units, there is an estimated loss of good reputation of $0.50 for each glove that one
a person wants to buy, but cannot do so because it is out of stock.
In addition, you must place the order now for the upcoming summer season and cannot wait until
observe how the demand for this type of item varies before ordering, I couldn't place several orders.
It is requested:

12
Prepared by Ing. Denisse Zeledón Torres.

Develop a results matrix and select the alternative that provides the highest expected profit.
using the VME criterion.

States of nature
E1 E2 E3
Alternatives D= 100 D = 150 D= 200

A1 (S= 100) 200 175 150

A2 (S= 200) 0 300 600

A3 ( S= 300 ) 150 150 450

$ 175
$ 300
VME (A1) = 150 (0.3) + 150 (0.4) + 450(0.3) = $ 150

Since action A2 is the one that provides us with the highest NPV at $300, it will be the selected alternative.
ask for 200 gloves.

DEMAND STOCK ANALYSIS.

It consists of the level of stock to be maintained for an item of random demand when for
any reason can only place a single order or request, meaning that in the case of
that if there are missing items, it cannot be reordered or requested again

STOCK TABLE

SD S=D SD
Revenue S S D

- cost CVV(S) CVV(S) CVV(S)

Other costs CVV(S) CVV(S) CVV(S)
Loss of prestige -(D-S) ------- ------
Value of ------ ------- (S-D)
recovery.
Utility S-CVV(S)-(D-S) S-CVV(S) D-CVV(S)+(S-D)

1.4 Expected profit with complete information:

It is the greatest benefit that would be obtained from the future that is going to happen.

GEIC=∑ MAYOR XIJ x pij (PROFIT TABLE)

13
Prepared by Eng. Denisse Zeledón Torres.

1.5- The Expected Value of Perfect Information (EVPI)

It refers to knowing the largest fee we are willing to pay to obtain the information.
perfect. In general terms.
VEIP = (Expected return with perfect information) - (Expected return in
Normal conditions)
VEIP = GEIC- VME (PROFIT TABLE)

VEIP: It is the upper limit of the amount one would be willing to pay to improve knowledge.
of the state of nature that will occur. Millions of dollars are literally spent on various
market research projects and other verification resources (geological tests,
quality control experiments, etc.) to determine which state of nature will occur in a
wide variety of applications. The expected value of perfect information indicates the anticipated amount
what will be gained in any similar effort, with what imposes a higher quota than the
amount that can be spent on collecting information.

CALCULATION FOR A COST TABLE.

COST OF COMPLETE INFORMATION = ∑ MINOR XIJ x PIJ

CIC==∑ LESS XIJ x PIJ

RISK COST = VME - CIC

CR=VME-CIC

Example No. 6
The daily demand for loaves of bread in a grocery store can be one of the values.
next: 100, 120 or 130 loaves. The store owner limits his options to providing one of
the three indicated levels. He always sells everything he has, if they have more in stock than what
You can sell on the same day, you must get rid of the leftover loaves of bread at a price of
discount of 55 cents/loaves, but has less stock than the demand, this will result in a loss
of prestige that I value at 15 cents for each unsold unit. Assuming that the merchant
pays 60 cents for a loaf of bread and sells it for $1.05, determine the supply level.
optimized through the use of each of the indicated decision models.
A.- Suppose that the following probabilities for the states of nature are specified:
P (100) = 0.2, P (120) = 0.5, P (130) = 0.3
1.- Prepare a compensation table
2.- What is the optimal decision if the maximum criterion is used?
3.- What is the optimal decision if the maximin criterion is used?
4.- What is the optimal decision if the Laplace criterion is used?
5.- Generate a table of damages
6.- What is the optimal decision if the minimax regret criterion is used?
B.- Suppose you are in a decision with no knowledge of the states of nature.
1.- Find the decision that maximizes the expected net return in córdobas.
2.- Find the decision that minimizes expected loss.
C.- Determine how much you would be willing to pay to have knowledge about the demand that will occur.

14
Prepared by Eng. Denisse Zeledón Torres.

TABLE OF REMUNERATIONS
Demand
Offer 100 120 130
100 45 42 40.5
120 44 54 52.5
130 43.5 53.5 58.5

A2: Maximax decision criterion (Maximize the maximum returns per decision)

Select the highest Performance and from these select the highest for each alternative or offer.
Demand
Offer 100 120 130 Criterion
Maximax
100 45 42 40.5 45
Level of
120 44 54 52.5 54
provisioning
130 43.5 53.5 58.5 58.5 optimal

A3: Maximin decision criterion (maximize minimum returns)

Select the least performance and from this select the highest for each alternative or offer

Demand
Offer 100 120 130 Criterion
Minimax
100 45 42 40.5 40.5
Level of
120 44 54 52.5 44
supply
130 43.5 53.5 58.5 43.5 optimal

A4: Place criterion

This criterion assigns equal probability weight to each state of nature. Since there are three states
from nature then P = 1/N = 1/3
1/3
This Criterion selects the highest VME

Demand
Offer 100 120 130 Criterion
P = 1/3 P = 1/3 P = 1/3 Laplace
(VME)
100 45 42 40.5 42.5

120 44 54 52.5 50.1667

Level of
130 43.5 53.5 58.5 51.833 supply
óptimo

VME (100) = 1/3 (45) + 1/3 (42) + 1/3 (40.5) = $ 42.5

15
Prepared by Eng. Denisse Zeledón Torres.

VME (120) = 1/3 (44) + 1/3 (54) + 1/3 (52.5) = $ 50.1667

VME (130) = 1/3(43.5) + 1/3 (53.5) + 1/3 (58.5) = $ 51.8333

A5: Table of damages (Opportunity cost table)

CO (Ai,Ej) = Maximum Xij-Xij (in the profit matrix): Select the highest in the column and it
subtract the current element.

OPPORTUNITY COST TABLE

Demand
Offer 100 120 130

100 0 12 18

120 1 0 6

130 1.5 0.5 0

A6: Minimax decision criterion

Minimize the maximum losses, by decision. Select the maximum value for each alternative and
then choose the least harm.

OPPORTUNITY COST TABLE

Demand
Offer 100 120 130 Criterion
minimax
100 0 12 18 18

120 1 0 6 6
Level of
130 1.5 0.5 0 1.5 supply
óptimo

B1 Maximize expected monetary value under risk

VME (X) = ∑Xi P (Xi) from the payment table

Si P (100) = 0.2, P (120) = 0.5, P (130) = 0.3

VME (100) = 45(0.2) +42(0.5) + 40.5 (0.3) = $ 42.15

VME (120) = 44 (0.2) + 54(0.59) + 52.5 (0.3) = $ 51.55
VME (130) = 43.5 (0.2) + 53.5 (0.5) + 58.5 (0.3) = $ 53 The greatest VME
Decision: The optimal supply level is to offer 130 loaves of bread.

B2Offer Minimize expected loss Low risk

PE(X) = ∑ XiP(X) from the opportunity cost table

100 = 0(0.2) + 12(0.5) + 18(0.3) = $11.4

PE (120) = 1 (0.2) + 0 (0.5) + 6 (0.3) = $ 2
PE (130) = 1.5 (0.2) 0.5 (0.5) + 0 (0.3) = $ 0.55
Decision: The optimal supply level is to provide 130 loaves of bread

16
Prepared by Eng. Denisse Zeledón Torres.

Note: It is always the case that VME(X) = PE(X) in decision.

Value of Perfect Information

Amount one is willing to pay to receive information.

VEIP = VENT–VME
Where VEIP: EXPECTED VALUE OF PERFECT INFORMATION
VENT: EXPECTED VALUE WITH THE NEW DEAL
VME: EXPECTED VALUE WITH THE ORDINARY SEQUENCE (Found in B1)
The highest result is selected from the Compensation Table for each state of nature.
VEIP = 45 (0.2) + 54 (0.5) +- 58.5 (0.3) = $ 53.55
$53

VEIP = 53.55–53 = $0.55

He is willing to pay $0.55 to receive information.
It is always the case that VEIP(X) = PE(X)

1.6.-Self-Assessment Activities No.1

1.- The finance department of an organization has calculated the profits from the strategy of
Launch of a new product to the market as shown in the following table α = 0.6
The amounts are given in thousands.

State of Nature
Strategy E1 E2 E3
A1 400 900 -500
A2 -200 700 0
A3 -100 200 300

Determine the alternative that should be selected for each of the following criteria.
Maximax
b) Maxi min
c) Minimax
d) Laplace
A shoe store sells one particular style for $15 per pair. If it remains unsold at the end
the product will be sold at half price during the season. In the event that the opposite occurs, it is
To say how many are missing, a loss of $1 will be assumed for each unsold pair. The problem that
The warehouse owner faces the question of how much to request from his supplier. The supplier presented him with the...
next quote.
If you order 200 pairs The cost is $10 per pair.
If you order 300 pairs, the cost is $7 per pair.
If you order 400 pairs, the cost is $6/pair.
According to the experience of the store owner, he expects to sell 200, 300, and 400 pairs.
Build the result matrix and through Laplace's criterion determine how much should be.
to ask.
3.-Procesa is a food processing company that grows its own crops.
past experience of the company, with the planting of three types of crops in different areas of the country,

17
Prepared by Eng. Denisse Zeledón Torres.

the following profits have been obtained given in the following result matrix, for the three
good weather

States of
nature
Strategy P1 = 0,55 P2 = 0,35 P3 = 0,10
A1 50,000 65,000 15,000
A2 70,000 45,000 18,000
A3 45,000 25,000 14,000

a-Which strategy is better according to the VME criterion?

If the probabilities for the different states of nature are N1: 1/2; N2: 1/4; and N3: 1/4, what
It will be the best strategy according to the VME.
c- calculate in each the geic, veip.

4.-Condominio S.A. is planning the development of a set of condominiums near Granada.

The land that he intends to buy will cost $600,000. The development of the common area will cost another $400,000.
Units will cost $30,000 each and are expected to be sold for $80,000.
The problem is deciding how many units to build, considering that: High demand is obtained for
The units, 40 of them can be sold at full price. It is thought that the probability of a
High demand is 0.35. A medium demand is obtained; only 30 can be sold at full price.
There will be a DUMPING of the remaining units with a loss of $5000 per unit.
The probability of an average demand is 0.55.
The demand turns out to be small, only 20 could be sold at full price and the rest will have to
to sell each at a loss of $5000.
The question is. Should Condominios S.A build 0, 20, 30 or 40 units?
5.- Using the results from the utility matrix shown in the following table.

D= 100 D= 150 D= 200

A1 200 175 150
A2 0 300 600
A3 -150 150 450
a.- Find the best alternative according to the Maxi min criterion.
b.- Find the best alternative according to the Maximax criterion.
c.- Find the best alternative according to the Minimax criterion
d.- Find the best alternative according to the Laplace criterion

6.- You represent a major importing company for slippers. Your supplier presents you with the
costs of a particular style. The order must be placed today.

If you order 100,000 units the cost is $3 per unit

If you order 200,000 units, the cost is 1.5 per unit.
If you order 300,000 units, the cost is 1.7 per unit.
The selling price is $5 if it remains unsold by the end of summer, it must be sold for $4. You did
an analysis and believes that the demand for slippers will be 100,000, 150,000, and 200,000 units.

a-Build the matrix of results or economic consequences and through the Maximax criterion that
quantity you would request in your order.

18
Prepared by Eng. Denisse Zeledón Torres.

What would be the best decision according to the VME.

c- what would be the geic and the veip.

1.8.- Decision Trees

It is a graphical resource for analyzing decisions under risk, that is, problems in which there have been
specified the probabilities of the states of nature. They were created to be used in problems in
those that have a sequence of decisions, each of which leads to one among several
uncertain results.
Decision tree-based analysis is a useful tool in decision-making regarding
investment is the acquisition or sale of physical property, the management of projects, people and
new product strategies.
Often decision-making problems involve making a sequence of decisions, that is,
there are occasions when the entire set of decisions at different points in time must
to be evaluated simultaneously with the initial decision making.

1.8.1 Components and structure

Decision trees have:
Branches: are represented by lines
Decision nodes: from them, the decision branches come out and they are represented with
Nodes of uncertainty: from them the branches of events emerge and are represented with

A decision tree has the following structure:

1st Decision Point 2nd decision point

X2

E2/P2
B X3
D1
E3/P3 X4 D4 X5 E6/P6
D2 X7
A
C D5
D F
E4/P4
E7/P7
X8
E5/P5
D3 D6 X6
X1
E
E8/P8 X9
D7 G
E9/P9 X10

1.8.2 Analysis: Monetary Value Criterion The construction of the tree is done from left to right and
The analysis begins from the far right of the tree and moves through the probability nodes.
and decision nodes until reaching the 1st decision point. Calculating each payment at the end of the branches.
The expected value is also calculated at each decision node. Then at each decision point,
select the alternative with the optimal expected value

19
Prepared by Eng. Denisse Zeledón Torres.

The optimal expected value is the value that provides us with the greatest profit or the lowest cost.

At each decision node, the expected value is calculated.

At each node or decision point, the alternative with the optimal expected value is selected.
This form of backtracking in the branches to the first decision point, optimizing as
we march forward only with the best option, while we eliminate all the
inferior, it is known as the process of backward induction.

Example # 1 For the structured decision tree above and giving monetary results for
for each event and the respective probabilities we have:

A decision tree has the following structure:

$50,000
P1=0.50

0.50
B -$10,000
D1 $38,000
P3=0.10 $10,000 D4
D2 P6= 0.30 $ 40,000
A C D D5
P4= 0.4 F
$30,000
P5=0.5 P7= 0.70
D3 D6 $1,5000
$0
E
P8= 0.50 $50,000
D7 G
P9= 0.50 $ 20,000

$33,000
VME (G) = 0.50 (50,000) + 0.50 (20,000) = $35,000

Second decision point

We compare VME (F) $33,000 with the economic consequence in D4. $38,000. Since D4 is greater, the
the decision is D4.
We compare VME (G) $35,000 with the economic consequence in D6 $15,000. Since VME (G) is greater.
the decision is D7

20
Prepared by Eng. Denisse Zeledón Torres.

First decision point

VME (B) = 0.50 (50,000) + 0.50 (-10,000) = $20,000
$33,700

We compare VME (B) $20,000, VME (C) $33,700 and the economic consequence in D3 $0.
Since VME (C) is the highest, then D2 is the optimal alternative.

Example # 2
Let's assume that our company has to decide whether to continue the regional distribution of a product.
or expands it to a national distribution. This represents a decision point for the company.
The casual events that can affect the national or regional distribution decision consist of
to know if there will be a high national demand for the product, a medium national demand, or a
limited.
If there is a high demand, they could expect profits of 4 million dollars, while they could
expect profits of 2 million dollars or 0.5 million with medium or limited demand
respectively.
For a regional distribution, the following profits can be forecasted:
If the regional demand is large, the company can obtain 2 million dollars. On the other hand, if the
regional demand is medium or limited, profits of 1.8 and 1.5 million dollars are calculated,
respectively.
The probabilities of occurrence of the three types of demand are 0.5 for high demand, 0.25 and
0.25 for a medium demand and a limited demand, respectively. What is the best?
alternative?

Solution:
We built the decision tree and calculated the NPV for each alternative.

$4 million

DA/0.50

DM/0.25
B $2M
National Distribution

DB/0.25
$ 0.5 M
A
DA/0.50 $2 million

Regional Distribution DM/0.25

C $1.8 million

DB/0.25
$1.5 million

VME (B) = 0.5 (4 ) + 0.25 ( 2 ) + 0.25 (0.5) = $ 2.625 M

VME (C) = 0.5 (2) + 0.25 (1.8) + 0.25 (1.5) = $ 1.825 M

The best alternative will be: A National distribution, expecting a profit of $2,625,000

21
Prepared by Eng. Denisse Zeledón Torres.

1.9. Self-Assessment Activities No.2

1.-Rick O'Shea is an independent trucker operating from Tucson. He has the option to take a...
shipment to Denver or take another different shipment to Salt Lake. If you select the
Embarkation to Denver has a 90% chance of finding a return shipment to Tucson.
If you do not find a return shipment, you will come back empty. If you choose the shipment to Salt Lake, you have a
50% chance of finding a return shipment to Tucson. Below are your
remunerations.
BOARDING NO RETURN
RETURN
Salt Lake $4,000 $3,500
Denver 3,850 3,350

a) Draw the decision tree for this problem

b) Using the expected return criterion in dollars, which city should Rick go to?
A company is trying to decide whether to build a new branch in a shopping mall.
open, shopping mall closed or in a remote place. All with approximately the same
initial investment. The values of profit and probabilities are summarized in the following tables

Earnings table:
SALES
For under of the Average Above average
average
Outdoor center 100,000 200,000 400,000
Closed Center 200,000 400,000 600.00

Secluded place 50,000 100,000 300,000

Probability Table SALES
For below of the Average Above average
average

Outdoor center 0.15 0.60 0.25

Closed Center 0.35 0.50 0.15
Secluded place 0.30 0.40 0.30
a) Build the Decision Tree
b) Calculate the expected gain of each node
c) Identify the optimal decision

3.- Maple is planning the introduction of a new product. The cost to be in a position to
factory one of the components of the product is very high so Maple is considering buying
this component instead of manufacturing it. However, once prepared for manufacturing the
component, the unit variable cost of Maple would be low compared to the purchase price of the
component. In the table, the materials manager has calculated the net profit in thousands of dollars for
three levels of demand

DEMAND
DECISION LOW MEDIUM HIGH
PROBABILITY 0.4 0.3 0.3

22
Prepared by Eng. Denisse Zeledón Torres.

MAKE THE COMPONENT 11 32 53

BUY THE COMPONENT 15 30 45

Identify the optimal decision

Coloco currently has an asset of $150,000 and wants to decide whether to sell it or not.
chocolate-flavored soda, it has 2 options:

Option I: conduct the market test or not. If you conduct the market test, you must
consider local success or local failure with probabilities of 0.6 and 0.4. If it is local success
You should consider if it sells nationally or does not sell nationally. If it sells at
At the national level, it will have a profit of $180,000, and if it does not sell, $120,000. If it happens
The local failure and not selling will have the same profits. But with a probability of 0.85
and 0.15 if it sells or not.

Option II: not conducting the market test and considering success or failure with a
profit of $100,000 and $150,000 with equal probability. Indicate the sequence of
decision.

As a byproduct of another research, the company CREMOSA found a substance that can
to be used as a tanning cream.
A major company in the skincare industry has offered to buy the rights to
the cream for $20,000 and they would later develop the product commercially.
CREMOSA is considering developing the product on its own. This effort is estimated to cost
$10,000 would have half the chances of being a success.
If the product were successfully developed, several companies would try to buy the rights. CREMOSA
think that there is a 0.4 chance of receiving $80,000 and a 0.60 chance of receiving $45,000 for the
rights.
Another option after developing the product would be for CREMOSA itself to market it. It is considered
that the possible yields of this alternative are $10,000, $50,000, and $150,000 with probabilities
respective of 0.3, 0.5 and 0.2
If CREMOSA fails in its attempt to develop the product, think that it could still be sold.
rights for $50,000.
a.- Build a decision tree for that situation
b.- Develop an optimal action plan for CREMOSA.

The owner of a property with 32 apartments is considering the possibility of changing the
heating system using fuel oil through an electric system. The owner of the
the property decides to review its costs and its income to decide. The information is the
next:

Before installation: utility: $72000

After installation: utility: $55000

23
Prepared by Eng. Denisse Zeledón Torres.

To rent the other 8 homes, the owner has two means: a: to present
advertisements in the newspapers, which will cost $500

b: put a rental advertisement at the main entrance. The possibilities of renting

housing

Advertising medium Probability of leasing Increase in income

A 70% $25000

20% $12500

10% $0

B 50% $25000

30% $12500

20% 0

a- Build a decision tree.

24