Page 1 of 1

Error: Input Polytope must be bounded

Posted: 12 Jan 2018, 01:20
by UserCplex
The question got resolved and this post of mine is a bit tangential. Hence, I have chosen to delete it.

Thanks.

Re: Error: Input Polytope must be bounded

Posted: 12 Jan 2018, 10:58
by UserCplex
The question got resolved and this post of mine is a bit tangential. Hence, I have chosen to delete it.

Thanks.

Re: Error: Input Polytope must be bounded

Posted: 12 Jan 2018, 16:06
by joswig
Indeed, there is a stupid typo in our code where the parser reads the bounds. This will be fixed in the release of 3.2, which is due within the next two weeks. Thanks for pointing this out. As a temporary fix for now please convert your bounds to "honest" inequalities and equations.

However, your input polytope is of dimension 36, and it lives in 105 dimensions. Thus, most likely, you will not be able to enumerate all its lattice points. Neither with polymake nor with any other software. Recall that deciding whether or not a polytope (given in terms of inequalities) contains one lattice point is NP-complete. You can try, but you would be exceptionally lucky if it should work.

Re: Error: Input Polytope must be bounded

Posted: 12 Jan 2018, 16:18
by UserCplex
The question got resolved and this post of mine is a bit tangential. Hence, I have chosen to delete it.

Thanks.

Re: Error: Input Polytope must be bounded

Posted: 12 Jan 2018, 16:44
by gawrilow
how does polymake ensure that the LATTICE_POINTS enumerated are exhaustive enough?
There is no such term "enough" in polymake's philosophy. Either the property is complete and correct or it does not exist at all. So in your case if the problem turns out to be intractably large, it will just run forever or stop due to lack of memory, but it will never give you a "partially true" answer. This is the fundamental difference to commercial software aimed at solving a business problem, where a slightly suboptimal solution obtained in reasonable time is accepted for obvious reasons.