Deep Reinforcement Learning with an Unbounded Action Space.
Ji HeJianshu ChenXiaodong HeJianfeng GaoLihong LiLi DengMari OstendorfPublished in: CoRR (2015)
Keyphrases
- action space
- reinforcement learning
- state space
- markov decision processes
- state and action spaces
- real valued
- continuous state
- action selection
- state action
- control policies
- continuous state spaces
- stochastic processes
- function approximation
- function approximators
- reinforcement learning methods
- markov decision problems
- optimal policy
- multi agent
- heuristic search
- learning algorithm
- dynamic programming
- markov decision process
- single agent
- reinforcement learning algorithms
- markov chain
- pairwise
- control problems
- optimal control
- markov random field
- supervised learning
- search space