Finite horizon risk sensitive MDP and linear programming.
Atul KumarVeeraruna KavithaNandyala HemachandraPublished in: CDC (2015)
Keyphrases
- risk sensitive
- finite horizon
- linear programming
- markov decision processes
- average cost
- dynamic programming
- optimal policy
- markov decision problems
- linear program
- markov decision process
- infinite horizon
- state space
- finite state
- control policies
- reinforcement learning
- policy iteration
- optimal control
- np hard
- reward function
- average reward
- action space
- objective function
- partially observable
- optimal solution
- multistage
- sufficient conditions
- decision processes
- decision making
- computational complexity
- long run
- decision problems
- finite number