> there is something in your code that I can't explain. > The first thing is that I added a loop counter (a double variable = 0 > and a loop++ command) and a xprintf to see the loops and the numerical > instability disappeared. Well, this seems very strange.
I found the bug that causes numerical instability. A long time ago Xypron fixed some my bug by adding a fragment to reoptimize lp, however, he changes column types directly in the glp_prob object, and this leads to inconsistency, because column statuses are not changed appropriately. Glp_simplex does not check that assuming that all information in glp_prob is correct. Will fix this bug in the next release. _______________________________________________ Bug-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-glpk
