An On-Line Planner for POMDPs with Large Discrete Action Space: A Quantile-Based Approach.
Erli WangHanna KurniawatiDirk P. KroesePublished in: ICAPS (2018)
Keyphrases
- action space
- continuous state
- continuous state spaces
- continuous action
- state space
- markov decision processes
- reinforcement learning
- real valued
- partially observable markov decision processes
- heuristic search
- control policies
- policy search
- action selection
- state and action spaces
- finite state
- optimal policy
- markov decision process
- initial state
- state action
- machine learning
- stochastic processes
- belief state
- domain independent
- dynamic programming
- markov decision problems
- decision problems
- orders of magnitude
- ai planning
- policy iteration
- domain specific
- cooperative
- dynamical systems