Switched Linear Systems Meet Markov Decision Processes: Stability Guaranteed Policy Synthesis.
Bo WuMurat CubuktepeUfuk TopcuPublished in: CDC (2019)
Keyphrases
- linear systems
- markov decision processes
- optimal policy
- sufficient conditions
- policy iteration
- markov decision process
- average reward
- state space
- finite horizon
- infinite horizon
- dynamical systems
- state and action spaces
- action space
- partially observable
- finite state
- reward function
- average cost
- decision processes
- reinforcement learning
- dynamic programming
- decision problems
- long run
- discounted reward
- expected reward
- markov decision problems
- policy evaluation
- total reward
- partially observable markov decision processes
- decision theoretic planning
- reinforcement learning algorithms
- transition matrices
- sparse linear systems
- continuous state spaces
- stationary policies
- fixed point
- actor critic
- learning algorithm
- support vector