About

UG is a generic framework to parallelize branch-and-bound based solvers (e.g., MIP, MINLP, ExactIP) in a distributed or shared memory computing environment.

• Exploits powerful performance of state-of-the-art "base solvers", such as SCIP, Xpress, etc.
• Without the need for base solver parallelization

Base solvers and communication libraries are abstracted within UG. A parallel solver instantiated by UG framework is named:

• ug[Base solver, Communication library]

For questions about UG please write to Yuji Shinano ug@zib.de For licensing questions, please see the license section of the web page and the contact provided there.

News

License

UG is distributed under the Gnu Lesser General License. You are allowed to retrieve and use UG. If you modify the source code yourself, you have to publish it under the LGPL along with a copy of it (contained in the source distribution). If you want to use UG commercially or if you are interested in maintenance and support, please contact us by sending an email to Thorsten Koch.

Due to the inclusion of third-party code, and the possibility to link against various dependencies, a build of UG can touch additional licenses. See here for more details.

Any publication for which UG is used must include an acknowledgment and a reference to the article:

Yuji Shinano, Tobias Achterberg, Timo Berthold, Stefan Heinz, and Thorsten Koch,
ParaSCIP – a parallel extension of SCIP,
Competence in High Performance Computing 2010, editors: C. Bischof, H.-G. Hegering, W. E. Nagel, and G. Wittum, pages 135–148, Springer, 2012.

Here is the corresponding bibtex entry.

Generally, for citations, please refer to the OPUS entry of the corresponding version or use the citation file.

• Version 1.0.0: CITATION.cff and DOI to OPUS entry.
• Version 0.9.1: CITATION.cff and DOI to OPUS entry.

Download

The latest version of ParaSCIP and FiberSCIP is included in the SCIP Optimization Suite. Please refer to the provided README file for help on installing the software.

Projects

Current projects are as follows:

ParaSCIP and FiberSCIP can handle both MIPs and MINLPs and are a parallel framework for solving mixed integer (linear and nonlinear) programs. They have deterministic mode for debugging.

Reference

Yuji Shinano, Tobias Achterberg, Timo Berthold, Stefan Heinz, Thorsten Koch, ParaSCIP -- a parallel extension of SCIP.
Yuji Shinano, Stefan Heinz, Stefan Vigerske, Michael Winkler, FiberSCIP -- A shared memory parallelization of SCIP.

Cooperation

• Norddeutscher Verbund für Hoch- und Höchstleistungsrechnen (HLRN)
• Oak Ridge National Laboratory
• FICO Xpress
• GAMS
• マス・フォア・インダストリ研究所　九州大学 (Institute of Mathematics for Industry, Kyushu University)
• 統計数理研究所 (The Institute of Statistical Mathematics)
• Oakland University
• Lawrence Livermore National Laboratory
• Georgia Institute of Technology

Imprint and privacy statement

© 2025 by Zuse Institute Berlin (ZIB). For the imprint and privacy statement we refer to the Imprint of ZIB.