Optimal decisions for continuous time Markov decision processes over finite planning horizons.
Peter BuchholzIryna DohndorfDimitri ScheftelowitschPublished in: Comput. Oper. Res. (2017)
Keyphrases
- state variables
- state space
- markov decision processes
- optimal decisions
- planning problems
- heuristic search
- state and action spaces
- infinite horizon
- partially observable
- decision theoretic planning
- optimal policy
- planning under uncertainty
- markov chain
- dynamic programming
- reinforcement learning
- globally optimal
- classical planning
- decision making
- decision theory
- stationary policies
- finite state
- markov decision problems
- policy iteration
- probabilistic planning
- action space
- dynamical systems
- partially observable markov decision process
- preference elicitation
- belief state
- initial state
- average cost
- markov decision process
- reward function
- search space
- transition matrices
- partially observable markov decision processes
- ai planning
- optimal control
- decision problems
- factored mdps
- action selection
- decision theoretic
- finite number
- theoretical framework
- sufficient conditions
- real time dynamic programming