Solving Markov decision processes with downside risk adjustment.
Abhijit GosaviAnish ParulekarPublished in: Int. J. Autom. Comput. (2016)
Keyphrases
- markov decision processes
- transition matrices
- risk sensitive
- semi markov decision processes
- finite state
- optimal policy
- dynamic programming
- state space
- reinforcement learning
- policy iteration
- policy iteration algorithm
- markov decision problems
- factored mdps
- decision theoretic planning
- planning under uncertainty
- finite horizon
- partially observable
- reachability analysis
- average reward
- stochastic shortest path
- reinforcement learning algorithms
- infinite horizon
- decision processes
- action space
- state abstraction
- average cost
- action sets
- linear programming
- policy evaluation
- state and action spaces
- planning problems
- model based reinforcement learning
- decision problems
- machine learning