Policy-Based Reinforcement Learning in the Generalized Rock-Paper-Scissors Game.
Mali Imre GergelyGabriela CzibulaPublished in: ESANN (2023)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- agent learns
- markov decision process
- action selection
- markov games
- game playing
- reinforcement learning problems
- function approximators
- state space
- reinforcement learning algorithms
- state and action spaces
- control policy
- temporal difference learning
- function approximation
- temporal difference
- partially observable environments
- average reward
- game theoretic
- decision problems
- markov decision processes
- actor critic
- educational games
- game theory
- computer games
- game play
- serious games
- partially observable
- action space
- dynamic programming
- video games
- optimal control
- game design
- virtual world
- policy evaluation
- state action
- markov decision problems
- policy gradient
- policy gradient methods
- transition model
- reinforcement learning methods
- stochastic games
- nash equilibria
- reward function
- model free
- nash equilibrium
- control policies
- continuous state
- approximate dynamic programming
- policy iteration
- continuous state spaces
- inverse reinforcement learning
- optimal strategy
- learning experience
- learning algorithm
- machine learning