Globally Optimal Hierarchical Reinforcement Learning for Linearly-Solvable Markov Decision Processes.
Guillermo InfanteAnders JonssonVicenç GómezPublished in: CoRR (2021)
Keyphrases
- globally optimal
- markov decision processes
- hierarchical reinforcement learning
- state abstraction
- reinforcement learning
- average reward
- reward function
- policy iteration
- markov decision process
- state space
- optimal policy
- graph cuts
- reinforcement learning algorithms
- finite state
- decision theoretic planning
- dynamic programming
- transition matrices
- model free
- partially observable
- np hard
- average cost
- infinite horizon
- action space
- computational complexity
- random walk
- markov decision problems
- markov random field
- higher order