CAQL: Continuous Action Q-Learning.
Moonkyung RyuYinlam ChowRoss AndersonChristian TjandraatmadjaCraig BoutilierPublished in: ICLR (2020)
Keyphrases
- continuous action
- continuous state and action spaces
- reinforcement learning
- policy search
- continuous state
- reinforcement learning algorithms
- action space
- state space
- function approximation
- partially observable markov decision processes
- cooperative
- learning algorithm
- multi agent
- model free
- state action
- optimal policy
- action selection
- reward function
- learning rate
- state variables
- robot navigation
- neural network
- state dependent
- control policies
- dynamical systems
- dynamic programming
- machine learning