Important Notes

This scipoptsuite version implements the backend for my other project at https://github.com/avrech/learning2cut.git.

Installation

• add export SCIPOPTDIR=/path/to/installation/dir to your .bashrc
• go to scipoptsuite-6.0.2-avrech and run the following commands in the terminal

$ cmake -Bbuild -H. -DCMAKE_BUILD_TYPE=Debug -DCMAKE_INSTALL_PREFIX=$SCIPOPTDIR
$ cmake --build build
$ cd build
$ make install

Added Functionality

The following files in scip/src/scip/ were modified to enable cut selection control via PySCIPOpt:

• scip.h, pub_lp.h, struct_sepastore.h, sepastore.h
• solve.c, sepastore.c, lp.c

The following functionality was added:

• get the recently selected cut names
• force selecting a specified group of cuts
• reset maxcuts and maxcutsroot at the beginning of each separation round, to avoid undesired enoughcuts case.

In order to use this functionality, please clone PySCIPOpt from https://github.com/ds4dm/PySCIPOpt.git and install the ml-cutting-planes branch inside your virtualenv.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
*                                                                           *
*                  This file is part of the program and library             *
*         SCIP --- Solving Constraint Integer Programs                      *
*                                                                           *
*    Copyright (C) 2002-2019 Konrad-Zuse-Zentrum                            *
*                            fuer Informationstechnik Berlin                *
*                                                                           *
*  SCIP is distributed under the terms of the ZIB Academic License.         *
*                                                                           *
*  You should have received a copy of the ZIB Academic License              *
*  along with SCIP; see the file COPYING. If not email to scip@zib.de.      *
*                                                                           *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

The SCIP Optimization Suite consists of the following five software tools:

1. ZIMPL - the Zuse Institute Mathematical Programming language
2. SoPlex - the Sequential Object Oriented Simplex
3. SCIP - Solving Constraint Integer Programs
4. GCG - Generic Column Generation
5. UG - Ubiquity Generator Framework

We provide two different systems to compile the code: the traditional Makefile system and the new CMake build system. Be aware that generated libraries and binaries of both systems might be different and incompatible. For further details please refer to the INSTALL file of SCIP and the online documentation.

Content

1. CMake

2. Makefiles

• Creating a SCIP binary and the individual libraries
• Creating a single library containing SCIP, SoPlex, and ZIMPL
• Creating GCG and UG

CMake

Ensure that you're using an up-to-date CMake installation. CMake will automatically configure the code according to your environment and the available third-party tools, like GMP, etc. Create a new directory for the build, for instance inside the scipoptsuite directory.

mkdir build
cd build
cmake ..
make

For further information please refer to the online documentation of SCIP or the INSTALL in the SCIP subdirectory.

Makefiles

Creating a SCIP binary and the individual libraries

SCIP uses the libraries of ZIMPL and SoPlex to be able to read ZIMPL models and solve the subproblem LP relaxations with SoPlex.

In order to compile the whole bundle, just enter

make

within the SCIP Optimization Suite main directory. If you are using a gcc compiler with version less 4.2 you have to compile with the following additional flags

make LPSOPT=opt-gccold ZIMPLOPT=opt-gccold OPT=opt-gccold

If all goes well, you should get a final message

** Build complete. ** Find your binary in "/scipoptsuite-/scip-/bin". ** Enter "make test" to solve a number of easy instances in order to verify that SCIP runs correctly.

If this is not the case, there are most probably some libraries missing on your machine or they are not in the right version. In its default build, the SCIP Optimization Suite needs the following external libraries:

• the Z Compression Library (ZLIB: libz.a or libz.so on Unix systems) Lets you read in .gz compressed data files.
• the GNU Multi Precision Library (GMP: libgmp.a or libgmp.so on Unix systems) Allows ZIMPL to perform calculations in exact arithmetic.
• the Readline Library (READLINE: libreadline.a or libreadline.so on Unix systems) Enables cursor keys and file name completion in the SCIP shell.

You can disable those packets using the following make arguments:

• GMP=false (disables GMP support)
• ZLIB=false (disables ZLIB support)
• READLINE=false (disables READLINE support)

You can also disable ZIMPL by specifying ZIMPL=false as a make argument. Note, however, that this disables the ZIMPL file reader in SCIP and you can no longer read in ZIMPL models (the input files with a .zpl extension).

Since ZIMPL requires GMP, it is automatically disabled, if GMP is disabled.

Since the GMP is not installed on every machine, and the READLINE library is sometimes only existing in an old version, these are the two most frequent candidates for build problems. The following should work on most machines:

make GMP=false READLINE=false

Note that on some MAC systems, GMP is installed under /sw/include and /sw/lib. If these are not contained in the library and include paths, you have to add them explicitly.

If this still does not work, you should try the following, which is the most compatible method to build the SCIP Optimization Suite:

make ZLIB=false GMP=false READLINE=false LPSOPT=opt-gccold OPT=opt-gccold

Note, however, that in this case, you cannot

• read compressed (.gz) input files from SCIP and SoPlex
• read ZIMPL models from SCIP
• use the counting of solutions feature exactly for large numbers of solutions
• use the ZIMPL binary to transform ZIMPL (.zpl) models into MIP instances of .lp or .mps type
• use the cursor keys and command line completion in SCIP

For more information how to install the components of the SCIP Optimization Suite and for details concerning special architectures and operating systems, see the INSTALL file of SCIP and the documentation of each component.

Creating a single library containing SCIP, SoPlex, and ZIMPL

In case you need a single library which contains the whole package, that is SCIP, SoPlex, and ZIMPL, you can use the command:

make scipoptlib

This will create a single library containing SCIP, SoPlex, and ZIMPL. It is placed in the lib directory. This library is statically linked. If you need a shared library use the command:

make scipoptlib SHARED=true

Creating GCG and UG

In case for the both SCIP extensions GCG and UG, you can easily compile these using the commands

make gcg
make ug