On 2013-12-06, kcrisman <kcris...@gmail.com> wrote: > In http://ask.sagemath.org/question/3282/ it was revealed that in Sage, the > MILP interface defaults (like most such software) to assuming nonnegative > variables as a constraint, which they has to be explicitly undone. It > looks like the (possibly unrelated) linear_program command also does this. linear_program() is an old interface to cvxopt. It should go away and be replaced by a cvxopt backend for "main" Sage LP. (something I meant to write for years, but never got to)
> This is natural in that world. However, Nathann did a little digging upon > my request and discovered that Matlab and Maple apparently (?) don't assume > this. the assumption x>=0 goes back to the old times when LPs were normally in the "standard" form max c^T x subject to Ax=b, x>=0. > > So... should we change our interface? Or should we keep it the same as it > was? I can think of arguments for both. Input would be appreciated, > especially from those who have actually used this functionality (perhaps in > contexts not "about" LP). I suppose one can add a parameter to the method creaing variables to LPs, and allowing one to specify nonnegativity. This, combined with the deprecation, should allow for an easy change - indeed the default assumption x>=0 is kind of weird for outsiders... Dima > > - kcrisman > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
