Online Markov Decision Processes Configuration with Continuous Decision Space.
Davide MaranPierriccardo OlivieriFrancesco Emanuele StradiGiuseppe UrsoNicola GattiMarcello RestelliPublished in: AAAI (2024)
Keyphrases
- markov decision processes
- action space
- decision processes
- state space
- optimal policy
- reinforcement learning
- transition matrices
- finite state
- dynamic programming
- decision theoretic planning
- reachability analysis
- policy iteration
- risk sensitive
- average reward
- partially observable
- continuous state
- continuous state spaces
- infinite horizon
- reinforcement learning algorithms
- planning under uncertainty
- markov decision process
- finite horizon
- decision problems
- decision making
- state and action spaces
- factored mdps
- average cost
- model based reinforcement learning
- markov decision problems
- heuristic search
- machine learning
- interval estimation
- action sets
- reward function
- decision process
- influence diagrams
- real valued
- linear programming
- learning algorithm