Thompson Sampling for Partially Observable Linear-Quadratic Control.
Taylan KarginSahin LaleKamyar AzizzadenesheliAnima AnandkumarBabak HassibiPublished in: ACC (2023)
Keyphrases
- partially observable
- linear quadratic
- dynamical systems
- optimal control
- infinite horizon
- state space
- reinforcement learning
- partial observability
- decision problems
- markov decision processes
- closed loop
- partial observations
- control system
- vector valued
- control strategy
- orders of magnitude
- belief state
- sufficient conditions
- probability distribution
- feature space