The Optimal Observability of Partially Observable Markov Decision Processes: Discrete State Space.
Mohammad RezaeianBa-Ngu VoJamie S. EvansPublished in: IEEE Trans. Autom. Control. (2010)
Keyphrases
- state space
- partially observable markov decision processes
- continuous state spaces
- partially observable
- dynamic programming
- partial observability
- belief state
- dynamical systems
- continuous state
- belief space
- optimal policy
- markov decision processes
- reinforcement learning
- stochastic domains
- continuous action
- planning under uncertainty
- decision problems
- finite state
- heuristic search
- partially observable stochastic games
- planning problems
- action space
- partially observable markov decision process
- optimal control
- average reward
- control policies
- markov decision process
- orders of magnitude
- particle filter
- optimal solution
- goal state
- dynamic systems
- search space
- decision making
- fully observable
- markov chain
- state variables
- worst case
- partially observable markov