Risk-sensitive safety analysis via state-space augmentation.
Margaret P. ChapmanMichael FaussH. Vincent PoorKevin M. SmithPublished in: CoRR (2021)
Keyphrases
- risk sensitive
- safety analysis
- state space
- markov decision processes
- markov decision problems
- safety critical
- control system
- nuclear power plant
- optimal control
- reinforcement learning
- dynamic programming
- optimal policy
- fault model
- symbolic model checking
- utility function
- partially observable
- action space
- markov chain
- reinforcement learning algorithms
- heuristic search
- finite state
- orders of magnitude
- control policies
- model free
- state variables
- planning problems
- dynamical systems
- average cost
- search space
- partial observability
- markov decision process
- optimality criterion
- infinite horizon
- reward function
- embedded systems
- particle filter
- expected utility
- linear programming
- planning under uncertainty
- belief space
- decision support system
- learning agent
- belief state
- transition probabilities
- decision theoretic
- agent architecture
- real time