Thompson Sampling Based Monte-Carlo Planning in POMDPs.
Aijun BaiFeng WuZongzhang ZhangXiaoping ChenPublished in: ICAPS (2014)
Keyphrases
- monte carlo
- partially observable markov decision processes
- importance sampling
- predictive state representations
- adaptive sampling
- monte carlo simulation
- temporal difference
- partially observable
- monte carlo methods
- belief state
- planning problems
- reinforcement learning
- belief space
- motion planning
- markovian decision
- markov chain monte carlo
- finite state
- optimal strategy
- matrix inversion
- markov decision problems
- stochastic approximation
- monte carlo method
- variance reduction
- markov decision processes
- decision problems
- particle filter
- dynamic programming
- dynamical systems
- heuristic search
- optimal policy
- planning domains
- global illumination
- policy gradient
- decision theoretic
- search space
- search algorithm
- learning algorithm