Reinforcement Learning based on MPC/MHE for Unmodeled and Partially Observable Dynamics.
Hossein Nejatbakhsh EsfahaniArash Bahari KordabadSébastien GrosPublished in: ACC (2021)
Keyphrases
- partially observable
- reinforcement learning
- dynamical systems
- dynamic model
- state space
- partial observability
- markov decision processes
- partially observable domains
- decision problems
- markov decision problems
- optimal control
- partially observable environments
- partial observations
- infinite horizon
- hidden state
- reward function
- reinforcement learning algorithms
- learning algorithm
- function approximation
- dynamic programming
- optimal policy
- belief space
- action models
- partially observable markov decision processes
- fully observable
- belief state
- dynamic systems
- markov decision process
- action selection
- machine learning
- partially observable markov decision process
- transfer learning
- closed loop
- mathematical model
- model free
- planning domains
- utility function
- multi agent