On convergence of the stochastic subgradient method with on-line stepsize rules

Andrzej Ruszczyński, Wojciech Syski

Research output: Contribution to journalArticlepeer-review

7 Scopus citations

Abstract

A stochastic subgradient method for solving convex stochastic programming problems is considered. New on-line rules for determining stepsizes are derived from the concept of local regularized improvement functions. Convergence with probability 1 of the whole method is established.

Original languageEnglish (US)
Pages (from-to)512-527
Number of pages16
JournalJournal of Mathematical Analysis and Applications
Volume114
Issue number2
DOIs
StatePublished - Mar 1986
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Analysis
  • Applied Mathematics

Fingerprint Dive into the research topics of 'On convergence of the stochastic subgradient method with on-line stepsize rules'. Together they form a unique fingerprint.

Cite this