Mirror Descent Linearized Augmented Lagrangian Methods for Nonconvex Constrained Stochastic Zeroth-Order Optimization
A framework of mirror descent linearized augmented Lagrangian methods that employs two-point stochastic zeroth-order gradient estimators and exploits non-Euclidean mirror descent geometry is proposed to solve nonconvex constrained stochastic zeroth-order optimization problems with exact constraints and stochastic objec...
Qian-Kun Shi, HanNuo Yuan, Xiao Wang et al.
· 1 citation