filmov
tv
Decision Analysis using Microsoft Excel Part 1

Показать описание
Decision Making Under Uncertainty involves making choices without knowing the probabilities of various outcomes. This is different from risk (where probabilities are known). In uncertain environments, you rely on decision criteria or strategies to make the most rational choice.
Here are the main strategies used when facing uncertainty:
1. Maximax (Optimistic Approach)
Goal: Maximize the maximum possible payoff.
Assumes: The best will happen.
Use When: You're risk-seeking or in a highly competitive, aggressive environment.
2. Maximin (Pessimistic or Conservative Approach)
Goal: Maximize the minimum payoff.
Assumes: The worst will happen.
Use When: You're risk-averse or when stakes are high.
3. Laplace Criterion (Equal Likelihood)
Goal: Choose based on average payoff, assuming all states are equally likely.
Use When: You have no info about probability.
Formula:
Expected Payoff = Average of all outcomes per alternative
4. Hurwicz Criterion (Weighted Optimism)
Goal: A compromise between maximax and maximin
Formula:
H = α(best Payoff) + (1 - α)(Worst Payoff)
α = coefficient of optimism (between 0 and 1)
Use when you're partially optimistic, and want flexibility.
5. Minimax Regret (Savage Criterion)
Goal: Minimize the maximum regret.
Regret = Best payoff in state - Actual payoff of a choice
Use When: You want to avoid feeling regret from not choosing the best.
Steps:
Build a regret table.
Find the maximum regret for each alternative.
Choose the one with the smallest maximum regret.
Here are the main strategies used when facing uncertainty:
1. Maximax (Optimistic Approach)
Goal: Maximize the maximum possible payoff.
Assumes: The best will happen.
Use When: You're risk-seeking or in a highly competitive, aggressive environment.
2. Maximin (Pessimistic or Conservative Approach)
Goal: Maximize the minimum payoff.
Assumes: The worst will happen.
Use When: You're risk-averse or when stakes are high.
3. Laplace Criterion (Equal Likelihood)
Goal: Choose based on average payoff, assuming all states are equally likely.
Use When: You have no info about probability.
Formula:
Expected Payoff = Average of all outcomes per alternative
4. Hurwicz Criterion (Weighted Optimism)
Goal: A compromise between maximax and maximin
Formula:
H = α(best Payoff) + (1 - α)(Worst Payoff)
α = coefficient of optimism (between 0 and 1)
Use when you're partially optimistic, and want flexibility.
5. Minimax Regret (Savage Criterion)
Goal: Minimize the maximum regret.
Regret = Best payoff in state - Actual payoff of a choice
Use When: You want to avoid feeling regret from not choosing the best.
Steps:
Build a regret table.
Find the maximum regret for each alternative.
Choose the one with the smallest maximum regret.