Markov Decision Processes with Arbitrary Reward Processes.
Jia Yuan YuShie MannorNahum ShimkinPublished in: EWRL (2008)
Keyphrases
- markov decision processes
- average reward
- reinforcement learning
- reward function
- total reward
- optimal policy
- discounted reward
- finite state
- decision processes
- state space
- expected reward
- policy iteration
- reinforcement learning algorithms
- transition matrices
- decision theoretic planning
- dynamic programming
- planning under uncertainty
- stationary policies
- partially observable
- model based reinforcement learning
- average cost
- reachability analysis
- risk sensitive
- factored mdps
- finite horizon
- infinite horizon
- long run
- state abstraction
- learning algorithm
- markov decision process
- action sets
- optimal solution
- model free
- action space