9.1 Multiagent decision making and Games in Normal Form Flashcards
Multiagent decision making:
Group of agents ___ within an ___ and make simultaneous
coexist
Environment
Decisions
Game theory:
Mathematical study of ___ among ___, ___ agents
Interactions
Independent
Self-interested
The dominant theory for modeling agent’s interests is ___
Utility theory
In Game Theory, the success of an agent depends on the
___ of ___
Decisions
All other agents
Assumptions of Game Theory:
- Participant agents are ___
- Agents reason ___ (they take into account the other agent’s decision)
Rational
Strategically
Normal-form games (rr s´Static game or Strategic game):
- Each agent chooses a ___ and then receives a ___ that depends on the ___
- The joint action is the ___ of the game
- Although payoffs are common knowledge, an agent does not know the ___
- A game’s solution is a ___ of the outcome using the assumptions that all agents are ___ and ___
Single action Payoff Joint action Outcome Actions of the other agents Prediction Rational and strategic
In game theory, we assume that a rational agent will never choose a ___
Suboptimal action
Z is a suboptimal action (or strictly dominated action) in relation to Y if the ___ of Y is always ___ independently of the ___ of the other agents
Utility
Higher
Actions
Solution Concepts:
The problem of ___ about games and identifying certain ___ of ___
Reasoning
Subsets of outcomes
The iterated ___ of ___ actions is a ___ solution concept for games
Elimination
Strictly dominated
First
Drawbacks of iterated elimination of strictly dominated actions:
- Each step requires a further assumption about what the agents
know about each other’s ___
- Very imprecise prediciton about the play of the game, for instance, if no actions are eliminated then ___
Rationality
Anything could happen
Nash equilibrium:
Joint ___ from where no agent can ___ improve his ___. Hence no agent has ___ to deviate
Action
Unilaterally
Payoff
Incentive
In a Nash equilibrium, each agent’s actions is an
___ response to the other agents’ ___
Optimal
Actions
The cost of Nash equilibrium is ___ in the number of agents
Exponencial
If iterated elimination of strictly dominated actions eliminates all but a single joint action Z, then this joint action is the unique ___ of the game
Nash equilibrium