On the convex formulations of robust Markov decision processes.
Julien Grand-ClémentMarek PetrikPublished in: CoRR (2022)
Keyphrases
- markov decision processes
- state space
- reinforcement learning
- transition matrices
- optimal policy
- policy iteration
- finite state
- reachability analysis
- dynamic programming
- planning under uncertainty
- decision processes
- finite horizon
- decision theoretic planning
- model based reinforcement learning
- average cost
- action space
- average reward
- reinforcement learning algorithms
- risk sensitive
- action sets
- stochastic shortest path
- partially observable
- factored mdps
- semi markov decision processes
- fixed point
- data mining
- infinite horizon
- optimal control