Developing reinforcement learning for adaptive co-construction of continuous high-dimensional state and action spaces.
Masato NagayoshiHajime MuraoHisashi TamakiPublished in: Artif. Life Robotics (2012)
Keyphrases
- state and action spaces
- action space
- reinforcement learning
- high dimensional
- markov decision processes
- state space
- real valued
- action selection
- markov decision problems
- feature space
- function approximation
- single agent
- partially observable markov decision process
- optimal policy
- markov decision process
- average reward
- search space
- function approximators
- machine learning