Multi-weighted Markov Decision Processes with Reachability Objectives.
Patricia BouyerMauricio GonzálezNicolas MarkeyMickael RandourPublished in: GandALF (2018)
Keyphrases
- markov decision processes
- state space
- optimal policy
- transition matrices
- reinforcement learning
- finite state
- dynamic programming
- policy iteration
- decision theoretic planning
- partially observable
- risk sensitive
- planning under uncertainty
- factored mdps
- reachability analysis
- markov decision process
- decision processes
- state abstraction
- semi markov decision processes
- model based reinforcement learning
- finite horizon
- action space
- reinforcement learning algorithms
- average cost
- multiple objectives
- search space
- average reward
- infinite horizon
- state and action spaces
- action sets
- discounted reward
- heuristic search
- multistage