Exceptional Hermite Polynomials and Calogero-Moser Pairs

Luke Paluso, Alex Kasman · 2025

Read Exceptional Hermite Polynomials and Calogero-Moser Pairs by Luke Paluso, Alex Kasman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-21935.txt.

Introduction

Luke Paluso and Alex Kasman connect two corners of modern mathematics: exceptional Hermite polynomials and Calogero-Moser pairs. The polynomials belong to George Wilson's adelic Grassmannian, a space with two faces. One face speaks of differential conditions; the other, of Calogero-Moser pairs, pairs of matrices with a tight commutation rule.

Earlier work placed the generating functions of each exceptional Hermite family inside the Grassmannian by the first route. No one had tried the second. Drawing on the first author's thesis, this note gives a new formula for the polynomials written in Calogero-Moser terms. A theorem then uses the match to build explicit, finitely supported distributions that wipe the polynomials out.

Catalog record

CREATORLuke Paluso, Alex Kasman
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-21935

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Exceptional Hermite Polynomials and Calogero-Moser Pairs Luke Paluso, Alex Kasman Abstract . There are two equivalent descriptions of George Wilson’s adelic Grassmannian G ​ r a ​ d Gr^{ad} , one in terms of differential “conditions” and another in terms of Calogero-Moser Pairs. The former approach was used in the 2020 paper by Kasman-Milson which found that each family of Exceptional Hermite Polynomials has a generating function which lives in G ​ r a ​ d Gr^{ad} . This suggests that Calogero-Moser Pairs should also be useful in the study of Exceptional Hermite Polynomials, but no researchers have pursued that line of inquiry prior to the first author’s thesis. The purpose of this note is to summarize highlights from that thesis, including a novel formula for Exceptional Hermite Polynomials in terms of Calogero-Moser Pairs and a theorem utilizing this correspondence to produce explicit finitely-supported distributions which annihilate them. (First published as Contemporary Mathematics Vol. 822 pp. 167–180 https://doi.org/10.1090/conm/822/16481. © 2025 Amer. Math. Society.) 1. Introduction The mathematical objects mentioned in the title have separately been the subjects of many papers. Section 2 below will provide additional details, but for the sake of this introduction let us just briefly recall: • A pair of N\times N constant complex matrices (X,Z) whose commutator differs from the identity by a rank one matrix are called a Calogero-Moser Pair (CM Pair). Interest in them first arose in the 1970s to explain the integrable dynamics of the Calogero-Moser particle system using Hamiltonian reduction [ 7 ] . In 1998, Wilson showed that the rank one bispectral wave functions of the KP Hierarchy can all be written in terms of CM Pairs. [ 11 ] . • Exceptional orthogonal polynomials, a generalization of the classical orthogonal polynomials sharing many of their important properties despite lacking polynomials of certain degrees, were first defined in 2009 [ 3 ] and the particular example of the Exceptional Hermite Polynomials (XHPs) was first considered in 2014 [ 2 ] . It is not clear from those descriptions that there is any meaningful connection between XHPs and CM Pairs and they have not appeared together in any prior papers. However, it was shown in 2020 that each family of XHPs has a generating function which is a rank one bispectral wave function of the KP Hierarchy [ 6 ] . Consequently, it follows from Wilson’s observation that the generating functions for XHPs can be written in terms of CM Pairs. Finding a formula for XHPs in terms of CM Pairs and applications for the correspondence between them was the thesis project assigned to the first author by the second author. This note constitutes a research announcement aiming to convey the main results found in [ 8 ] . The thesis and its references should be consulted for details and complete proofs. 2. Background Material 2.1. Partitions and Young Diagrams For a fixed natural number N\in\mathbb{N} , a partition of N is a non-increasing sequence of non-negative integers \lambda=\left\{\lambda_{n}\right\}_{n\in\mathbb{N}} such that \displaystyle|\lambda|=\sum^{\infty}_{n=1}\lambda_{n}=N. Since the elements of the sequence must eventually be zero, a partition can be equivalently written as a finite list by omitting all of the 0 entries. The length of the partition \ell=\ell(\lambda) , is the number of non-zero entries in the sequence. For example, the partition \lambda=\{4,3,2,0,0,\ldots\} is a partition of |\lambda|=9 which has length \ell(\lambda)=3 and hence could also be written \lambda=\{4,3,2\} . Following [ 6 ] we define some sets of integers corresponding to the choice of a partition \lambda , (2.1) \displaystyle\mathcal{M}^{(\lambda)} \displaystyle=\left\{m_{i}=\lambda_{i}-i:i\in\mathbb{N}\right\} (2.2) \displaystyle\mathcal{K}_{n}^{(\lambda)} \displaystyle=\left\{m_{i}+n:1\leq i\leq n\right\} (2.3) \displaystyle\mathcal{I}^{(\lambda)} \displaystyle=\mathbb{Z}\setminus\mathcal{M}^{(\lambda)}+|\lambda|=\left\{z+|\lambda|\;:\;z\in\mathbb{Z}\setminus\mathcal{M}^{(\lambda)}\right\}. \mathcal{M}^{(\lambda)} is known as the Maya Diagram of a partition \lambda . The set defined in ( 2.2 ) depends on the choice of an integer n and will be of special interest in the case that n=\ell(\lambda) is the length of the partition: \mathcal{K_{\ell}}^{(\lambda)} is known as the Index Set of a partition and we will just write it as \mathcal{K}^{(\lambda)}=\left\{k_{1},\dots,k_{\ell}\right\} . It is also noteworthy that \mathcal{I}^{(\lambda)}\subset\mathbb{N} . A Young Diagram is a finite array of boxes arranged in left-justified rows, with the rows in non-increasing order according to their lengths. There is a natural bijection between partitions and Young Diagrams which associates to \lambda=\{\lambda_{1},\lambda_{2},\ldots\} the diagram whose i^{th} row has \lambda_{i} boxes. An alternate method for representing partitions is the Frobenius notation . Consider the Young Diagram corresponding to a partition \lambda . The number of boxes to the right of the box in the i^{th} row and i^{th} column in the Young Diagram is denoted \alpha_{i} . Similarly, \beta_{i} is the number of boxes below the box in the i^{th} row and i^{th} column. The partition corresponding to that Young Diagram can then be denoted \lambda=\left\{(\alpha_{1},\dots,\alpha_{N^{\prime}})|(\beta_{1},\dots,\beta_{N^{\prime}})\right\} where N^{\prime} is the highest integer j so that \lambda_{j}\geq j and where the Young Diagram has a box in position (j,j) . The Young Diagram for the partition \lambda={4,3,2} is shown in Figure 2.1 . Figure 1. \lambda=\{4,3,2\}=\left\{(3,1)|(2,1)\right\} Allboxestotherightoftheboxinthefirstcolumnandfirstrowarecoloredredcorrespondingwithα _1=3 .Similarlythenumberofboxesbelowtheboxinthefirstcolumnandfirstrowarecoloredblue,correspondingwith β _1=2 .Similarlywecanconclude α _2= β _2=1 .Hence,inFrobeniusnotationthepartitioncorrespondingtothatYoungDiagramis λ ={ (3 , 1) — (2 , 1)} . 2.2. Bivariate Exceptional Hermite Polynomials Before we get to the XHPs, we begin by defining the classical Hermite polynomials. Readers may already be familiar with them and inclined to skip over this definition, but note that here we work with them in an unusual bivariate form. For the purposes of this paper, the Hermite polynomials H_{n}(x,y) are defined by the generating function e^{xz+yz^{2}}=\sum^{\infty}_{n=0}H_{n}\frac{z^{n}}{n!}. In the special case y=-1/4 they yield the standard univariate Hermite polynomials. Moreover, for any negative value of y they are orthogonal satisfying \displaystyle\langle H_{n},H_{m}\rangle_{H_{y}}=2(-\pi y)^{\frac{1}{2}}(-2y)^{n}n!\delta_{n,m}. with respect to the inner product \displaystyle\langle f(x),g(x)\rangle_{H_{y}}=\int_{-\infty}^{\infty}f(x)g(x)e^{\frac{x^{2}}{4y}}dx. To each partition \lambda , we associate a family of bivariate Exceptional Hermite Polynomials (briefly referred to as the \lambda -XHPs) [ 6 ] . This is the set \left\{H_{n}^{(\lambda)}\ :\ n\in\mathcal{I}^{(\lambda)}\right\} of polynomials defined by the formula (2.4) \displaystyle H_{n}^{(\lambda)}=\frac{\text{Wr}(H_{k_{\ell}},H_{k_{\ell-1}}\dots,H_{k_{1}},H_{n-N+\ell})}{\prod_{i<j}(k_{i}-k_{j})\prod_{i}(n-N+\ell-k_{i})} \displaystyle\text{for}\;\;\;n\in\mathcal{I}^{(\lambda)}, where the Wronskian involves differentiation with respect to the variable x , N=|\lambda| , and where \mathcal{I}^{(\lambda)} and index set \mathcal{K}^{(\lambda)}=\{k_{1},\ldots,k_{\ell}\} are as in Section 2.1 . The polynomial H_{n}^{(\lambda)}(x,y) is homogeneous of weighted degree n where \text{deg}(x)=1 and \text{deg}(y)=2 . Consequently, the set includes polynomials of every non-negative degree except those in \mathcal{K}_{N}^{(\lambda)} . Whenever \lambda is an even partition (which is a partition whose distinct entries all appear an even number times) and y<0 is any negative number then the \lambda -XHPs are orthogonal with respect to a Hermite-like inner product: (2.5) \displaystyle\langle H^{(\lambda)}_{n},H^{(\lambda)}_{m}\rangle_{\lambda}=\int_{-\infty}^{\infty}H^{(\lambda)}_{n}H^{(\lambda)}_{m}\frac{e^{\frac{x^{2}}{4y}}}{\left(\tau^{(\lambda)}(x,y)\right)^{2}}\,dx=\nu_{n-N}(y)\delta_{n,m} where \tau^{(\lambda)}(x,y)=\frac{\text{Wr}(H_{k_{\ell}},\dots,H_{k_{1}})}{\prod_{i<j<\ell}(k_{i}-k_{j})}\hbox{ and }\nu_{n}(y)=2(-\pi y)^{\frac{1}{2}}(-2y)^{m}\frac{(m+\ell)!}{\prod_{i=1}^{\ell}(m-m_{i}(\lambda))}. Remark 2.1 . If the partition is not even, the \lambda -XHPs are still orthogonal under a modified version of the inner product above where the only change is that the path of integration in the complex plane is shifted in order to avoid the roots of \tau^{(\lambda)}(x,y) [ 4 ] . 2.3. The Adelic Grassmannian Gr^{ad} In his two papers that introduced and studied the so-called “adelic Grassmannian”, George Wilson provided two seemingly different but entirely equivalent descriptions [ 10 , 11 ] . They are each best described as a procedure for making functions of a certain type, in one case using Wronskians and linear functionals and in the other using matrix multiplication

Read the complete text