On the Complexity of Partially Observed Markov Decision Processes.
Dima BuragoMichel de RougemontAnatol SlissenkoPublished in: Theor. Comput. Sci. (1996)
Keyphrases
- markov decision processes
- partially observed
- finite state
- optimal policy
- expected reward
- transition matrices
- reinforcement learning
- decision theoretic planning
- state space
- markov decision process
- policy iteration
- infinite horizon
- model based reinforcement learning
- dynamic programming
- decision problems
- partially observable
- reachability analysis
- average cost
- factored mdps
- average reward
- reinforcement learning algorithms
- risk sensitive
- computational complexity
- planning under uncertainty
- decision processes
- action space
- finite horizon
- probabilistic planning
- state and action spaces
- action sets
- stochastic shortest path
- partially observable markov decision processes
- reward function
- interval estimation