Discounted deterministic Markov decision processes and discounted all-pairs shortest paths.
Omid MadaniMikkel ThorupUri ZwickPublished in: SODA (2009)
Keyphrases
- markov decision processes
- optimal policy
- state space
- infinite horizon
- reinforcement learning
- average reward
- finite state
- dynamic programming
- average cost
- decision theoretic planning
- stationary policies
- policy iteration
- reinforcement learning algorithms
- model based reinforcement learning
- finite horizon
- reachability analysis
- transition matrices
- planning under uncertainty
- partially observable
- action space
- factored mdps
- shortest path
- decision processes
- decentralized control
- action sets
- discounted reward
- markov decision process
- state abstraction
- reward function
- semi markov decision processes
- collaborative filtering
- search algorithm