Nonapproximability Results for Partially Observable Markov Decision Processes.
Christopher LusenaJudy GoldsmithMartin MundhenkPublished in: J. Artif. Intell. Res. (2001)
Keyphrases
- partially observable markov decision processes
- finite state
- reinforcement learning
- dynamical systems
- decision problems
- markov decision processes
- dynamic programming
- belief state
- planning under uncertainty
- optimal policy
- belief space
- continuous state
- state space
- partial observability
- stochastic domains
- multi agent
- partially observable stochastic games
- sequential decision making problems
- planning problems
- partially observable domains
- data mining
- partially observable
- markov random field
- approximate solutions
- dec pomdps
- partially observable markov
- infinite horizon
- partially observable markov decision process
- initial state
- orders of magnitude
- particle filter
- linear programming
- decision trees
- learning algorithm