Correct method is_sparse_paving
#39382
Conversation
|
Couldn't we just check that both M and M* (the dual of M) are paving - that's one of definitions in the literature. |
|
Whichever definition you use, please add a reference in the usual way, not here. |
|
Documentation preview for this PR (built with commit d894970; changes) is ready! 🎉 |
|
I agree with Dima's suggestion. Also, don't forget the |
Yes, we can. I tried to use an alternative characterization to avoid computation on the dual matroid (whose rank may be much larger). See the first example in the
Thanks. |
It was brought to my attention that the method `is_sparse_paving` is incorrect. See sagemath#36962 (comment). The method should only check the symmetric differences of `r`-element circuits rather than all (`r`- and `r+1`-element) circuits. The algorithm used is based on a somewhat unusual definition which can be found in https://arxiv.org/pdf/math/0404200. URL: sagemath#39382 Reported by: gmou3 Reviewer(s): Travis Scrimshaw
It was brought to my attention that the method
is_sparse_pavingis incorrect. See #36962 (comment).The method should only check the symmetric differences of
r-element circuits rather than all (r- andr+1-element) circuits.The algorithm used is based on a somewhat unusual definition which can be found in https://arxiv.org/pdf/math/0404200.