Toward breaking the curse of dimensionality: an FPTAS for stochastic dynamic programs with multidimensional actions and scalar states.
Nir HalmanGiacomo NanniciniPublished in: CoRR (2018)
Keyphrases
- stochastic dynamic
- state transitions
- initial state
- perceptual aliasing
- action sequences
- partial knowledge
- state transition
- multi dimensional
- multidimensional data
- internal states
- state action
- multidimensional space
- vector valued
- plan recognition
- approximation algorithms
- situation calculus
- human actions
- state space
- state information
- external world
- evolutionary algorithm
- multidimensional signals
- goal state
- reasoning about actions
- action selection
- computer programs
- decision theoretic
- special case