Effect of Reward Function Choices in Risk-Averse Reinforcement Learning.
Shuai MaJia Yuan YuPublished in: CoRR (2016)
Keyphrases
- reward function
- reinforcement learning
- risk averse
- reinforcement learning algorithms
- markov decision processes
- optimal policy
- risk neutral
- state space
- partially observable
- inverse reinforcement learning
- decision makers
- transition model
- stochastic programming
- function approximation
- utility function
- markov decision process
- multiple agents
- inventory level
- temporal difference
- portfolio management
- multi agent
- multistage
- graphical models
- control policies
- learning algorithm
- transition probabilities
- generative model
- action space
- control policy
- average cost
- expected utility
- markov decision problems
- state variables
- optimal control