Mathematics > Optimization and Control
[Submitted on 31 May 2015 (v1), last revised 25 Dec 2016 (this version, v14)]
Title:Searching equillibriums in Beckmann's and Nesterov--de Palma's models
View PDFAbstract:In this paper we propose and develop classical Frank--Wolf algorithm for Beckmann's type models. This is not new, but we investigate details that allows us to speed up. We also consider stable dynamic like models. First model of this type was proposed 15 years ago by Yu. Nesterov and A. DePalma. We propose randomized dual averaging method with special (sum-type) randomization. For both of the problems we obtain the rates of convergences. It seems that this estimations to be unimprovable without additional assumption about problem formulation.
Submission history
From: Alexander Gasnikov [view email][v1] Sun, 31 May 2015 21:42:19 UTC (829 KB)
[v2] Tue, 2 Jun 2015 21:11:24 UTC (828 KB)
[v3] Thu, 9 Jul 2015 20:25:41 UTC (1,126 KB)
[v4] Fri, 7 Aug 2015 20:31:06 UTC (1,125 KB)
[v5] Mon, 19 Oct 2015 17:04:47 UTC (1,093 KB)
[v6] Sun, 25 Oct 2015 11:35:13 UTC (1,095 KB)
[v7] Thu, 12 Nov 2015 14:58:54 UTC (1,027 KB)
[v8] Tue, 12 Jan 2016 19:36:22 UTC (1,032 KB)
[v9] Thu, 4 Feb 2016 15:39:37 UTC (1,032 KB)
[v10] Mon, 4 Apr 2016 04:58:16 UTC (991 KB)
[v11] Thu, 12 May 2016 16:06:20 UTC (994 KB)
[v12] Thu, 18 Aug 2016 11:46:56 UTC (1,194 KB)
[v13] Sun, 18 Dec 2016 13:51:02 UTC (1,202 KB)
[v14] Sun, 25 Dec 2016 16:00:57 UTC (1,203 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.