LPI: Software Verification with Local Policy Iteration - (Competition Contribution).
Egor George KarpenkovPublished in: TACAS (2016)
Keyphrases
- policy iteration
- markov decision processes
- model free
- reinforcement learning
- fixed point
- sample path
- least squares
- optimal policy
- finite state
- markov decision process
- temporal difference
- average reward
- optimal control
- infinite horizon
- linear programming
- policy evaluation
- state space
- markov decision problems
- function approximation
- sufficient conditions
- bayesian networks
- markov chain
- dynamic programming
- cost function
- graphical models