Risk-Sensitive Policy with Distributional Reinforcement Learning.
Thibaut ThéateDamien ErnstPublished in: CoRR (2022)
Keyphrases
- risk sensitive
- reinforcement learning
- control policies
- markov decision processes
- optimal policy
- markov decision problems
- model free
- optimal control
- policy iteration
- action space
- state space
- partially observable
- markov decision process
- control policy
- average cost
- infinite horizon
- reinforcement learning algorithms
- function approximation
- dynamic programming
- reward function
- average reward
- finite horizon
- action selection
- decision processes
- finite state
- utility function
- temporal difference
- control strategies
- function approximators
- multi agent
- long run
- decision problems
- multistage
- linear programming
- machine learning
- control system
- learning algorithm
- stochastic processes
- real valued
- transition probabilities
- bayesian networks