Variance Reduced Value Iteration and Faster Algorithms for Solving Markov Decision Processes.
Aaron SidfordMengdi WangXian WuYinyu YePublished in: SODA (2018)
Keyphrases
- markov decision processes
- policy iteration
- stochastic shortest path
- factored mdps
- state space
- finite state
- transition matrices
- policy iteration algorithm
- reachability analysis
- semi markov decision processes
- reinforcement learning
- optimal policy
- reinforcement learning algorithms
- average reward
- dynamic programming
- model based reinforcement learning
- partially observable markov decision processes
- infinite horizon
- learning algorithm
- policy evaluation
- markov decision problems
- fixed point
- model free
- planning under uncertainty
- decision theoretic planning
- markov decision process
- partially observable
- dynamical systems
- markov chain
- least squares
- convergence rate