Adaptive co-construction of state and action spaces in reinforcement learning.
Masato NagayoshiHajime MuraoHisashi TamakiPublished in: Artif. Life Robotics (2011)
Keyphrases
- state and action spaces
- reinforcement learning
- markov decision processes
- action space
- state space
- markov decision problems
- partially observable markov decision process
- average reward
- function approximation
- dynamic programming
- neural network
- decision theoretic
- multi agent
- partially observable
- single agent
- action selection
- decision processes
- markov decision process
- partially observable markov decision processes
- optimal control
- real valued
- linear programming
- least squares
- search algorithm