POMDPs in Continuous Time and Discrete Spaces.
Bastian AltMatthias SchultheisHeinz KoepplPublished in: NeurIPS (2020)
Keyphrases
- dynamical systems
- partially observable markov decision processes
- state space
- reinforcement learning
- partially observable
- belief state
- markov chain
- optimal control
- markov decision processes
- predictive state representations
- dynamic programming
- finite state
- point based value iteration
- iterative learning control
- markov processes
- planning under uncertainty
- belief space
- optimal policy
- continuous state
- distributed constraint optimization
- markov decision problems
- partially observable markov decision process
- policy iteration algorithm
- dec pomdps
- continuous time bayesian networks
- policy search
- average reward
- machine learning
- model free
- multi agent
- search algorithm