Approximate Policy Iteration for Markov Decision Processes via Quantitative Adaptive Aggregations.
Alessandro AbateMilan CeskaMarta KwiatkowskaPublished in: ATVA (2016)
Keyphrases
- markov decision processes
- policy iteration
- markov games
- optimal policy
- reinforcement learning
- finite state
- state space
- planning under uncertainty
- markov decision process
- reinforcement learning algorithms
- markov decision problems
- average cost
- dynamic programming
- model free
- average reward
- partially observable
- finite horizon
- fixed point
- action space
- decision problems
- least squares
- temporal difference
- stochastic games