Risk-sensitive planning in partially observable environments.
Janusz MareckiPradeep VarakanthamPublished in: AAMAS (2010)
Keyphrases
- risk sensitive
- partially observable environments
- partially observable
- markov decision processes
- markov decision problems
- optimal control
- reinforcement learning algorithms
- model free
- reinforcement learning
- decision theoretic
- state space
- utility function
- partially observable markov decision processes
- infinite horizon
- planning problems
- expected utility
- optimality criterion
- policy iteration
- average cost
- decision problems
- control policies
- optimal policy
- heuristic search
- finite state
- dynamical systems
- decision theory
- average reward
- decision makers
- learning algorithm
- temporal difference
- random walk
- planning domains
- markov decision process
- action space
- function approximation