Conditional Value-at-Risk for Reachability and Mean Payoff in Markov Decision Processes.
Jan KretínskýTobias MeggendorferPublished in: CoRR (2018)
Keyphrases
- markov decision processes
- state space
- finite state
- reinforcement learning
- optimal policy
- decision theoretic planning
- transition matrices
- policy iteration
- game theory
- reinforcement learning algorithms
- action space
- heuristic search
- average cost
- reachability analysis
- risk sensitive
- decision processes
- dynamic programming
- finite horizon
- average reward
- markov decision process
- infinite horizon
- factored mdps
- partially observable
- decision diagrams
- state abstraction
- planning problems
- dynamical systems
- stochastic games
- nash equilibrium
- semi markov decision processes