11
$\begingroup$

We have two conjectured generalizations of the question asked at a Hankel matrix of involution numbers by Tewodros Amdeberhan. Let $n!!=1!\,2!\cdots n!$.

Conjecture 1. Let $I_k$ denote the number of involutions in the symmetric group $\mathfrak{S}_k$. Then the Smith normal form of the matrix $[I_{i+j}]_{i,j=0}^n$ has diagonal entries $0!, 1!, \dots, n!$,

Conjecture 2. Let $J_k=k!\sum_{\lambda\vdash k}s_\lambda$, where $s_\lambda$ is a Schur function. Set $J_0=1$. When the symmetric function $\det[J_{i+j}]_{i,j=0}^n$ is expanded in terms of power sums, then every coefficient is an integer divisible by $n!!$. (Tewodros' question is equivalent to the coefficient of $p_1^{n(n+1)}$ being equal to $n!!$.)

Is there a nice formula or combinatorial interpretation of the coefficients in Conjecture 2?

$\endgroup$
1
  • $\begingroup$ This is a really cool extension of the discussion on involution. Glad to see it. $\endgroup$ Feb 8, 2017 at 17:28

1 Answer 1

3
$\begingroup$

On Conjecture 1: As remarked by Johann Cigler in the linked question (and shown in the references linked there) the matrix $M:=\left[I_{i+j}\right]_{i\ge0\atop j\ge0}$ diagonalises as $M=U^TDU$ with $D:=\operatorname{diag(k!)}$, and $U$ an upper triangular integer coefficients matrix with unit diagonal elements, hence with integer coefficient inverse: so this is also the SNF of $M$.

For example:

$$ \left[ \begin {array}{ccccc} 1&1&2&4&10\\ 1&2&4&10& 26\\ 2&4&10&26&76\\ 4&10&26&76&232 \\ 10&26&76&232&764\end {array} \right] =$$$$= \left[ \begin {array}{ccccc} 1&&&&\\ 1&1&&& \\ 2&2&1&&\\ 4&6&3&1& \\ 10&16&12&4&1\end {array} \right] \left[ \begin {array}{ccccc} 1& & & & \\ &1& & & \\ & &2& & \\ & & &6& \\ & & & &24\end {array} \right] \left[ \begin {array}{ccccc} 1&1&2&4&10\\ &1&2&6& 16\\ & &1&3&12\\ & & &1&4 \\ & & & &1\end {array} \right] $$

$\endgroup$
5
  • $\begingroup$ This diagonalisation of $M$ is precisely the same as in my answer to the linked Amdeberhan’s question. $\endgroup$ Feb 9, 2017 at 15:57
  • $\begingroup$ (Actually I can't see the diagonalisation there, but yes, there is a uniqueness result :) ) $\endgroup$ Feb 9, 2017 at 16:43
  • 1
    $\begingroup$ $\sum_k\binom{n}{k}I_{n-k} \binom{m}{k}I_{m-k} k!=I_{m+n} $ is equivalent with the fact that if $A$ is the matrix with entries $\binom{i}{j}I_{i-j}$ and $D$ is the diagonal matrix with elements $i!$ then $M=ADA^T.$ $\endgroup$ Feb 9, 2017 at 16:59
  • $\begingroup$ The identity Johann Cigler mentioned is equation (2.7) in his link: arxiv.org/pdf/1406.2356.pdf $\endgroup$ Feb 9, 2017 at 18:34
  • $\begingroup$ ok thank you guys $\endgroup$ Feb 9, 2017 at 19:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.