Trial-Based Heuristic Tree Search for MDPs with Factored Action Spaces.
Florian GeißerDavid SpeckThomas KellerPublished in: SOCS (2020)
Keyphrases
- tree search
- state space
- action space
- markov decision processes
- state and action spaces
- reinforcement learning
- dynamic programming
- tree search algorithm
- continuous state spaces
- heuristic search
- continuous state
- optimal policy
- orders of magnitude
- game tree
- markov chain
- planning problems
- search space
- markov decision process
- partially observable
- markov decision problems
- search algorithm
- continuous action
- real valued
- infinite horizon
- search tree
- particle filter
- reinforcement learning algorithms
- finite state
- branch and bound
- initial state
- reward function
- function approximators