## Warnings computing number of vertices and facets of polytope

Questions and problems about using polymake go here.

Moderator: Moderators

AlfonsoC

### Warnings computing number of vertices and facets of polytope

Hello,

I defined a 12-dimensional bounded polytope in polymake via a list of inequalities, then I extracted its integer hull, and then I computed the product of the number of vertices and the number of facets. I obtain a numerical value, but before that, I get three warnings that I don't understand. My code and the warnings are below. What do the warnings mean? And more importantly, can I trust the numerical result? Thanks in advance!

Code:
$a= new Polytope(INEQUALITIES=>[ "some inequalities" ]);$b =$a-> LATTICE_POINTS;$b= new Polytope(POINTS=>$b);$n=($b-> N_VERTICES)*($b-> N_FACETS);
print \$n;

Warnings:
WARNING: rule latte.integer_points: N_LATTICE_POINTS : CONE_AMBIENT_DIM , VERTICES failed: can't parse output from LattE's 'count'
polymake: WARNING: rule latte.integer_points: N_LATTICE_POINTS : CONE_AMBIENT_DIM, CONE_DIM, FACETS | INEQUALITIES failed: can't parse output from LattE's 'count'
polymake: WARNING: rule latte.ehrhartpoly: LATTICE, EHRHART_POLYNOMIAL_COEFF : CONE_AMBIENT_DIM, FACETS | INEQUALITIES failed: could not parse output from latte at /usr/share/polymake/apps/polytope/rules/latte.rules line 231.

gawrilow
Main Author
Posts: 331
Joined: 25 Dec 2010, 17:40

### Re: Warnings computing number of vertices and facets of polytope

The reason for the warning could be either that the program 'count' is suddenly missing on your computer despite earlier successful configuration of latte.rules, or that some dependency of it is missing, in particular it needs some binaries from the cdd-test package. To find out the exact reason, you can start polymake with -dd option and repeat your input. It will show you what exactly it attempts to start and also preserve the temporary input files so that you can run 'count' manually and watch for its diagnostic. Finally, polymake will show you the eventually successful rule providing the number of lattice points, supposedly it will be a normaliz function. I reckon, you can trust its results.