POMDP inference and robust solution via deep reinforcement learning: An application to railway optimal maintenance.
Giacomo ArcieriCyprien HoelzlOliver SchweryDaniel StraubKonstantinos G. PapakonstantinouEleni N. ChatziPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- optimal solution
- state space
- control policy
- closed form
- approximate dynamic programming
- optimal policy
- dynamic programming
- optimal control
- markov decision processes
- multi agent
- continuous state
- robust optimization
- reward function
- function approximation
- worst case
- bayesian networks
- policy evaluation
- markov decision process
- continuous state spaces
- estimation error
- partially observable
- model free
- machine learning
- dynamical systems
- reinforcement learning algorithms
- partially observable markov decision processes
- policy iteration
- finite state
- reinforcement learning methods
- hidden state
- dynamic environments
- learning algorithm