Global Versus Local Constructive Function Approximation for On-Line Reinforcement Learning.
Peter VamplewRobert OllingtonPublished in: Australian Conference on Artificial Intelligence (2005)
Keyphrases
- function approximation
- reinforcement learning
- temporal difference
- mountain car
- temporal difference learning
- tile coding
- temporal difference learning algorithms
- learning tasks
- function approximators
- state space
- state action space
- reinforcement learning algorithms
- model free
- radial basis function
- transfer learning
- action selection
- optimal policy
- policy gradient
- artificial neural networks
- multi agent
- machine learning