Anytime Algorithms for Solving Possibilistic MDPs and Hybrid MDPs.
Kim BautersWeiru LiuLluís GodoPublished in: FoIKS (2016)
Keyphrases
- markov decision problems
- anytime algorithms
- markov decision processes
- expected utility
- reinforcement learning
- state space
- optimal policy
- partially observable
- linear programming
- dynamic programming
- decision theoretic
- planning under uncertainty
- markov decision process
- queueing networks
- policy iteration
- utility function
- finite state
- average cost
- real time systems
- action space
- decision processes
- transition probabilities
- dynamical systems
- decision makers
- reward function
- infinite horizon
- pareto optimal
- initial state
- data sets
- linear program
- computational efficiency
- orders of magnitude
- learning algorithm