Bundle Methods for Non-smooth Convex Optimization over Simple Domains

Wim Stefanus van Ackooij () and Welington Luis de Oliveira ()
Additional contact information
Wim Stefanus van Ackooij: Électricité de France (EDF R&D)
Welington Luis de Oliveira: Mines Paris - PSL

Chapter Chapter 11 in Methods of Nonsmooth Optimization in Stochastic Programming, 2025, pp 321-374 from Springer

Abstract: Abstract Once the main ideas on cutting-plane methods have been presented in the previous chapter, here we go further to present state-of-the-art algorithms for minimizing a non-smooth convex function f over a “simple” feasible set X ⊂ ℝ n $$X \subset \mathbb {R}^n$$ . This chapter goes deep into the theory of bundle methods, shedding light on the intuition behind each variant, their specificities, strengths, and weaknesses. For didactic purposes, we provide abridged versions of the considered bundle methods before presenting their more efficient forms capable of exploiting the possible additive structure of the objective function and even inexact oracles. As in the previous chapter, f may only be assessed by a first-order (inexact) oracle.

Keywords: Non-smooth optimization; Proximal bundle method; Level bundle method; Trust-region bundle method (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:isochp:978-3-031-84837-7_11

Ordering information: This item can be ordered from
http://www.springer.com/9783031848377

DOI: 10.1007/978-3-031-84837-7_11

Access Statistics for this chapter

More chapters in International Series in Operations Research & Management Science from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().