Where academic tradition
meets the exciting future

Diagonal Bundle Method for Nonsmooth Sparse Optimization

Napsu Karmitsa, Diagonal Bundle Method for Nonsmooth Sparse Optimization. TUCS Technical Reports 1116, TUCS, 2014.

Abstract:

We propose an efficient diagonal bundle method D-Bundle for sparse nonsmooth, possibly nonconvex optimization. The convergence of the proposed method is proved for locally Lipschitz continuous functions that are not necessary differentiable nor convex. The numerical experiments have been made using problems with up to million variables. The results to be presented confirm the usability of the D-Bundle especially for extremely large-scale problems.

Files:

Full publication in PDF-format

BibTeX entry:

@TECHREPORT{tKarmitsa_Napsu14a,
  title = {Diagonal Bundle Method for Nonsmooth Sparse Optimization},
  author = {Karmitsa, Napsu},
  number = {1116},
  series = {TUCS Technical Reports},
  publisher = {TUCS},
  year = {2014},
  keywords = {Nondifferentiable optimization, sparse problems, bundle methods, diagonal variable metric methods},
  ISBN = {978-952-12-3086-8},
}

Belongs to TUCS Research Unit(s): Turku Optimization Group (TOpGroup)

Edit publication