Parametrized Deep Q-Networks Learning: Reinforcement Learning with Discrete-Continuous Hybrid Action Space.
Jiechao XiongQing WangZhuoran YangPeng SunLei HanYang ZhengHaobo FuTong ZhangJi LiuHan LiuPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- action space
- continuous state spaces
- continuous state
- state space
- continuous action
- learning algorithm
- markov decision processes
- policy search
- reinforcement learning methods
- function approximation
- action selection
- state action
- learning tasks
- machine learning
- function approximators
- computational complexity
- partially observable
- stochastic processes
- optimal control
- multi agent
- decision making
- learning problems
- state and action spaces
- dynamic programming