Program Analysis with Local Policy Iteration.
George KarpenkovDavid MonniauxPhilipp WendlerPublished in: CoRR (2015)
Keyphrases
- policy iteration
- markov decision processes
- model free
- optimal policy
- fixed point
- reinforcement learning
- sample path
- least squares
- temporal difference
- finite state
- markov decision process
- policy evaluation
- average reward
- infinite horizon
- state space
- optimal control
- markov decision problems
- discounted reward
- convergence rate
- search algorithm
- neural network
- reward function
- graph cuts
- markov random field
- training data