01_Decision Making under Uncertainty Flashcards
Decision Analysis
3 Levels of Uncertainty
1. Decision under Certainty
- future is known and described by single scenario
- deterministic
2. Decision under Uncertainty
- multiple different scenarios for future
- probabilities are unknown
3. Decision under Risk
- multiple different scenarios for future
- probabilities are known
Number of goals
2 items
- Single Goal: single criterion decision
- Multiple Goals: multi-criteria decision
Number of Decision Makers
2 Alternatives
- single decision maker (DM)
- group decision maker
Time distinction
2 items
- Static decision
- Dynamic decision
Behaviour of Ppponent
2 Alternatives
-
Game against Nature
- randomly acting opponent -
Game Theory (Game against active antagonist)
- strategically acting opponent
5 Decision Rules
- Maximax-Rule
- Maximin-Rule
- Minimax-Regret-Rule
- Hurwicz-Rule
- Lapalce-Rule
- Dominance Criterion
- Efficient alternative
Dominance Criterion
- an action a(i) dominates another action a(j), if it generates an equal or better outcome in every scenario s(n)
Efficient alternative
- an action a(i) is efficient, if it is not dominated by another alternative
Exam Question
Determine the efficient actions. Motivate your answers.
practice wording
- a(1) outperforms a(3) in scenario s(1); in every other scenario a(1) does not perform worse than a(3)
- therefore a(1) dominates a(3)
- neither a(1) nor a(2) dominate the other one respectively
- all alternatives are efficient if
x ∈ {11,12}
3 Decision Situations
depending on Level of Uncertainty
- Deterministic
- Uncertainty (without probabilities)
- Risk (with probabilities)
5 Characterizations of Decision Situations
- Level of Uncertainty
- Number of Goals
- Number of Decision Makers
- Development over Time
- Type of Antagonist
Describe the Hurwicz-Rule und the Risk Attitude reflected by this rule
- Alternative which maximizes the linear combination of the maximum and minimum profit.
- for λ = 1 the rule equals the Maximax-Rule (risk seeking decision maker)
- for λ = 0 the rule equals the Maximin-Rule (risk averse decision maker)