Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] Re: Bug report: Polytope_distance_d package

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] Re: Bug report: Polytope_distance_d package


Chronological Thread 
  • From: Bernd Gärtner <>
  • To: <>
  • Subject: Re: [cgal-discuss] Re: Bug report: Polytope_distance_d package
  • Date: Tue, 13 Aug 2013 15:58:04 +0200

On 8/9/13 11:44 AM, jdumas wrote:
Hi,

Sorry to bump this thread again after a thousand years, but after reading
the doc and having had a glance at the source code, it seems that there's
currently no way to set the pricing stategy for the solver used by the
Polytope_distance_d class. It falls back on either QP_PARTIAL_DANTZIG or
QP_PARTIAL_FILTERED_DANTZIG.
You're right, I had overlooked this.


Could you provide a patch to pass Quadratic_program_options to the solver?
I'm working on a fix for the next release. For now, simply replace QP_PARTIAL_DANTZIG by QP_BLAND in the file Polytope_distance_d.h. If you're calling the package as in the example file polytope_distance_d.cpp (full exact computations), this should solve the issue. If you call it as in polytope_distance_d_fast_exact.cpp (filtered computations), you should switch to full exact computations.

Best,
Bernd.




Archive powered by MHonArc 2.6.18.

Top of Page