Epistemic Risk-Sensitive Reinforcement Learning.
Hannes ErikssonChristos DimitrakakisPublished in: CoRR (2019)
Keyphrases
- np hard
- risk sensitive
- reinforcement learning
- model free
- optimal control
- markov decision processes
- markov decision problems
- control policies
- linear programming
- reinforcement learning algorithms
- optimal solution
- optimal policy
- function approximation
- utility function
- policy iteration
- state space
- average cost
- temporal difference
- learning algorithm
- action space
- markov decision chains
- machine learning
- dynamic programming
- decision problems
- partially observable
- control strategies
- reward function
- average reward
- monte carlo
- multi agent
- markov chain
- markov decision process
- control strategy
- fixed point
- radial basis function
- decision theoretic