Scribd
Upload
12 views12 pages

Decision Theory

Uploaded by

Christe Maesse Gesto
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
12 views12 pages

Decision Theory

Uploaded by

Christe Maesse Gesto
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 12

DECISION THEORY

BUSINESS STATISTICS
Introduction to Decision Theory
• Decision Theory is a field of study that
provides a systematic framework for making
decisions in the face of uncertainty. In
business statistics, it helps decision-makers
choose the best course of action by
considering various factors and potential
outcomes.
Elements of Decision Making:
• Decision Makers: Individuals or entities
responsible for making decisions.
• Alternatives and Choices: Different options or
courses of action available.
• Outcomes and Consequences: Results or
consequences associated with each alternative.
• Uncertainty and Risk: The presence of
unknown or unpredictable factors.
Decision Criteria:
• Maximax Criterion: Select the alternative with the
highest possible payoff.
• Maximin Criterion: Choose the alternative with the
highest minimum payoff.
• Minimax Regret Criterion: Minimize the maximum
regret (difference between the best and actual
outcomes).
A payoff refers to the outcome or result associated with a
particular decision or strategy in a given situation. Payoffs are often
represented numerically and can be positive (indicating a gain or
benefit) or negative (indicating a loss or cost). The concept of
payoffs is fundamental in analyzing decision problems and strategic
interactions.
Key points related to payoffs:
• Numerical Representation: Payoffs are typically represented using numerical
values, which can be monetary (such as dollars, euros, etc.) or other quantifiable
units relevant to the specific context.
• Decision Alternatives: Each decision alternative or strategy in a decision problem
is associated with a set of possible payoffs. The decision-maker aims to choose
the alternative that maximizes positive payoffs or minimizes negative payoffs.
• States of Nature: Payoffs can depend on uncertain factors known as states of
nature. In decision-making under uncertainty, the likelihood of different states of
nature occurring is considered, and payoffs are often weighted by probabilities.
• Payoff Matrices and Tables: In decision analysis, payoff matrices or tables are
used to organize and present the payoffs associated with different combinations
of decisions and states of nature. These tools help decision-makers visualize and
analyze the consequences of their choices.
• Objective Function: In optimization problems, payoffs are part of the objective
function. The goal is to find the decision or strategy that maximizes or minimizes
the objective function, which is often expressed as a combination of payoffs.
Payoff Example:
Consider a decision problem involving the launch of a new
product. The decision-maker has two options: launch the product or
not launch it. The states of nature are "Market Success" and "Market
Failure." The payoffs, representing profits in thousands of dollars,
could be as follows:
Market is Successful Market Fails

Product is Launched Gain $50,000 Loose $10,000

Product is not launched Gain $20,000 Gain $5,000

In this case, the payoffs provide a clear indication of the


potential financial outcomes associated with each decision in
different states of nature. The decision-maker would then use
decision criteria, such as Maximax, Maximin, or Minimax Regret, to
make an informed choice based on these payoffs.
Examples Problems for Decision Criteria
Maximax Criterion:
• Scenario: Investment in Stocks A, B, or C. Potential
Profits (in thousands): A (50), B (30), C (40).
• Calculation: Choose the alternative (stock) with the
highest potential profit (Maximax).
Maximin Criterion:
• Scenario: Product Launch Strategies with Potential
Sales (in millions): Strategy 1 (5M), Strategy 2 (8M),
Strategy 3 (4M).
• Calculation: Choose the strategy with the highest
minimum potential sales (Maximin).
Examples Problems for Decision Criteria
Minimax Regret Criterion:
• Scenario: Manufacturing Processes with Regret
Matrix (in units): Process A, B, C. Regret values for
each pair.
• Calculation: Minimize regret by choosing the
process with the least maximum regret.
Decision Trees and Payoff Tables
Building Decision Trees:
• Decision trees visually represent decision
problems, showing the sequence of decisions
and possible outcomes. Nodes represent
decision points or chance events, branches show
possible paths, and leaves represent final
outcomes.
Decision Tree Analysis:
• EMV Calculation: Expected Monetary Value is
calculated by multiplying the payoff of each
outcome by its probability and summing them.
Example Problem: Decision Tree Analysis:
Suppose a company is deciding whether to invest in a new product line. The
decision is to be made based on two possible future states: "Success" and
"Failure." The company can choose between two alternatives: "Invest" or
"Do Not Invest." The payoffs (profits in millions) associated with each
combination of decision and future state are as follows:
• If the company invests and the new product line is successful, the payoff
is $5 million.
• If the company invests and the new product line fails, the payoff is -$2
million (a loss of $2 million).
• If the company does not invest and the new product line is successful,
the payoff is $1 million.
• If the company does not invest and the new product line fails, the payoff
is $0 (no profit, no loss).
• Now, let's assign probabilities to the future states:
• Probability of "Success" (S): 0.6
• Probability of "Failure" (F): 0.4
Success Failure
Payoff table: Invest $5 million -$2 million (loss)
Do not Invest $1 million $0

The decision tree for this scenario can be


represented as follows:
Decision Tree

Success 5M
Invest
Failure - 2M

Success 1M
Do not Invest
Failure 0
Now, let's calculate the expected payoffs for each decision:
• Invest:
– Expected Payoff = (Probability of Success * Payoff for Success) + (Probability of
Failure * Payoff for Failure)
– Expected Payoff = (0.6 * $5M) + (0.4 * (-$2M)) = $3M - $0.8M = $2.2M
• Do Not Invest:
– Expected Payoff = (Probability of Success * Payoff for Success) + (Probability of
Failure * Payoff for Failure)
– Expected Payoff = (0.6 * $1M) + (0.4 * $0) = $0.6M
Now, we compare the expected payoffs:
• Expected Payoff for Invest = $2.2M
• Expected Payoff for Do Not Invest = $0.6M
Since the expected payoff is higher for the "Invest" decision ($2.2M
> $0.6M), the rational decision based on decision theory would be to
invest in the new product line.
This analysis helps the company make an informed decision
considering both the potential payoffs and the probabilities of different
outcomes.

You might also like