Büchi Objectives in Countable MDPs.
Stefan KieferRichard MayrMahsa ShirmohammadiPatrick TotzkePublished in: CoRR (2019)
Keyphrases
- markov decision processes
- average cost
- reinforcement learning
- state space
- markov chain
- factored mdps
- optimal policy
- long run
- state and action spaces
- multi agent
- dynamic programming
- finite horizon
- finite number
- neural network
- semi markov decision processes
- initial state
- linear program
- linear programming
- finite state
- multiple objectives
- information gain
- policy iteration
- decision diagrams
- learning algorithm
- data sets
- real time dynamic programming
- factored markov decision processes