Control of exploitation-exploration meta-parameter in reinforcement learning.
Shin IshiiWako YoshidaJunichiro YoshimotoPublished in: Neural Networks (2002)
Keyphrases
- reinforcement learning
- control problems
- action selection
- exploration exploitation tradeoff
- robot control
- optimal control
- control system
- active exploration
- adaptive control
- state space
- control parameters
- meta level
- control strategies
- model based reinforcement learning
- reinforcement learning algorithms
- control strategy
- control policy
- temporal difference learning
- autonomous learning
- evolutionary algorithm
- data sets
- multi agent
- learning process
- model free