Q-Learning in Continuous State-Action Space with Noisy and Redundant Inputs by Using a Selective Desensitization Neural Network.
Takaaki KobayashiTakeshi ShibuyaMasahiko MoritaPublished in: J. Adv. Comput. Intell. Intell. Informatics (2015)
Keyphrases
- state action space
- function approximation
- neural network
- reinforcement learning
- radial basis function
- optimal policy
- function approximators
- actor critic
- state space
- artificial neural networks
- action space
- model free
- temporal difference
- temporal difference learning
- learning tasks
- reinforcement learning algorithms
- mountain car
- fuzzy logic
- genetic algorithm
- control policy
- reinforcement learning methods
- multi agent
- learning algorithm
- adaptive control
- support vector