Hypersimplex Bug

Questions and problems about using polymake go here.
charles.wang
Posts: 2
Joined: 03 Jul 2016, 00:42

Hypersimplex Bug

Postby charles.wang » 03 Jul 2016, 01:21

I've noticed that the commands hypersimplex(2,3) and hypersimplex(3,4) produce strange results. I expect hypersimplex(n-1,n) to give (n-1)-simplices, but get the following:

Code: Select all

polytope> print hypersimplex(2,3)->F_VECTOR; 3 6 polytope> print hypersimplex(3,4)->F_VECTOR; 4 10 8

I get similar (obviously incorrect) results for hypersimplex(1,3) and hypersimplex(1,4).

Can someone take a look at this? (I am using polymake version 3.0.)

Thanks!

User avatar
joswig
Main Author
Posts: 280
Joined: 24 Dec 2010, 11:10

Re: Hypersimplex Bug

Postby joswig » 04 Jul 2016, 11:10

Thanks for pointing this out. Indeed, this is a bug.

The reason is that (for convenience and speed) the client writes VERTICES and FACETS, but they do not fit if the hypersimplex degenerates to a simplex. In fact, a worse version of the same problem occurs for k=0 or k=d.

The issue is resolved in polymake's git repository, and so it will be correct in the next release. We can provide a patch for 3.0 if necessary.
The solution is (1) to not write FACETS and VERTICES_IN_FACETS in the simplex case, (2) to forbid parameters which lead to the degeneration to a point.

charles.wang
Posts: 2
Joined: 03 Jul 2016, 00:42

Re: Hypersimplex Bug

Postby charles.wang » 05 Jul 2016, 21:16

Okay, thanks! I won't need a patch for this -- I was just using these for some testing.


Return to “Helpdesk”