Near-optimal Reinforcement Learning in Factored MDPs: Oracle-Efficient Algorithms for the Non-episodic Setting.
Ziping XuAmbuj TewariPublished in: CoRR (2020)
Keyphrases
- factored mdps
- reinforcement learning
- markov decision processes
- state space
- approximate dynamic programming
- policy iteration
- markov decision problems
- context specific
- algebraic decision diagrams
- function approximation
- optimal policy
- reinforcement learning algorithms
- learning algorithm
- machine learning
- transition model
- dynamic programming
- planning under uncertainty
- partially observable
- markov decision process
- finite state
- stochastic processes
- action space
- markov chain
- multi agent
- hidden markov models
- supervised learning
- decision processes
- function approximators
- linear program
- basis functions
- reward function
- temporal difference
- random fields
- dynamical systems
- fixed point