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, and in 2010 by a larger initiative from academia and industry.

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. The library has now been released in its sixth edition.

MIPLIB 2017 – The 6th edition of the library

6 years after the release of MIPLIB 2010, an amazing progress of both solver and hardware technology has led to the former benchmark set being considered as solved. At the same time, new interesting and large MIP models have been developed in academia and industry, such that an update of the previous MIPLIB by new instances has been due.

MIPLIB 2017 is a collaborative effort by Arizona State University, COIN-OR, CPLEX, FICO, Gurobi, MathWorks, MIPCL, MOSEK, NUOPT, SAS, and Zuse Institute Berlin. To continue the diversity and quality standards of the previous editions, we collected interesting and challenging (mixed-)integer linear problems from all fields of Operations Research and Combinatorial Optimization, ideally built to model real-world problems.

For the first time, the compilation of the two instance sets of MIPLIB 2017 has been fully automated by solving a MIP model aiming at a diverse selection of as many different applications as possible, without overrepresenting any particular type of application or model.

Furthermore, we collected model files and supplementary data and made them available in addition to the raw instance format. Though an optional add-on, such model data provides researchers with richer information on the instances.

Last Update Nov 19, 2018 by Gregor Hendel
generated with R Markdown
© 2018 by Konrad-Zuse-Zentrum für Informationstechnik Berlin (ZIB)