A FORMULA FOR CONNECTED BOSONIC N N -POINT FUNCTIONS FOR BKP HIERARCHY

XUHUI ZHANG, JIAN ZHOU · 2025

Read A FORMULA FOR CONNECTED BOSONIC N N -POINT FUNCTIONS FOR BKP HIERARCHY by XUHUI ZHANG, JIAN ZHOU (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-2506-00451.txt.

Introduction

Xuhui Zhang and Jian Zhou prove a formula, in a 2025 preprint, for a class of functions tied to the BKP hierarchy. The BKP hierarchy is a well-known system in math physics, one of a family of equation systems with rich structure. Its solutions are encoded in a special function called a tau-function. The paper concerns the connected n-point functions built from that tau-function, written in bosonic form.

The authors reach their formula by placing the BKP system inside the larger KP hierarchy. That embedding lets them transfer known results across. Their formula differs in shape from one published earlier by Wang and Yang. The main work of the paper is the proof that the two formulas always agree. For specialists, the result gives a second, sometimes handier route to the same quantities.

Catalog record

CREATORXUHUI ZHANG, JIAN ZHOU
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE66 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-00451

Browse the full library · About the Tower of Babel

Full text

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

A FORMULA FOR CONNECTED BOSONIC N N -POINT FUNCTIONS FOR BKP HIERARCHY XUHUI ZHANG, JIAN ZHOU Abstract We present a formula for the connected n n -point functions of a tau-funtion of the BKP hierarchy by embedding BKP hierarchy into KP hierarchy. This formula is different from the one given by Wang and Yang [ WY22 ] . We prove that these two formulae are equivalent. Keywords : KP hierarchy, BKP hierarchy, connected bosoinc n n -piont function 1 Introduction It is an important problem to find tau-functions and their corresponding (connected) n -point functions in mathematical physics. For the topological 2D gravity theory whose partition function is given by the Witten-Kontsevich tau-function [ Wit93 , Kon92 ] , the one-point function, two-function, and the three-function are found by Witten, Dijkgraaf and Zagier respectively. Okounkov [ Oko02 ] expressed the n -point functions in this case as n -dimensional error-function-type integrals and gave a new derivation of Witten’s KdV equations. In [ LX11 ] , Liu and Xu found a recursion formula for the n -point functions, and they used it to prove the Faber conjecture, see [ LX07 ] and [ LX09 ] . Buryak et al . proved a new formula in [ BSSZ15 ] and Alexandrov et al. [ AHIS21 ] gave yet another derivation of the Witten Conjecture/Kontsevich Theorem. It involves a summation over all permutations \tau of \{1,\dots,n\} such that \tau(1)=1 and as in the Okounkov formula, an integral over n -dimensional space. More recently, a different type of formulae for n -point functions for various theories governed by integrable hierarchies for tau-functions have been derived using matrix resolvent. In [ BDY16 ] , Bertola, Dubrovin and Yang found an explicit connected n -point functions of an arbitrary tau-function of the KdV hierarchy in terms of matrix resolvents and a summation over permutations. Dubrovin-Yang [ DY17 ] used the similar method to find the connected n -point functions of tau-functions of Toda lattice hierarchy in terms of matrix resolvents of the difference Lax operator, and as an application, they obtained explicit generating series for connected GUE correlators. See [ DYZ21 ] and [ Yan20 ] for another proof and some new formulae. Moreover, Mattia and Yang [ CY22 ] extended the matrix-resolvent method for computing the connected n -point functions of tau-functions of the Ablowitz-Ladik hierarchy. Inspired by the above work, yet another type of formulae for n -point function has been derived. In [ Zho13a ] , the second author found an explicit formula for the fermionic representation of Witten-Kontsevich tau-function. Later, Balogh and Yang [ BY17 ] proved that the coefficients given in [ Zho13a ] are the affine coordinates for the point of the Sato Grassmannian corresponding to the Witten-Kontsevich tau function. Combining the above ideas, Zhou [ Zho15a ] gave a formula for the connected bosoinc n -point function of KP hierarchy directly in terms of the affine coordinates. In particular a new formula for the n -point functions of the Witten-Kontsevich tau-function in terms of Airy functions is obtained by this approach. The key ingredient in deriving the formula in [ Zho15a ] is the boson-fermion correspondence applied to the study of KP hierarchy. This has played a crucial role in Kyoto school approach to integrable hierarchies [ MJD00 ] . They emphasized on the symmetries of the integrable hierarchies, and this leads naturally to infinite-dimensional Lie algebras and representations [ JM83 , KW88 , DS85 ] . In particular, the KP hierarchy corresponds to the basic representation of the Lie algebra of type A_{\infty} , and the BKP hierarchy corresponds to the spin representation of the Lie algebra of type B_{\infty} . In [ JM83 ] various reductions of the KP hierarchy have been considered by considering the fermionic constructions of the representations of the infinite-dimensional Lie algebras. Such constructions are obtained by taking the invariant subspaces of suitable involutions on the fermionic Fock space. This has inspired Zhou [ Zho15b ] to propose the approach to various integrable hierarchies by embedding them into the KP hierarchy. This can be achieved by suitable embedding of the Lie algebras of their transformation groups into the Lie algebra of type A_{\infty} . On the other hand, for the Lie algebras in [ JM83 ] , their corresponding integrable hierarchies can also be studied by applying the corresponding boson-fermion correspondences. A priori, the above two approaches should lead to equivalent results, however the appearance of the results may be different, and to establish their equivalence can be a challenging problem. In this paper, we will address such issues for the formulae for n -point functions of the BKP hierarchy. Using the boson-fermion correspondence of type B, Wang and Yang [ WY22 ] modified Zhou’s method to get a formula for the connected bosoinc n -point function of tau-function of BKP hierarchy. More recently they generalize to the cases of connected (n,m) -point functions of 2D Toda lattice hierarchy [ WY+23b ] and diagonal 2-BKP hierarchy [ WY23a ] . In this paper we first apply the approach proposed in [ Zho15b ] to derive another formula for the BKP case from the formula for the KP hierarchy, and then we prove that our formula is equivalent to the formula in [ WY22 ] . Our method for deriving the formula for n -point functions is clearly applicable to other reductions of the KP hierarchy. It has the advantage of not depending on the specific boson-fermion correspondence for the reduction. Given a \tau -function \tau^{BKP} of the BKP hierarchy, it is well known that there is a \tau -function \tau^{KP} of the KP hierarchy such that \tau^{KP}(t_{1},0,t_{3},0,t_{5},0\ldots)=\tau^{BKP}(t_{1},t_{3},t_{5},\ldots)^{2}. See [ DJKM82 ] . The first main result of this paper is the following theorem. For definitions and notations, see § 2 . Theorem 1.1 . Let \tau^{BKP}({\boldsymbol{t}}) be a \tau -function of the BKP hierarchy and \tau^{KP}(\boldsymbol{t}) be the \tau -function of the KP hierarchy such that \tau^{KP}(t_{1},0,t_{3},0,t_{5},0\ldots)=\tau^{BKP}(t_{1},t_{3},t_{5},\ldots)^{2}. Define F^{BKP}(\boldsymbol{t}):=\log\tau^{BKP}(\boldsymbol{t}) . Then 1. The generating series of the affine coordinates of \tau^{KP}(\boldsymbol{t}) in the KP sense is given by \displaystyle A^{KP}(w,z)= \displaystyle 2\left(\sum_{m=0}^{\infty}(-1)^{m+1}a_{m+1,0}^{BKP}w^{-m-1}z^{-1}\right. (1.1) \displaystyle+ \displaystyle\left.\sum_{m=0}^{\infty}\sum_{n=1}^{\infty}(-1)^{m+1}(a_{m+1,n}^{BKP}+a_{m+1,0}^{BKP}a_{0,n}^{BKP})w^{-m-1}z^{-n-1}\right) where a_{n,m}^{BKP} are the affine coordinates of \tau^{BKP}(\boldsymbol{t}) . 2. Let A^{BKP} be the genrating series of the affine coordinates of \tau^{BKP}(\boldsymbol{t}) in the BKP sense. The relationship between A^{KP} and A^{BKP} is given by A^{BKP}(w,z)=\frac{1}{4}(zA^{KP}(w,-z)-wA^{KP}(z,-w)). (1.2) 3. The connected bosonic n -point functions of F^{BKP}(\boldsymbol{t}) are given by \displaystyle\sum_{i_{1},\ldots,i_{n}\geq 1,odd}\frac{\partial^{n}F^{BKP}}{\partial t_{i_{1}}\cdots\partial t_{i_{n}}}\Bigg|_{\boldsymbol{t}=0}\prod_{k=1}^{n}z_{k}^{-i_{k}-1} (1.3) \displaystyle= \displaystyle\frac{(-1)^{n-1}}{2^{n+1}}\sum_{\begin{subarray}{c}\sigma:n-cycles\epsilon_{1},\ldots,\\ \epsilon_{n}\in\{\pm 1\}\end{subarray}}\prod_{i=1}^{n}\hat{A}^{KP}(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)})-\delta_{n,2}i_{z_{1},z_{2}}\frac{z_{1}^{2}+z_{2}^{2}}{2(z_{1}^{2}-z_{2}^{2})^{2}}. where \hat{A}^{KP} is defined by \hat{A}^{KP}(z_{i},z_{j})=\begin{cases}i_{z_{i},z_{j}}\frac{1}{z_{i}-z_{j}}+A^{KP}(z_{i},z_{j}),&i<j,\\ A^{KP}(z_{i},z_{j}),&i=j,\\ i_{z_{j},z_{i}}\frac{1}{z_{i}-z_{j}}+A^{KP}(z_{i},z_{j}),&i>j,\end{cases} (1.4) and here we use the following notation: \displaystyle i_{x,y}\frac{1}{(x+y)^{m}} \displaystyle=\sum_{k\geq 0}\binom{-m}{k}x^{-m-k}y^{k}, (1.5) \displaystyle i_{x,y}\frac{x^{2}+y^{2}}{(x^{2}-y^{2})^{2}} \displaystyle=\sum_{n=0}^{\infty}(2n+1)x^{-2n-2}y^{2n}. (1.6) Define \hat{A}^{BKP}(w,z)=A^{BKP}(w,z)-\frac{1}{4}-\frac{1}{2}\sum_{i=1}^{\infty}(-1)^{i}w^{-i}z^{i}. (1.7) The formula of connected bosonic n -point functions for BKP hierarchy given by Wang and Yang [ WY22 , Theorem 4.1] is \displaystyle\sum_{i_{1},\ldots,i_{n}>0,odd}\frac{\partial^{n}\log\tau^{BKP}}{\partial t_{i_{1}}\cdots\partial t_{i_{n}}}\Bigg|_{\boldsymbol{t}=0}z_{1}^{-i_{1}}\cdots z_{n}^{-i_{n}} (1.8) \displaystyle= \displaystyle-\delta_{n,2}z_{1}z_{2}i_{z_{1},z_{2}}\frac{z_{1}^{2}+z^{2}_{2}}{2(z_{1}^{2}-z_{2}^{2})^{2}}+\sum_{\begin{subarray}{c}\sigma:n-cycles,\\ \epsilon_{2},\ldots,\epsilon_{n}\in\{\pm 1\}\end{subarray}}(-\epsilon_{2}\cdots\epsilon_{n})\prod_{i=1}^{n}\xi(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)}), where \epsilon_{1}=1 and \xi(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)})=\begin{cases}\hat{A}^{BKP}(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)}),&\sigma^{i-1}(1)<\sigma^{i}(1),\\ A^{BKP}(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)}),&\sigma^{i-1}(1)=\sigma^{i}(1),\\ -\hat{A}^{BKP}(-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)},\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)}),&\sigma^{i-1}(1)>\sigma^{i}(1).\end{cases} (1.9) The above two formulae ( 1.8 ) and ( 1.3 ) are different. Our second main result is that they are equivalent. Theorem 1.2 . In the above notations, we have \displaystyle\sum_{\begin{subarray}{c}\sigma:n-cycles,\\ \epsilon_{2},\ldots,\epsilon_{n}\in\{\pm 1\}\end{subarray}}(-\epsilon_{2}\cdots\epsilon_{n})\prod_{i=1}^{n}\xi(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},-\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)}) (1.10) \displaystyle= \displaystyle\frac{(-1)^{n-1}}{2^{n+1}}\sum_{\begin{subarray}{c}\sigma:n-cycles\\ \epsilon_{1},\ldots,\epsilon_{n}\in\{\pm 1\}\end{subarray}}\prod_{i=1}^{n}\hat{A}^{KP}(\epsilon_{\sigma^{i-1}(1)}z_{\sigma^{i-1}(1)},\epsilon_{\sigma^{i}(1)}z_{\sigma^{i}(1)}). Surprisingly, the proof of this equivalence is very technical and tedious. We will need the

Read the complete text