Multi-Objective Approaches to Markov Decision Processes with Uncertain Transition Parameters.
Dimitri ScheftelowitschPeter BuchholzVahid HashemiHolger HermannsPublished in: VALUETOOLS (2017)
Keyphrases
- markov decision processes
- multi objective
- reinforcement learning
- finite state
- optimal policy
- policy iteration
- planning under uncertainty
- finite horizon
- reinforcement learning algorithms
- state space
- transition matrices
- dynamic programming
- infinite horizon
- decision processes
- action sets
- factored mdps
- average cost
- evolutionary algorithm
- model based reinforcement learning
- reachability analysis
- decision theoretic planning
- reward function
- genetic algorithm
- multiple objectives
- decision making
- average reward
- markov decision process
- pareto optimal
- multi objective optimization