Symblicit algorithms for optimal strategy synthesis in monotonic Markov decision processes.
Aaron BohyVéronique BruyèreJean-François RaskinPublished in: SYNT (2014)
Keyphrases
- markov decision processes
- policy iteration
- optimal strategy
- optimal policy
- factored mdps
- finite state
- state space
- learning algorithm
- computational complexity
- reinforcement learning
- transition matrices
- planning under uncertainty
- decision problems
- dynamic programming
- cost function
- learning rate
- partially observable markov decision processes
- action space
- decision theoretic planning
- lower bound