Q-learning in Continuous Action Space by Extending EVA.
Toi TsunedaDaiki KuyoshiSatoshi YamanePublished in: CANDAR (Workshops) (2020)
Keyphrases
- action space
- state space
- continuous state spaces
- reinforcement learning
- action selection
- state action
- reinforcement learning methods
- markov decision processes
- continuous state
- single agent
- real valued
- state and action spaces
- dynamic programming
- heuristic search
- optimal policy
- control policies
- multi agent
- stochastic processes
- reinforcement learning algorithms
- function approximation
- markov chain
- markov decision problems
- state variables
- policy iteration
- cooperative
- temporal difference
- model free
- machine learning
- markov decision process
- planning problems
- function approximators
- dynamical systems
- learning algorithm
- belief state
- control problems
- search space
- multi agent systems
- search algorithm
- decision making