Greatest fixed points of probabilistic min/max polynomial equations, and reachability for branching Markov decision processes.
Kousha EtessamiAlistair StewartMihalis YannakakisPublished in: Inf. Comput. (2018)
Keyphrases
- min max
- fixed point
- markov decision processes
- policy iteration
- polynomial equations
- stationary points
- state space
- dynamical systems
- dynamic programming
- sufficient conditions
- reinforcement learning
- optimal policy
- finite state
- decision theoretic planning
- belief propagation
- reinforcement learning algorithms
- probabilistic model
- action space
- partially observable
- generative model
- posterior probability
- bayesian networks
- objective function
- markov decision process
- search space
- variational inequalities
- average cost
- markov random field