'ROI' Access to 'MIPLIB' 2010 Benchmark Instances

The mixed integer programming library 'MIPLIB' (see < http://miplib.zib.de/>) is commonly used to compare the performance of mixed integer optimization solvers. This package provides functions to access 'MIPLIB' from the 'R' Optimization Infrastructure ('ROI'). More information about 'MIPLIB' can be found in the paper by Koch et al. available at < http://mpc.zib.de/index.php/MPC/article/viewFile/56/28>. The 'README.md' file illustrates how to use this package.

ROI.models.miplib

MIPLIB 2010

library(ROI)
library(ROI.models.miplib)

MIPLIB 2010 is a library of mixed integer programming problems which can be obtained from http://miplib.zib.de/. ROI.models.miplib provides utility functions to download MIPLIB 2010 and convert it into ROI optimization problems. Since MIPLIB 2010 is kind of big only functions for downloading and accessing the library are provided but not the library itself.

Download MIPLIB

There are two versions of MIPLIB available at http://miplib.zib.de/ a smaller version for benchmarking (around 94 MB) and the full version (around 1.3 GB) which contains all benchmark instances.

Download the benchmark instances.

miplib_download_benchmark(quiet = FALSE)

Download the all instances.

miplib_download_all(quiet = FALSE)

By default the optimization problems are stored in the package folder as .rds-files and can be accessed via the function miplib.

Download MIPLIB 2010 Metainfo

At http://miplib.zib.de/ also solutions to the problem instances are provided which can easily be obtained by executing.

miplib_download_metainfo()

Using MIPLI 2010

After downloading MIPLIB 2010 it can be accessed by the function miplib.

miplib()
## or if a non-default folder name was provided at the download step
miplib(folder = folder_name)

lists all the names of the (downloaded) MIPLIB 2010 instances.

miplib(x = "all")
## or if a non-default folder name was provided at the download step
miplib(x = "all", folder = folder_name)

gives all the (downloaded) MIPLIB instances.

miplib(x = "metainfo")
## or if a non-default folder name was provided at the download step
miplib(x = "metainfo", folder = folder_name)

gives the available meta information.

miplib(x = "rmine6")
## or if a non-default folder name was provided at the download step
miplib(x = "rmine6", folder = folder_name)

returns the rmine6 optimization problem as an ROI optimization problem.

ROI_solve( miplib(x = "iis.100.0.cov") )

References

• 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, Kati Wolter (2011).
  MIPLIB 2010
  URL http://mpc.zib.de/index.php/MPC/article/view/56/28

• R. E. Bixby, E. A. Boyd, R. R. Indovina (1992).
  MIPLIB: A Test Set of Mixed Integer Programming Problems