Discounted Markov Decision Processes and Algorithms for Solving Stochastic Control Problem on Networks.
Dmitrii LozovanuStefan PicklPublished in: CTW (2011)
Keyphrases
- markov decision processes
- policy iteration
- optimal policy
- factored mdps
- transition matrices
- finite state
- stochastic shortest path
- dynamic programming
- state space
- reinforcement learning
- finite horizon
- infinite horizon
- markov decision process
- markov decision problems
- learning algorithm
- reinforcement learning algorithms
- average cost
- average reward
- partially observable markov decision processes
- optimal control
- action space
- convergence rate
- least squares