MIPLIB - Mixed Integer Problem Library

In response to the needs of researchers for access to real-world mixed integer programs a group of researchers Robert E. Bixby, E.A. Boyd, and R.R. Indovina created in 1992 the MIPLIB, an electronically available library of both pure and mixed integer programs. This was updated in 1996 by Robert E. Bixby, Sebastian Ceria, Cassandra M. McZeal, and Martin W.P. Savelsbergh. The library was updated again in 2003 by Alexander Martin, Tobias Achterberg, and Thorsten Koch.

Since its introduction, MIPLIB has become a standard test set used to compare the performance of mixed integer optimizers. Its availability has provided an important stimulus for researchers in this very active area.

MIPLIB 2010

Again, more than 7 years have past since the last update of the MIPLIB. The MIPLIB 2010 is the first MIPLIB release that has been assembled by a large group from academia and from industry, all of whom work in integer programming. There was mutual consent that the concept of the library had to be expanded in order to fulfill the needs of the community. The new version comprises 361 instances sorted into several groups. This includes the main benchmark test set of 87 instances, which are all solvable by today's codes, and also the challenge test set with 164 instances, many of which are currently unsolved. For the first time, we include scripts to run automated tests in a predefined way. Further, there is a solution checker to test the accuracy of provided solutions using exact arithmetic.

The people involved in the creation of MIPLIB 2010 are: Thorsten Koch, Tobias Achterberg, Erling Andersen, Oliver Bastert, Timo Berthold, Robert E. Bixby, Emilie Danna, Gerald Gamrath, Ambros M. Gleixner, Stefan Heinz, Andrea Lodi, Hans Mittelmann, Ted Ralphs, Domenico Salvagnin, Daniel E. Steffy, and Kati Wolter.

A reference to this library (in bibtex format), can be found here: KochEtAl2011

Detailed information

Download

With MIPLIB test script (recommended):

benchmark instances: tgz (94 MB), zip (94 MB)
all instances: tgz (1.4 GB), zip (1.4 GB)
test script only: tgz (34 KB), zip (50 KB)

Without MIPLIB test script:

benchmark instances: tgz (94 MB), zip (94 MB)
all instances: tgz (1.4 GB), zip (1.4 GB)
unchanged contributed instances

Older versions of the scripts can be found here, unchanged release versions of the instances that were corrected after the release can be found here.

News

(a complete and more detailed changelog can be found here)

Mar 2013	queens-30 moved from open to hard.
Mar 2013	Infeasibility of neos-807456 proven, moved from open to hard.
Jan 2013	bg512142 solved, moved from open to hard.
Nov 2012	shs1023 solved, moved from open to hard.
Nov 2012	rail03 solved, moved from open to hard.
Okt 2012	pigeon-13 solved, moved from open to hard.
Aug 2012	bnatt400 moved from hard to easy.
Aug 2012	germany50-DBM solved, moved from open to hard.
Jul 2012	set3-15, dc1c, berlin_5_8_0 solved, moved from open to hard.
Jun 2012	hanoi5 and dg012142 solved, moved from open to hard.
Jun 2012	Solution files added for easy and hard instances.
May 2012	g200x740i, p100x588b, ns1696083, vpphard2, and uc-case3 solved, moved from open to hard, maxgasflow and uct-subprob moved from hard to easy
Apr 2012	neos-631710 solved, moved from open to easy, blp-ic97 moved from hard to easy
Mar 2012	r80x800 solved, moved from open to hard.
Mar 2012	First feasible solution found for ns1663818, infeasibility of lrsa120 proven.
Feb 2012	First feasible solution found for in.
Feb 2012	Added information about the feasibility of instances and new lists of feasible instances, infeasible instances, and instances with unknown feasibility.
Feb 2012	buildingenergy solved, moved from open to easy, toll-like solved, moved from open to hard
Jan 2012	11 open instances solved, b2c1s1, maxgasflow, opm2-z10-s2, opm2-z11-s8, opm2-z12-s14, opm2-z12-s7, rmatr200-p10, satellites3-40-fs, satellites3-40, and wnq-n100-mw99-14 moved to hard; transportmoment moved to easy.
Jan 2012	corrected some instances to comply with the mps-format, version to 1.0.3 of the script released: added support for lp_solve and GLPK
Aug 2011	seymour-disj-10 solved, moved from open to hard
Jul 2011	Change of solution statuses: 6 instanced moved from hard to easy
Jul 2011	d10200 and rococoC11-011100 solved, moved from open to hard
Jun 2011	Change of solution statuses: blp-ic97 and atm20-100 moved from open to hard, 11 instances moved from hard to easy.
Jun 2011	ns2122603 turned out to be feasible, removed from infeasible test set.
Jun 2011	Version 1.0.2 of MIPLIB 2010 scripts released: adapted for Gurobi 4.5.0 and higher
May 2011	Version to 1.0.1 of the script released: added compare result scripts which can be called via the Makefile, e.g., make cmpres SOLVER="*" TEST=benchmark.
May 2011	MIPLIB 2010 final version released.
Feb 2011	MIPLIB 2010 beta version released.
Oct 2010	Submission closed.
Mar 2010	Call for submissions.

Keeping it up-to-date

The current maintainers of MIPLIB are Timo Berthold (berthold Klammeraffe zib.de), Gerald Gamrath (gamrathzib.de), Dan Steffy (steffyzib.de), and Thorsten Koch (kochzib.de).

If you find optimal solutions to formerly unsolved problems, please notify us.
We are still short on references to some of the instances. If you know an article, which has results, insights or history for one of the instances, please send us the reference, so we could include it.

Solvability of the instances

MIPLIB 2010: number of instances classified as easy, hard, and not solved

Number of solved MIPLIB 2010 instances that currently classify as easy, hard, and not solved.
'Easy' means, that the instance could be solved within one hour using a commercial MIP-solver, 'hard' stands for instances, that were solved, but not in the previous conditions.

Older versions of MIPLIB and related links

Disclaimer

Most of the data files on this site have been converted, some optimal solutions are gathered from papers and some data was typed in by hand or generated by automatic solution extraction programs. While we took every effort to make no mistakes, we can not guarantee that everything is correct. If you find any errors or have doubts about a solution, please contact us.