Proposes a Riemannian quasi-Newton bundle method with subgradient aggregation and limited-memory updates for nonsmooth optimization, proving global convergence to stationary points under a new semismoothness assumption.
C omputa- tional Optimization and Applications 72(2), 391–412 (2019)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A restricted memory quasi-Newton bundle method for nonsmooth optimization on Riemannian manifolds
Proposes a Riemannian quasi-Newton bundle method with subgradient aggregation and limited-memory updates for nonsmooth optimization, proving global convergence to stationary points under a new semismoothness assumption.