Meta-learning linear quadratic regulators: A policy gradient MAML approach for model-free LQR.
Leonardo Felipe TosoDonglin ZhanJames AndersonHan WangPublished in: L4DC (2024)
Keyphrases
- linear quadratic
- model free
- meta learning
- policy gradient
- optimal control
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- learning tasks
- closed loop
- average reward
- model selection
- dynamical systems
- vector valued
- machine learning algorithms
- machine learning
- reinforcement learning methods
- dynamic programming
- temporal difference
- transfer learning
- learning problems
- learning algorithm
- state space
- rl algorithms
- decision trees
- policy iteration
- gaussian model
- supervised learning
- multi agent
- markov decision processes
- stochastic games
- image features
- learning experience
- data mining