Embracing Risk in Reinforcement Learning: The Connection between Risk-Sensitive Exponential and Distributionally Robust Criteria.
Erfaun NooraniJohn S. BarasPublished in: ACC (2022)
Keyphrases
- control strategy
- risk sensitive
- optimal control
- control strategies
- mathematical model
- reinforcement learning
- risk neutral
- infinite horizon
- robust optimization
- markov decision processes
- model free
- markov decision chains
- function approximation
- state space
- machine learning
- optimal policy
- reinforcement learning algorithms
- average cost
- policy iteration
- control policies
- multi agent