A Dissipativity Theory for Undiscounted Markov Decision Processes.
Sébastien GrosMario ZanonPublished in: CoRR (2021)
Keyphrases
- markov decision processes
- policy iteration
- state space
- average reward
- finite state
- reinforcement learning
- optimal policy
- infinite horizon
- transition matrices
- average cost
- decision theoretic planning
- decision processes
- reachability analysis
- partially observable
- action space
- reinforcement learning algorithms
- dynamic programming
- risk sensitive
- planning under uncertainty
- model based reinforcement learning
- stochastic games
- finite horizon
- markov decision process
- action sets
- state and action spaces
- semi markov decision processes
- reward function
- factored mdps
- total reward
- state abstraction
- search algorithm
- fixed point
- data mining