> We use CVXOPT to solve some convex optimization problems with GLPK and > itself LP solver. But we have found some solutions computed by GLPK are > incorrect. GLPK considers the problem is primal infeasible while CVXOPT > original LP solver can find a optimal solution. The detailed bug report > and corresponding test file are in the attachments. Looking forwards to > your reply. Thank you very much.
Thank you for your report. Unfortunately, I cannot reproduce the bug/error, because I don't have cvxopt. Please, post the lp instance in any format that glpsol understands or simply in a human-readable format. Most likely incorrect results are caused by tiny constraint coefficients (like 6.169139520277579e-17) that make the instance extremely badly scaled. Try to replace such tiny coefficients by exact zero, this may help. Andrew Makhorin _______________________________________________ Bug-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-glpk
