Certified policy synthesis for general Markov decision processes: An application in building automation systems.
Sofie HaesaertNathalie CauchiAlessandro AbatePublished in: Perform. Evaluation (2017)
Keyphrases
- markov decision processes
- optimal policy
- policy iteration
- markov decision process
- infinite horizon
- finite horizon
- average reward
- action space
- partially observable
- state space
- transition matrices
- reinforcement learning
- reward function
- average cost
- dynamic programming
- finite state
- decision processes
- decision theoretic planning
- state and action spaces
- decision problems
- total reward
- policy iteration algorithm
- risk sensitive
- markov decision problems
- planning under uncertainty
- factored mdps
- expected reward
- model based reinforcement learning
- policy evaluation
- reinforcement learning algorithms
- discounted reward
- reachability analysis
- objective function
- probabilistic planning
- stationary policies
- control policies
- continuous state spaces
- long run
- real valued