We describe an algorithm for minimizing convex, not necessarily smooth, functions of several variables, based on a descent direction finding procedure that inherits some characteristics both of standard bundle method and of Wolfe’s conjugate subgradient method. This is obtained by allowing appropriate upward shifting of the affine approximations of the objective function which contribute to the classic definition of the cutting plane function. The algorithm embeds a proximity control strategy. Finite termination is proved at a point satisfying an approximate optimality condition and some numerical results are provided.

A method for convex minimization based on translated first-order approximations

GAUDIOSO, MANLIO;GORGONE, ENRICO
2017

Abstract

We describe an algorithm for minimizing convex, not necessarily smooth, functions of several variables, based on a descent direction finding procedure that inherits some characteristics both of standard bundle method and of Wolfe’s conjugate subgradient method. This is obtained by allowing appropriate upward shifting of the affine approximations of the objective function which contribute to the classic definition of the cutting plane function. The algorithm embeds a proximity control strategy. Finite termination is proved at a point satisfying an approximate optimality condition and some numerical results are provided.
Bundle methods; Convex optimization; Nonsmooth optimization; Applied mathematics
File in questo prodotto:
File Dimensione Formato  
reprint.pdf

non disponibili

Tipologia: versione editoriale
Dimensione 504 kB
Formato Adobe PDF
504 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11584/212504
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 5
  • ???jsp.display-item.citation.isi??? 4
social impact