A Globally Convergent Policy Gradient Method for Linear Quadratic Gaussian (LQG) Control.

Tomonori Sadamoto Fumiya Nakamata

Published in: CoRR (2023)

Keyphrases

linear quadratic
gradient method
optimal control
closed loop
vector valued
dynamical systems
gaussian model
control system
control strategy
optimization methods
convergence rate
dynamic programming
step size
reinforcement learning
multiscale
negative matrix factorization
mathematical model