Q-learning for history-based reinforcement learning.
Mayank DaswaniPeter SunehagMarcus HutterPublished in: ACML (2013)
Keyphrases
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- model free
- state space
- optimal policy
- learning algorithm
- temporal difference learning
- multi agent reinforcement learning
- continuous state and action spaces
- markov decision processes
- temporal difference
- state action space
- reinforcement learning methods
- supervised learning
- stochastic approximation
- control problems
- multi agent
- rl algorithms
- learning agent
- machine learning
- eligibility traces
- learning capabilities
- transfer learning
- continuous state
- cooperative
- reward shaping
- action space
- learning agents
- markov decision process
- dynamic programming
- markov decision problems
- policy evaluation
- multiagent learning
- policy search
- fixed point
- action selection
- radial basis function