From Perturbation Analysis to Markov Decision Processes and Reinforcement Learning.
Xi-Ren CaoPublished in: Discret. Event Dyn. Syst. (2003)
Keyphrases
- markov decision processes
- reinforcement learning
- reinforcement learning algorithms
- state space
- optimal policy
- finite state
- dynamic programming
- action space
- partially observable
- policy iteration
- state and action spaces
- action sets
- transition matrices
- model based reinforcement learning
- planning under uncertainty
- markov decision process
- finite horizon
- average reward
- reachability analysis
- function approximation
- control problems
- decision processes
- model free
- learning algorithm
- machine learning
- average cost
- infinite horizon
- initial state
- partially observable markov decision processes
- reward function
- approximate dynamic programming
- actor critic
- state abstraction
- decentralized control
- temporal difference
- total reward
- optimal control
- linear program