00 mi piace00 non mi piace

5 visualizzazioni30 paginedecision theory

Oct 08, 2018

© © All Rights Reserved

PPTX, PDF, TXT o leggi online da Scribd

decision theory

© All Rights Reserved

5 visualizzazioni

00 mi piace00 non mi piace

decision theory

© All Rights Reserved

Sei sulla pagina 1di 30

Decision Theory

certainty cannot be assumed.

among several different courses of action in an

attempt to optimize his decision process.

Steps in Decision Making

2. List all the possible alternative events that may

occur.

3. Determine all possible outcomes

4. Identify the profit/loss of each combination of

alternative events and outcomes.

5. Choose one mathematical decision theory model.

6. Apply the model and make the decision.

Decision Making Criteria

Maximax Criterion

Maximin Criterion

Mimimax Regret Criterion

Maximax Criterion

(Optimistic Approach)

payoffs for the different decisions starting with the

identification of the maximum payoffs of each

alternative decision.

Maximin Criterion

(Conservative Approach)

minimum payoffs given the various decisions that

are possible. It is a two step process once the payoff

table has been formulated.

The first step is to identify the minimum payoff for

each decision.

The second step is to select the largest minimum

payoff.

Minimax Regret Criterion

construction of an opportunity loss or regret matrix

prior to applying the minimax rule. Construct an

opportunity loss table by transforming each element

in the payoff table to an opportunity loss. The degree

of the oportunity loss for a given element is the loss

incurred by not selecting the optimal alternative

decision given a state of nature.

Converting a payoff table to an opportunity loss table is

composed of two step iterative process.

column. The element represents the best decision given

a particular state of nature. The second step is to

subtract each element in the column from the largest

element to compute the opportunity loss.

maximum regret (opportunity loss) for each desion and

the choose that decision with the smallest maximum

regret

Example 1:

plant next year on his land: corn, wheat, or rice.

The return from each crop will be determined by

whether a new trade bill with Hongkong passes the

Senate. The profit the farmer will realize from each

crop given the two possible results on the trade bill

is shown in the following payoff table. Determine

the best crop to plant using the following decision

criteria.

Crop Trade Bill

Pass fail

Corn 2,000,000 700,000

1,400,000 1,200,00

Rice

A. (Maximax Criterion)

Pass fail

Corn 2,000,000 700,000

1,400,000 1,200,00

Rice

maximum payoff

Step 2: identify the highest maximum payoff. The

minimum payoff can be found.

Corn 2,000,000

Rice 1,400,000

payoff of 2,000,000

B. Maximin Criterion

Step 1: Select the column with minimum payoff.

Pass fail

Corn 2,000,000 700,000

1,400,000 1,200,00

Rice

Minimum payoff

Step 2: Identify the lowest minimum payoff.

Pass fail

Corn 2,000,000 700,000

Rice

C. Minimax Regret Criterion

column.

Pass fail

Corn 2,000,000 700,000

1,400,000 1,200,00

Rice

Step 2: Subtract every element from the

largest payoff in both columns

Pass

Corn = 2,000,000 – 2,000,000 = 0

Wheat = 2,000,000 – 1,100,000= 900,000

Rice= 2,000,000 -1,400,000 = 600,000

Fail

Corn= 1,200,000 – 700,000 = 500,000

Wheat= 1,200,000 – 1,200,000 = 0

Step 3: select the largest amount in each

row.

Crop Trade Bill Crop Minimax

Payoffs

Pass fail

Corn 0 500,000 Corn 500,000

600,000 0

Rice 600,000

Rice

Step 4: Identify the maximum regrets

Crop Minimax Payoffs

Corn 500,000

Wheat 900,000

minimum

Rice 600,000

in regret

Expected Value/Mathematical Expectations

Mathematical Expectation, of a discrete random

variable is very important ion characterizing its

probability distribution. Originally, the concept of

expected value was introduced with reference to

games of chance where, if a player stands to win an

amount k with probability P, his expected value or

expectation is defined as the k(P).

The expected value of a discrete random variable of

probability distribution is the theoretical average of

the variable. The formula is

the symbol E(x) is used to expected value

Example 1

prize of 500,000 or a second prize of 200,000 with

probabilities 0.0001 and 0.0005. what should be a

fair price to pay for the ticket?

solution:

E(x)= ∑[ x ● P(x)]

E(x)= (500,000)(0.0001) + (200,000)(0.0005)

E(x)= 50 + 100

E(x) = 150

The fair price of the ticket is 150

Example 2

gambler will be paid php 9.00; if 2 or a 4 turns up he

will lose php 9.00; and if 6 is showing, he will win

12.00. Determine the expectation of the gambler.

Take note that there are six faces of a regular die

which serves as the possible outcome, therefore the

probabilities of A, B, and C are as follows.

E(x)= ∑[ x ● P(x)]

E(x) = [9 ● P(A)] – [9 ● P(B)] + [ 12 ● P(C)]

E(x) = 9 (3/6) – 9 (2/6) + 12(1/6)

E(x) = 4.5 – 3+2

E(x) = 3.50

Decision trees

branches. Though, rather that determining the

probability of each branch (outcome) as a probability

tree in a decision tree, the user computes the

expected value of each outcome and makes a

decision based on these expected values.

The circles and squares are referred as nodes. The

squares are decision nodes, and the braches

emanating from a decision node reflect the

alternative decisions possible at that point.

The decision tree represents the sequence of events

in a decision situation.

Decision variable- is a variable whose value

represents a potential decision on the part of the

decision maker.

Posterior probabilities

statistical dependence forms the necessary

foundation for an area of probability known as

Beyesian Analysis

additional information can sometimes enable one

to alter the marginal probabilities of the occurrence

of an event. The altered probabilities are reffered to

as Posterior Probabilities.

Example 1.

acquired a shoe company and is contemplating the

future one of its major plants located in Marikina.

Three alternative decisions are being considered: (1)

expand the plant and produce lightweight, durable

shoes for possible sales to the department store, a

market with little foreign competition: (2) maintain the

status quo at the plant, continuing production of shoes

that are subject to heavy foreign competition. (3) sell

the plant now. I

If one of the first two alternatives is chosen, the

plant will still be sold at the end of the year. The

amount of profit that could be earned by selling the

plant in a year depends on foreign market

conditions, including the status of a trade embargo

bill in Congress. The following payoff table

describes this decision situation.

States of Nature

Decision Good foreign Competitive Bad foreign Competitive

Condition Condition

Expand 30,000,000 20,000,000

Maintain status quo 50,000,000 -7,500,000

Sell Now 15,000,000 15,000,000

probability of 0.75 that good foreign competitive

conditions will exist and a probability of 0.25 that

bad conditions will exist. Determine the best

decision using expected value and expected

opportunity loss.

information

D. Develop a decision tree for this decision situation, with

expected values at the probability nodes.

report on the political and market situation in the future.

The report will be either positive (P) or negative (N),

indicating either good (g) or bad (b) future foreign

competitive situation. The conditional probability of each

report outcome given each state of nature is as follows.

0.70

probability obtained in part d.

## Molto più che documenti.

Scopri tutto ciò che Scribd ha da offrire, inclusi libri e audiolibri dei maggiori editori.

Annulla in qualsiasi momento.