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