What is decidable about partially observable Markov decision processes with ω-regular objectives.
Krishnendu ChatterjeeMartin ChmelikMathieu TracolPublished in: J. Comput. Syst. Sci. (2016)
Keyphrases
- partially observable markov decision processes
- finite state
- dynamical systems
- reinforcement learning
- planning under uncertainty
- belief state
- decision problems
- continuous state
- dynamic programming
- belief space
- optimal policy
- state space
- markov decision processes
- partial observability
- planning problems
- partially observable stochastic games
- partially observable domains
- stochastic domains
- multi agent
- partially observable
- markov chain
- partially observable markov decision process
- heuristic search
- domain specific
- predictive state representations
- point based value iteration
- partially observable markov
- sequential decision making problems
- dec pomdps
- infinite horizon
- special case
- situation calculus
- lower bound
- decision trees