Memory-two strategies forming symmetric mutual reinforcement learning equilibrium in repeated prisoner's dilemma game.
Masahiko UedaPublished in: CoRR (2021)
Keyphrases
- mixed strategy
- reinforcement learning
- repeated games
- nash equilibrium
- equilibrium strategies
- game theory
- pure strategy
- nash equilibria
- game theoretic
- games with incomplete information
- optimal strategy
- solution concepts
- equilibrium point
- stochastic games
- incomplete information
- exploration exploitation dilemma
- function approximation
- subgame perfect equilibrium
- model free
- markov games
- evolutionary game theory
- extensive form games
- fictitious play
- game design
- memory requirements
- reinforcement learning algorithms
- computer games
- average reward
- two player games
- machine learning
- learning algorithm
- video games
- perfect information
- markov decision processes
- learning agents
- exploration strategy
- decision problems
- learning process
- multi agent systems
- multi agent
- state space
- temporal difference learning