Ext operators for wreath Macdonald polynomials

Seamus Albion Ferlinc, Joshua Jeishing Wen · 2025

Read Ext operators for wreath Macdonald polynomials by Seamus Albion Ferlinc, Joshua Jeishing Wen (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-2508-10772.txt.

Introduction

Seamus Albion Ferlinc and Joshua Jeishing Wen built new operators for a class of symmetric polynomials, in a short paper from twenty twenty-five. The objects are wreath Macdonald polynomials, a ringed form of a classic family. The team gives a vertex operator for them, in the spirit of a known one by Carlsson, Nekrasov, and Okounkov. As a use, they prove a modular formula of Nekrasov-Okounkov type for rank three and up. That formula had been guessed by Walsh and Warnaar.

Catalog record

CREATORSeamus Albion Ferlinc, Joshua Jeishing Wen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-10772

Browse the full library · About the Tower of Babel

Full text

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

Ext operators for wreath Macdonald polynomials Seamus Albion Ferlinc, Joshua Jeishing Wen Abstract . We introduce a wreath Macdonald polynomial analogue of the Carlsson–Nekrasov–Okounkov vertex operator. As an application, we prove a modular ( q , t ) (q,t) -Nekrasov–Okounkov formula for r ⩾ 3 r\geqslant 3 originally conjectured by Walsh and Warnaar. 2020 Mathematics Subject Classification Primary: 05E05, 33D52; Secondary: 14D21, 17B69. 1. Introduction 1.1. Ext operators The term “Ext operator” comes from the geometry of moduli of sheaves on surfaces, which will not appear in the main body of the paper. We elaborate here on these geometric origins; the discussion extends to more general moduli of sheaves on more general surfaces, but we will confine ourselves to the Hilbert scheme of points on \mathbb{C}^{2} . Namely, for each n , we have the following identification of closed points: \mathrm{Hilb}_{n}(\mathbb{C}^{2})\cong\left\{\mathcal{I}\subset\mathbb{C}[x,y]\,|\,\dim\mathbb{C}[x,y]\big/\mathcal{I}=n\right\}. Thus, \mathrm{Hilb}_{n}(\mathbb{C}^{2}) parameterizes certain ideal sheaves. We set \mathrm{Hilb}\coloneq\bigsqcup_{n}\mathrm{Hilb}_{n}(\mathbb{C}^{2}). After choosing coordinates, the two-dimensional torus T acts naturally on \mathbb{C}^{2} . Given two equivariant sheaves \mathcal{F} and \mathcal{G} on \mathbb{C}^{2} , one can define their T -equivariant K -theoretic Euler characteristic \chi_{\mathbb{C}^{2}}(\mathcal{F},\mathcal{G})=\sum_{i}(-1)^{i}\mathrm{Ext}^{i}_{\mathbb{C}^{2}}(\mathcal{F},\mathcal{G}). Carlsson, Nekrasov, and Okounkov [ CO12 , CNO14 ] consider the virtual class E of vector bundles on \mathrm{Hilb}\times\mathrm{Hilb} defined as follows: at the point corresponding to a pair of ideals (\mathcal{I}_{1},\mathcal{I}_{2})\in\mathrm{Hilb}\times\mathrm{Hilb} , the fiber of E is E(\mathcal{I}_{1},\mathcal{I}_{2})=\chi_{\mathbb{C}^{2}}(\mathcal{O}_{\mathbb{C}^{2}},\mathcal{O}_{\mathbb{C}^{2}})-\chi_{\mathbb{C}_{2}}(\mathcal{I}_{1},\mathcal{I}_{2}). The first summand \chi_{\mathbb{C}^{2}}(\mathcal{O}_{\mathbb{C}^{2}},\mathcal{O}_{\mathbb{C}^{2}}) is just a normalization term; we call E the Ext bundle . Lifting the T -action on \mathbb{C}^{2} naturally to \mathrm{Hilb} , the torus fixed points correspond to monomial ideals \{\mathcal{I}_{\lambda}\} , which can be indexed by partitions. In [ CO12 ] , the authors compute (1.1) \sum_{k}(-u)^{k}{\textstyle\bigwedge^{k}}E(\mathcal{I}_{\lambda},\mathcal{I}_{\mu})=\prod_{\square\in\lambda}\left(1-uq^{-a_{\mu}(\square)}t^{l_{\lambda}(\square)+1}\right)\prod_{\square\in\mu}\left(1-uq^{a_{\lambda}(\square)+1}t^{-l_{\mu}(\square)}\right). Here, a_{\mu}(\square) and l_{\mu}(\square) are the (generalized) arm and leg lengths of the box; we define these quantities in Subsection 2.2.1 . The product on the right-hand-side of ( 1.1 ) is the Nekrasov factor , a combinatorial quantity that has played a large role in the AGT correspondence [ AGT10 , AFH+11 ] . Either from Haiman’s celebrated proof of the Macdonald positivity conjecture [ Hai01 ] or from Grojnowski and Nakajima’s Heisenberg action [ Gro96 , Nak97 ] , it is well-known that one should study cohomology or K -theory of \mathrm{Hilb} in terms of the ring of symmetric functions \Lambda [ Mac15 ] . In K -theory, the class of the skyscraper sheaf [\mathcal{O}_{\mathcal{I}_{\lambda}}] corresponds to the modified Macdonald polynomial H_{\lambda} . One of the key results from [ CO12 , CNO14 ] is a description of the left-hand-side of ( 1.1 ) in terms of natural operators on symmetric functions. Thus, they are able to reproduce ( 1.1 ) using certain operators applied to H_{\lambda} and H_{\mu} . 1.2. Wreath Macdonald polynomials This paper is concerned with a generalization of [ CNO14 ] where one replaces the Macdonald polynomials with wreath Macdonald polynomials . Recall that under the Frobenius characteristic, we can use \Lambda to describe the representation rings of symmetric groups \Sigma_{n} : \Lambda\cong\bigoplus_{n}\mathrm{Rep}(\Sigma_{n}). Defined by Haiman [ Hai03 ] , the wreath Macdonald polynomials are characters for the wreath product \mathbb{Z}/r\mathbb{Z}\wr\Sigma_{n} . Fixing r and letting n vary, we also have a wreath Frobenius characteristic [ Mac15 , Appedix I.B] : \Lambda^{\otimes r}\cong\bigoplus_{n}\mathrm{Rep}(\mathbb{Z}/r\mathbb{Z}\wr\Sigma_{n}). The wreath Macdonald polynomials are also indexed by (single) partitions; we will denote them again by \{H_{\lambda}\} . The decomposition of a partition into its r -core and r -quotient (cf. Subsection 2.3 below) plays a role in their definition, and letting \lambda range over all of those with the same r -core, we obtain a basis of \mathbb{C}(q,t)\otimes\Lambda^{\otimes r} . The geometric story told in Subsection 1.1 above also has a wreath analogue. By work of Gordon [ Gor08 ] and Bezrukavnikov and Finkelberg [ BF14 ] (cf. [ Hai03 ] as well), the wreath Macdonald polynomials are also intimately tied to the K -theory of \mathrm{Hilb} . Namely, consider the action of \mathbb{Z}/r\mathbb{Z} on \mathbb{C}^{2} via matrices \begin{pmatrix}e^{\frac{k\pi i}{r}}&0\\ 0&e^{-\frac{k\pi i}{r}}\end{pmatrix}. Lifting this action to \mathrm{Hilb} , one can consider the fixed subvarieties \mathrm{Hilb}^{\mathbb{Z}/r\mathbb{Z}} . These are still smooth, and they still contain the monomial ideals \{\mathcal{I}_{\lambda}\} . The aforementioned work assign the class of the skyscraper sheaf \left[\mathcal{O}_{\mathcal{I}_{\lambda}}\right] , now in K_{T}(\mathrm{Hilb}^{\mathbb{Z}/r\mathbb{Z}}) , to H_{\lambda} . Restricted to \mathrm{Hilb}^{\mathbb{Z}/r\mathbb{Z}}\times\mathrm{Hilb}^{\mathbb{Z}/r\mathbb{Z}} , the Ext bundle E becomes a bundle of \mathbb{Z}/r\mathbb{Z} -modules, and we can take its invariants E^{\mathbb{Z}/r\mathbb{Z}} . Equation ( 1.1 ) becomes \sum_{k}(-u)^{k}{\textstyle\bigwedge^{k}}\bigg(E(\mathcal{I}_{\lambda},\mathcal{I}_{\mu})^{\mathbb{Z}/r\mathbb{Z}}\bigg)=N_{\lambda,\mu}(u), where (1.2) N_{\lambda,\mu}(u)\coloneq\prod_{\begin{subarray}{c}\square\in\lambda\\ h_{\mu,\lambda}(\square)\equiv 0\>\mathrm{mod}\>r\end{subarray}}\left(1-uq^{-a_{\mu}(\square)}t^{l_{\lambda}(\square)+1}\right)\prod_{\begin{subarray}{c}\square\in\mu\\ h_{\lambda,\mu}(\square)\equiv 0\>\mathrm{mod}\>r\end{subarray}}\left(1-uq^{a_{\lambda}(\square)+1}t^{-l_{\mu}(\square)}\right). Here, the quantity h_{\mu,\lambda}(\square) is the mixed hook length, defined in Subsection 2.2.1 . The expression N_{\lambda,\mu}(u) is the Nekrasov factor for ALE space, considered in [ AKM+18 , Equation (5.33)] . Our main result is the following. Theorem 1.1 . Let r\geqslant 3 . Define \mathsf{W}(u)\coloneq\Omega\left[\frac{(1-u^{-1})X^{(0)}}{(1-q\sigma^{-1})(t\sigma-1)}\right]\mathcal{T}\left[(1-uqt)X^{(0)}\right]. For \lambda and \mu with the same r -core, we have \left\langle H_{\mu}^{\dagger},\mathsf{W}(u)H_{\lambda}\right\rangle_{q,t}^{\prime}=u^{-|\mathrm{quot}(\mu)|}N_{\lambda,\mu}(u). All the notation in the theorem is introduced in Sections 2 and 3 . Our formula should also hold for r=2 . The only obstruction is that the results from [ Wen25a ] involving the quantum toroidal algebra are only proved for r\geqslant 3 ; the algebra for r=2 has different formulas. Our method of proof is philosophically identical to [ CNO14 ] . Although not obvious at first, it turns out that Theorem 1.1 is intimately tied to the reciprocity property for Macdonald polynomials. In [ CNO14 ] , this is manifested in their use of the Cherednik–Macdonald–Mehta identity [ Che97 ] . Here, we phrase this in terms of Tesler’s identity [ GT96 , GHT99 ] , which was proven for wreath Macdonald polynomials by Romero and the second author [ RW25b ] . 1.3. Nekrasov–Okounkov formulas The ordinary Nekrasov–Okounkov formula is the following expansion for an arbitrary complex power of the Dedekind eta function: (1.3) \sum_{\lambda}T^{|\lambda|}\prod_{\square\in\lambda}\bigg(1-\frac{z}{h_{\lambda}^{2}(\square)}\bigg)=\prod_{i\geqslant 1}(1-T^{i})^{z-1}. Note that h_{\lambda}(\square)=h_{\lambda,\lambda}(\square) is the ordinary hook length of the box \square . One may either take 0<T<1 and z\in\mathbb{C} or T\in\mathbb{C} with |T|<1 and z\in\mathbb{R} . This was discovered by Nekrasov and Okounkov [ NO06 , Equation (6.12)] in the course of their proof of Nekrasov’s conjecture [ Nek04 ] . It was found independently by Westbury as a hook-length formula for the D’Arcais polynomials [ Wes06 ] . Generalizations of the Nekrasov–Okounkov formula abound in the literature; see, for instance, [ CRV18 , DH11 , Han10 , HJ11 , Pét16b , Pét16a , RW16 , Wah22 , WW20 ] and references therein. One such generalization is Han’s modular version wherein the product in the summand is replaced by a product over boxes with hook length congruent to zero modulo r for a positive integer r [ Han10 , Theorem 1.4] . This was further expanded on by Han and Ji by way of their so-called “multiplication-addition theorem” [ HJ11 ] which allows for modular analogues of a vast swathe of identities involving hook-lengths. In a different direction, Rains and Warnaar [ RW16 , Theorem 1.3] and Carlsson and Rodriguez Villegas [ CRV18 , Theorem 1.0.2] independently proved a (q,t) -deformation of the Nekrasov–Okounkov formula. To state this, let (a_{1},\ldots,a_{n};q_{1},\ldots,q_{m})_{\infty}=\prod_{i=1}^{n}\prod_{j=1}^{m}\prod_{k=0}^{\infty}\big(1-a_{i}q_{j}^{k}\big) denote the infinite multiple q -Pochhammer symbol. Then (1.4) \sum_{\lambda}T^{|\lambda|}\prod_{\square\in\lambda}\frac{(1-uq^{a_{\lambda}(\square)+1}t^{l_{\lambda}(\square)})(1-u^{-1}q^{a_{\lambda}(\square)}t^{l_{\lambda}(\square)+1})}{(1-q^{a_{\lambda}(\square)+1}t^{l_{\lambda}(\square)})(1-q^{a_{\lambda}(\square)}t^{l_{\lambda}(\square)+1})}=\frac{(uqT,u^{-1}tT;q,t,T)_{\infty}}{(T,tT;q,t,T)_{\infty}}. For q=t this reduces to a q -analogue of the Nekrasov–Okounkov formula which may be found in [ DH11 ] or [

Read the complete text