Numerical analysis of continuous time Markov decision processes over finite horizons.
Peter BuchholzIngo SchulzPublished in: Comput. Oper. Res. (2011)
Keyphrases
- numerical analysis
- markov decision processes
- stationary policies
- state and action spaces
- infinite horizon
- state space
- finite state
- optimal policy
- dynamic programming
- image enhancement
- policy iteration
- optimal control
- reinforcement learning
- transition matrices
- finite horizon
- markov chain
- partially observable
- decision theoretic planning
- action sets
- dynamical systems
- average reward
- reachability analysis
- planning under uncertainty
- markov decision process
- decision processes
- average cost
- reinforcement learning algorithms
- action space
- model based reinforcement learning
- risk sensitive
- factored mdps
- interval estimation
- semi markov decision processes
- sufficient conditions
- linear program
- finite number
- reward function
- fixed point
- state abstraction
- probability distribution
- machine learning
- multi agent
- partially observable markov decision processes