ＯＲ学会中部支部研究会を以下のように開催いたします． 皆さまのご参加をお待ちしております．
【2016/05/19更新：講演概要を掲載】
—
日時: 6月22日(水) 17:00-19:00
場所: 名古屋大学 東山キャンパス 情報科学研究科棟1階 第4講義室
http://www.co.cm.is.nagoya-u.ac.jp/~yagiura/access-j.html
講演者: Gregor Hendel
Zuse Institute Berlin, 研究員
http://www.zib.de/members/hendel
講演題目: How to solve Integer Optimization Problems with SCIP
講演者: Yuji Shinano
Zuse Institute Berlin, 研究員
http://www.zib.de/members/shinano
講演題目: Towards Using over a Million CPU Cores to Solve Previously
Unsolved Mixed Integer Programming Problem Instances
——————————————————-
講演概要:
——————————————————-
Gregor Hendel
How to solve Integer Optimization Problems with SCIP
The general-purpose branch-and-cut solver SCIP is one of the
fastest noncommercial tools for solving integer optimization
problems. Its plugin-based system facilitates custom extensions
for specific applications. The goal of this talk is to provide
users with a basic understanding of the SCIP solving process
and to illustrate some ways to tackle custom projects. In the
first part of this talk, I will give an overview of the solving
process of SCIP and the role of the different plugin types. I
will introduce the most important plugin types of SCIP with
special focus on its available primal heuristics. The second part
deals with the customization of SCIP at the example of the famous
Traveling Salesman Problem. Three alternative approaches to
formulate and solve TSP’s will be discussed and their particular
advantages will be highlighted: via the modeling language ZIMPL,
the C/C++ callable library of SCIP, and the new Python interface.
——————————————————-
Yuji Shinano
Towards Using over a Million CPU Cores to Solve Previously
Unsolved Mixed Integer Programming Problem Instances
The Ubiquity Generator (UG) is a framework for the external
parallelization of MIP solvers. UG provides a systematic way
to develop a parallel solver that can run on large-scale
distributed memory computing environments. It was used to develop
ParaSCIP, a distributed memory, massively parallel version of
the open source academic solver SCIP. In this talk we present
a success story where we solve 14 open MIP instances from
MIPLIB2003 and MIPLIB2010 using ParaSCIP on up to 80,000 cores of
supercomputers. Finally, we introduce ParaXpress, for which one
of the fastest commercial MIP solvers, the FICO Xpress-Optimizer,
has been parallelized by UG. Combining the internal shared-memory
parallelization of Xpress and the external parallelization of UG,
we aim at a new order of magnitude for supercomputer core-usage
in MIP solving.
——————————————————-
—
