Intrinsically Motivated Hierarchical Policy Learning in Multi-objective Markov Decision Processes.
Sherif M. AbdelfattahKathryn E. MerrickJiankun HuPublished in: CoRR (2023)
Keyphrases
- markov decision processes
- multi objective
- reinforcement learning
- optimal policy
- partially observable
- state space
- evolutionary algorithm
- policy iteration
- markov decision process
- average reward
- stochastic games
- finite horizon
- state abstraction
- infinite horizon
- state and action spaces
- model based reinforcement learning
- learning algorithm
- hierarchical reinforcement learning
- markov decision problems
- genetic algorithm
- transition matrices
- average cost
- multi objective optimization
- policy evaluation
- supervised learning
- action space
- planning under uncertainty
- reward function
- decision makers
- dynamic programming