The Divergence of Reinforcement Learning Algorithms with Value-Iteration and Function Approximation
Michael FairbankEduardo AlonsoPublished in: CoRR (2011)
Keyphrases
- reinforcement learning algorithms
- function approximation
- reinforcement learning
- markov decision processes
- policy iteration
- model free
- state space
- temporal difference
- optimal policy
- reinforcement learning problems
- average reward
- temporal difference learning
- markov decision process
- dynamic programming
- reinforcement learning methods
- partially observable markov decision processes
- function approximators
- td learning
- learning tasks
- finite state
- control problems
- radial basis function
- policy search
- reward function
- data mining
- partially observable
- average cost
- least squares
- learning process
- infinite horizon
- markov chain
- transfer learning