13
$\begingroup$

Given a set of integer partitions $\{\lambda_1, \lambda_2,\dots \lambda_n\}$. Are there combinatorial criteria for deciding whether the associated Schur polynomials $s_{\lambda_1}, s_{\lambda_2},\dots s_{\lambda_n}$ are algebraically independent?

If this is too hard, are there any surprising necessary or sufficient criteria?

$\endgroup$
3
  • 1
    $\begingroup$ Are you interested in algebraic relations among Schur functions like: doi.org/10.37236/1560 (see Theorem 2)? So, this gives a pattern needed to be avoided for algebraic independence. If useful I could make an answer outlining the main theorem so the paper doesn't have to opened to see what's inside. $\endgroup$ Commented Dec 16, 2020 at 21:19
  • $\begingroup$ @JohnMachacek That's an interesting relation. I wonder if it is the case that all algebraic relations are built out of such identities (and Giambelli type formulas etc.). I think it's useful to have an answer with partial progress of the form "patterns to avoid". $\endgroup$ Commented Dec 16, 2020 at 22:04
  • 1
    $\begingroup$ Somewhat related blog post: realopacblog.wordpress.com/2019/09/22/… $\endgroup$ Commented Dec 17, 2020 at 0:07

1 Answer 1

4
$\begingroup$

I'll add further details to my comment and make it an answer. This will give algebraic relations between certain collections of six Schur polynomials. Hence, for algebraic independence it will give a necessary condition of avoiding these six Schur polynomials together. This relation comes from Bijective proofs for Schur function identities which imply Dodgson's condensation formula and Plücker relations by Fulmek and Kleber (see Theorem 2 of this paper). I will change notation from the linked paper to agree with the notation in the question. Take $a_1 \geq a_2 \geq \cdots \geq a_{r+1}$ if we let $$\lambda_1 = (a_1, a_2, \dots, a_r)$$ $$\lambda_2 = (a_2, a_2, \dots, a_{r+1})$$ $$\lambda_3 = (a_2, a_3, \dots, a_r)$$ $$\lambda_4 = (a_1, a_2, \dots, a_{r+1})$$ $$\lambda_3 = (a_2-1, a_3-1, \dots, a_{r+1}-1)$$ $$\lambda_4 = (a_1+1, a_2+1, \dots, a_r+1)$$ then $$s_{\lambda_1}s_{\lambda_2} = s_{\lambda_3}s_{\lambda_4} + s_{\lambda_5}s_{\lambda_6}.$$ Letting $\lambda_1 = \lambda_2 = \cdots = \lambda_{r+1} = c$ we find that $\lambda_1 = \lambda_2 = (c)^r$, $\lambda_3 = (c)^{r-1}$, $\lambda_4 = (c)^{r+1}$, $\lambda_5 = (c-1)^r$, and $\lambda_6 = (c+1)^r$ which are all rectangle shapes recovering a result of Kirillov (which is cited in the linked paper).

$\endgroup$
2
  • 2
    $\begingroup$ See also "Planar flows and quadratic relations over semirings" by Danilov, Karzanov, and Koshevoy: arxiv.org/abs/1102.2578. $\endgroup$ Commented Dec 17, 2020 at 21:18
  • 2
    $\begingroup$ And a somewhat related conjecture of Lam, Postnikov, and Pylyavskyy is discussed here: arxiv.org/abs/math/0608134. $\endgroup$ Commented Dec 17, 2020 at 21:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .