Action decoupled SAC reinforcement learning with discrete-continuous hybrid action spaces.
Yahao XuYiran WeiKeyang JiangLi ChenDi WangHongbin DengPublished in: Neurocomputing (2023)
Keyphrases
- action space
- continuous action
- reinforcement learning
- continuous state
- continuous state spaces
- policy search
- state space
- markov decision processes
- real valued
- state and action spaces
- action selection
- reinforcement learning methods
- stochastic processes
- state action
- control policies
- multi agent
- reinforcement learning algorithms
- model free
- skill learning
- learning algorithm
- function approximation
- function approximators
- mobile robot
- partially observable markov decision processes
- control problems
- partially observable
- optimal policy
- supervised learning
- learning agent
- probability distribution
- dynamic programming