The QSopt code is based on floating-point arithmetic and the results are subject to rounding errors, as is the case with other commerical and academic linear-programming solvers. For the exact solution of LP problems, we have created the QSopt_ex rational solver and callable library. This package makes use of the GNU MP Bignum Library to deliver the exact rational solution to LP instances having rational input.
This code may be used under the terms of the GNU General Public License (Version 2.1 or later) as published by the Free Software Foundation. Alternatively, use is granted for research purposes only. It is your choice of which of these two licenses you are operating under.
The source code for Qsopt_ex 2.6 (090408) can be downloaded here, as a gzipped tar file. This is a simplified version of the full development code. It has been tested under Linux (various distributions), Mac OS 10.5, Cygwin, Solaris 9, and AIX.
The QSopt_ex research is supported supported by Office of Naval Research Grant N00014-03-1-0040 and National Science Foundation Grant CMMI-0726370.
The authors of QSopt_ex are David Applegate, William Cook, Sanjeeb Dash, and Daniel Espinoza.
22 Jul 2008: Added support for IBM's AIX version of Unix.
09 Sep 2008: Corrected bug in port of LU factorization.
08 Apr 2009: Corrected copywrite statement, modfied formating.