Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


A COIN-OR Project

Projects such as this one are maintained by a small group of volunteers under the auspices of the non-profit COIN-OR Foundation and we need your help! Please consider sponsoring our activities or volunteering to help!

Latest Release

This file is auto-generated from config.yml using the generate_readme script. To make changes, please edit config.yml or the generation scripts here and here.

The COIN-OR Cut Generation Library (Cgl) is a collection of cut generators that can be used with other COIN-OR packages that make use of cuts, such as, among others, the linear solver Clp or the mixed integer linear programming solvers Cbc or BCP. Cgl uses the abstract class OsiSolverInterface (see Osi) to use or communicate with a solver. It does not directly call a solver.

Each cut generator is in a separate directory with its own maintainer. All generators are combined in one library when Cgl is compiled.

Available cut generators are:

CoinUtils is an open-source collection of classes and helper functions that are generally useful to multiple COIN-OR projects. These utilities include:

  • classes for storing and manipulating sparse matrices and vectors,
  • performing matrix factorization,
  • parsing input files in standard formats, e.g. MPS,
  • building representations of mathematical programs,
  • performing simple presolve operations,
  • warm starting algorithms for mathematical programs,
  • comparing floating point numbers with a tolerance
  • classes for storing and manipulating conflict graphs, and
  • classes for searching and storing cliques and odd cycles in conflict graphs, among others.

The project managers of Cgl are Robin Lougee (@rlougee) and Francois Margot.

Cgl is written in C++ and is released as open source under the Eclipse Public License 2.0.

It is distributed under the auspices of the COIN-OR Foundation.

The Cgl development site is


Code: DOI


Windows Builds

Linux and MacOS Builds


What follows is a quick start guide for obtaining or building Cgl on common platforms. More detailed information is available here.

Docker image

There is a Docker image that provides Cgl, as well as other projects in the COIN-OR Optimization Suite here


For newer releases, binaries will be made available as assets attached to releases in Github here. Older binaries are archived as part of Cbc here.

  • Linux (see for a complete listing):

    • arch:
      $ sudo pacman -S  coin-or-cgl
    • Debian/Ubuntu:
      $ sudo apt-get install  coinor-cgl coinor-libcgl-dev
    • Fedora/Redhat/CentOS:
      $ sudo yum install  coin-or-Cgl coin-or-Cgl-devel
    • freebsd:
      $ sudo pkg install math/cgl
    • linuxbrew:
      $ brew install cgl
  • Windows: The easiest way to get Cgl on Windows is to download an archive as described above.

  • Mac OS X: The easiest way to get Cgl on Mac OS X is through Homebrew.

    $ brew tap coin-or-tools/coinor
    $ brew install coin-or-tools/coinor/cgl
  • conda (cross-platform, no Windows for now):

    $ conda install coin-or-cgl

Due to license incompatibilities, pre-compiled binaries lack some functionality. If binaries are not available for your platform for the latest version and you would like to request them to be built and posted, feel free to let us know on the mailing list.


Source code can be obtained either by

  • Downloading a snapshot of the source code for the latest release version of Cgl from the releases page,
  • Cloning this repository from Github, or
  • Using the coinbrew script to get the project and all dependencies (recommended, see below).


Cgl has a number of dependencies, which are detailed in config.yml. Dependencies on other COIN-OR projects are automatically downloaded when obtaining the source with coinbrew. For some of the remaining third-party dependencies, automatic download scripts and build wrappers are provided (and will also be automatically run for required and recommended dependencies), while other libraries that are aeasy to obtain must be installed using an appropriate package manager (or may come with your OS by default).

BUILDING from source

These quick start instructions assume you are in a bash shell.

Using coinbrew

To download and build Cgl from source, execute the following on the command line.

chmod u+x coinbrew
./coinbrew fetch Cgl@master
./coinbrew build Cgl

For more detailed instructions on coinbrew, see The coinbrew script will fetch the additional projects specified in the Dependencies section of config.yml.

Without coinbrew (Expert users)

  • Download the source code, e.g., by cloning the git repo
  • Download and install the source code for the dependencies listed in config.yml
  • Build the code as follows (make sure to set PKG_CONFIG_PTH to install directory for dependencies).
./configure -C
make test
make install

Doxygen Documentation

If you have Doxygen available, you can build a HTML documentation by typing

make doxygen-docs

in the build directory. If Cgl was built via coinbrew, then the build directory will be ./build/Cgl/master by default. The doxygen documentation main file is found at <build-dir>/doxydoc/html/index.html.

If you don't have doxygen installed locally, you can use also find the documentation here.

Project Links

Information for Subproject Managers

A cut generator in Cgl must conform to the following:

  • Its main class CglCutGeneratorDeriv is derived from the class CglCutGenerator.
  • It has three related classes used for data, parameters and information with respect to the enumeration tree:
    • A class CglDataDeriv derived from CglData; it should contain pointers on all data used by the generator that might be obtained from an OsiSolverInterface object when calling generateCuts() with an OsiSolverInterface object as parameter. The class CglDataDeriv might be CglData if the latter is sufficient. An exception is made for generators needing information deemed too expensive to collect from the solver (for example the optimal Simplex tableau); in this case CglDataDeriv might still contain a pointer on the OsiSolverInterface object, but its use should be limited to obtaining the "expensive" information from the solver.
    • A class CglParamDeriv derived from CglParam. It should contain parameters of the generator that can be set by the user. The parameters in the class CglParamDeriv must be taken into account during the cut generation. The class CglParamDeriv might be CglParam if the latter is sufficient.
    • A class CglTreeInfoDeriv derived from CglTreeInfo. The class CglTreeInfoDeriv might be CglTreeInfo if the latter is sufficient.
  • The class CglCutGeneratorDeriv must have
    • A member of type CglParamDeriv used to store the current parameters.
    • A method getParam() that returns the object storing the current parameters.
    • A method generateCuts(const OsiSolverInterface & si, OsiCuts & cs, const CglTreeInfoDeriv info)
    • A method generateCuts(const CglDataDeriv &data, OsiCuts & cs, const CglTreeInfoDeriv info)
  • The data class CglDataDeriv must have methods getMember() and setMember() for each member of the class. Data members in CglData irrelevant for a generator are completely ignored. If a data member that is used by a generator is not available when generateCuts(const CglDataDeriv &data, OsiCuts & cs, const CglTreeInfoDeriv info) is called, the call is aborted, as if no cuts were found. A warning message might be printed.
  • The class CglParamDeriv must have methods getMember() and setMember() for each member of the class. All parameters must have default values. Each cut generator with a derived class is free to change the default values for all the members of CglParamDeriv, including those from CglParam.
  • Once an object of the cut generator class is created, it should be possible to call generateCuts() several times in a row without having to destroy and re-create the object.
  • By default, a successful call to generateCuts() should not generate any output. If an error occurs, a message might be printed.