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!
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:
-
Combinatorial cuts:
- CglAllDifferent
- CglBKClique
- CglClique
- CglKnapsackCover
- CglOddHole
- CglOddWheel
- CglZeroHalf
-
Flow cover cuts:
-
Gomory cuts and variants:
- CglGomory
- CglGMI
- CglRedSplit
- CglRedSplit2
-
Lift-and-project cuts:
-
Mixed integer rounding cuts and variants:
-
Strengthening:
- CglCliqueStrengthening
- CglDuplicateRow
- CglPreprocess
- CglProbing
- CglSimpleRounding
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 https://github.com/coin-or/Cgl.
What follows is a quick start guide for obtaining or building Cgl on common platforms. More detailed information is available here.
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 https://repology.org/project/coin-or-cgl/versions 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
- arch:
-
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).
These quick start instructions assume you are in a bash shell.
To download and build Cgl from source, execute the following on the command line.
wget https://raw.githubusercontent.com/coin-or/coinbrew/master/coinbrew
chmod u+x coinbrew
./coinbrew fetch Cgl@master
./coinbrew build Cgl
For more detailed instructions on coinbrew, see https://coin-or.github.io/coinbrew.
The coinbrew
script will fetch the additional projects specified in the Dependencies section of config.yml.
- Download the source code, e.g., by cloning the git repo https://github.com/coin-or/Cgl
- 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
make test
make install
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.
- Code of Conduct
- COIN-OR Web Site
- COIN-OR general discussion forum
- Cgl Discussion forum
- Report a bug
- Doxygen generated documentation
A cut generator in Cgl
must conform to the following:
- Its main class
CglCutGeneratorDeriv
is derived from the classCglCutGenerator
. - It has three related classes used for data, parameters and information with respect to the enumeration tree:
- A class
CglDataDeriv
derived fromCglData
; it should contain pointers on all data used by the generator that might be obtained from anOsiSolverInterface
object when callinggenerateCuts()
with anOsiSolverInterface
object as parameter. The classCglDataDeriv
might beCglData
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 caseCglDataDeriv
might still contain a pointer on theOsiSolverInterface
object, but its use should be limited to obtaining the "expensive" information from the solver. - A class
CglParamDeriv
derived fromCglParam
. It should contain parameters of the generator that can be set by the user. The parameters in the classCglParamDeriv
must be taken into account during the cut generation. The classCglParamDeriv
might beCglParam
if the latter is sufficient. - A class
CglTreeInfoDeriv
derived fromCglTreeInfo
. The classCglTreeInfoDeriv
might beCglTreeInfo
if the latter is sufficient.
- A class
- 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)
- A member of type
- The data class
CglDataDeriv
must have methodsgetMember()
andsetMember()
for eachmember
of the class. Data members inCglData
irrelevant for a generator are completely ignored. If a data member that is used by a generator is not available whengenerateCuts(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 methodsgetMember()
andsetMember()
for eachmember
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 ofCglParamDeriv
, including those fromCglParam
. - 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.