Extreme Risk Averse Policy for Goal-Directed Risk-Sensitive Markov Decision Process.
Valdinei FreireKarina Valdivia DelgadoPublished in: BRACIS (2016)
Keyphrases
- markov decision process
- goal directed
- risk sensitive
- risk neutral
- markov decision processes
- risk averse
- optimal policy
- utility function
- average cost
- infinite horizon
- state space
- finite horizon
- policy iteration
- optimal control
- control policies
- reinforcement learning
- markov decision problems
- finite state
- autonomous robots
- reward function
- transition probabilities
- partially observable
- dynamic programming
- decision makers
- average reward
- action space
- reinforcement learning algorithms
- stochastic programming
- decision processes
- initial state
- model free
- long run
- decision problems
- decision making
- expected utility
- machine learning
- portfolio management
- graphical models
- multi agent