An approach for dynamic selection of synthesis transformations based on Markov Decision Processes.
Tobias WelpAndreas KuehlmannPublished in: DATE (2011)
Keyphrases
- markov decision processes
- state space
- finite state
- optimal policy
- dynamic programming
- reinforcement learning
- transition matrices
- reinforcement learning algorithms
- decision theoretic planning
- planning under uncertainty
- reachability analysis
- infinite horizon
- average reward
- policy iteration
- average cost
- partially observable
- model based reinforcement learning
- risk sensitive
- factored mdps
- markov decision process
- decision processes
- finite horizon
- state abstraction
- reward function
- search algorithm
- action sets
- learning algorithm