Evolutionary policy iteration for solving Markov decision processes.
Hyeong Soo ChangHong-Gi LeeMichael C. FuSteven I. MarcusPublished in: IEEE Trans. Autom. Control. (2005)
Keyphrases
- markov decision processes
- policy iteration
- transition matrices
- markov decision problems
- policy iteration algorithm
- optimal policy
- finite state
- state space
- factored mdps
- sample path
- average reward
- infinite horizon
- model free
- approximate dynamic programming
- dynamic programming
- reinforcement learning
- markov decision process
- least squares
- fixed point
- decision theoretic planning
- policy evaluation
- planning under uncertainty
- stochastic shortest path
- partially observable
- average cost
- reinforcement learning algorithms
- reward function
- discounted reward
- finite horizon
- decision processes
- action space
- temporal difference
- machine learning
- stochastic games
- partially observable markov decision processes