Generating functions of $\mathcal{W}_{1+\infty}$ action on symmetric functions

Caleb Fernelius, Natasha Rozhkovskaya Department of Mathematics, Kansas State University, Manhattan, KS 66502, USA [email protected]

Abstract.

We describe the action of the infinite-dimensional Lie algebra $\mathcal{W}_{1+\infty}$ and its B-type analogue on Schur and Schur Q-functions, respectively, using formal distributions framework. We observe an interesting self-duality property possessed by these compact formulas.

2010 Mathematics Subject Classification:

Primary 17B69, Secondary 35Q51, 20G43, 05E05.

1. Introduction

The unique non-trivial central extension of the Lie algebra of differential operators on the circle, commonly known as the $\mathcal{W}_{1+\infty}$ algebra, gained prominence in the 1990’s for its role in two-dimensional quantum field theory and integrable systems. Several classes of representations of this infinite-dimensional Lie algebra have been well-studied (e.g. [4], [7], [11], [12], [13], [24] and in particular, review [3]). Recent interest in the structure of $\mathcal{W}_{1+\infty}$ comes from the connection of enumerative geometry to soliton integrable hierarchies. Studies in these areas revealed that certain generating functions for intersection numbers on moduli spaces of stable curves provide examples of tau-functions of famous integrable hierarchies. Classical results state that Schur symmetric functions serve as tau-functions of the KP hierarchy [26], and Schur $Q$ -functions serve as tau-functions of the BKP hierarchy [29], and other tau-functions of the corresponding hierarchies can be expressed as finite or infinite linear combinations of these classical solutions with coefficients satisfying certain algebraic constraints. Hence, the generating functions of intersection numbers that are tau-functions of the KP or the BKP hierarchies, can be also expressed through these celebrated families [1, 2, 21, 18, 17, 19]. In this context, the symmetries of the KP, KdV, BKP hierarchies are often described through the action of $\mathcal{W}_{1+\infty}$ operators and their analogues. In particular, in [19] the explicit formulas for the action of certain operators from Lie algebra $\mathcal{W}_{1+\infty}$ and its $B$ type analogue on the Schur functions and Schur $Q$ - functions respectively were found, generalizing particular cases studied in earlier papers, see cf. [19] for citations.

In this paper we continue the discussion of the structure of $\mathcal{W}_{1+\infty}$ and its $B$ type analogue in the language of formal distributions that serve as generating functions for actions of these algebras on families of symmetric functions. Due to close connection of representations of infinite-dimensional Lie algebras to the areas studied with the tools of vertex operators, this language is natural and provides additional insight in the structure of such actions. We discovered that in terms of generating functions formulas of action of $\mathcal{W}_{1+\infty}$ and its $B$ type analogue on Schur symmetric functions and Schur $Q$ -functions respectively are surprisingly compact, involve no derivations, but only multiplication operators, and exhibit an interesting duality, as explained in Remarks 6.1, 10.1. We also show that $\mathcal{W}_{1+\infty}$ has the structure of a conformal Lie algebra, and provide formulas for the action of generators of infinite-dimensional Lie algebras $\hat{a}_{\infty}$ , $\hat{o}_{\infty}$ in terms of formal distributions of operators acting on symmetric functions.

The paper is organized in the following way. In Sections 2, 3 we review the algebraic structure of the infinite-dimensional Lie algebras $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ in the language of formal distributions and their realizations through the Clifford algebra of charged free fermions. In Section 2.6 we describe the conformal Lie algebra structure of $\mathcal{W}_{1+\infty}$ . In Section 4 we collect necessary facts about symmetric functions. Sections 5, 6 describe the action of algebraic structures on the space of symmetric functions, with Theorem 6.1 as the main result. Particular examples and some connections with formulas deduced by other authors are outlined in the same section. In Section 7 we introduce $\mathcal{W}^{B}_{1+\infty}$ as a subalgebra of anti-involution invariants in $\mathcal{W}_{1+\infty}$ , and in Section 8 we express formal distributions of generators of this subalgebra and the infinite-dimensional Lie algebra $\hat{o}_{\infty}$ through the Clifford algebra of neutral fermions. In Section 9 we review necessary facts about Schur $Q$ -functions, and in Section 10 we prove Theorem 10.1 that provides actions of $\mathcal{W}^{B}_{1+\infty}$ and $\hat{o}_{\infty}$ on families of symmetric functions.

2. Lie subalgebra $\mathcal{W}_{1+\infty}$ of $\hat{a}_{\infty}$

2.1. Lie algebra $\hat{a}_{\infty}$

For our goals it is convenient to introduce $\mathcal{W}_{1+\infty}$ as a central extension of a matrix subalgebra, cf. [12]. Consider the Lie algebra $a_{\infty}$ of infinite matrices with a finite number of nonzero diagonals

a_{\infty}=\{(a_{ij})_{i,j\in{\mathbb{Z}}}|\quad a_{ij}=0\text{\quad for \quad}|i-j|>>0\}.

Any element of $a_{\infty}$ is a finite linear combination of matrices of the form $\sum_{i\in{\mathbb{Z}}}\lambda_{i}E_{i,i+k},$ and the Lie algebra structure

\left[\sum\lambda_{i}E_{i,i+k},\sum\mu_{i}E_{j,j+k}\right]=\sum(\lambda_{i}\mu_{i+k}-\lambda_{i+l}\mu_{i})E_{i,i+k+l}

comes from the commutation relations of generators, which are standard matrix units $\{E_{ij}\}_{i,j\in{\mathbb{Z}}}$ :

[E_{ij},E_{kl}]=\delta_{j,k}E_{il}-\delta_{l,i}E_{kj}.

The central extension $\hat{a}_{\infty}=a_{\infty}\oplus{\mathbb{C}}C$ is the Lie algebra with central element $C$ and generators $\{\hat{E}_{ij}\}_{i,j\in{\mathbb{Z}}}$ such that

\displaystyle[\hat{E}_{ij},\hat{E}_{kl}]=\delta_{j,k}\hat{E}_{il}-\delta_{l,i}\hat{E}_{kj}+\gamma(\hat{E}_{ij},\hat{E}_{kl})\,C,

(2.1)

and the cocycle $\gamma(\hat{E}_{ij},\hat{E}_{kl})$ has the values

\begin{cases}\gamma(\hat{E}_{ij},\hat{E}_{ji})=1,&\text{if}\quad i\leq 0,\quad j\geq 1,\\ \gamma(\hat{E}_{ij},\hat{E}_{ji})=-1,&\text{if}\quad i\geq 1,\quad j\leq 0,\\ \gamma(\hat{E}_{ij},\hat{E}_{kl})=0,&\text{otherwise.}\end{cases}

2.2. Formal distributions

Commutation relations of $\hat{a}_{\infty}$ and other algebraic structures below can be presented in the form of relations on generating functions. For this, following [14], [15], we recall the notion of a formal distribution. Let $W$ be a vector space. A $W$ - valued formal distribution is a bilateral series in the indeterminate $u$ with coefficients in $W$ :

\displaystyle a(u)=\sum_{n\in{\mathbb{Z}}}a_{n}u^{n},\quad a_{n}\in W.

A formal distribution in two and more indeterminates is defined similarly. The vector space of all $W$ -valued formal distributions in indeterminate $u$ is denoted as $W[[u,u^{-1}]]$ , and we also use the notation $W[u]$ for the space of polynomials, $W[[u]]$ for the space of power series, $W[u,u^{-1}]$ for the space of Laurent polynomials, and $W((u))$ for the space of formal Laurent series. Formal distributions can be added and multiplied by Laurent polynomials, but in general multiplication of two formal distributions is not a well-defined operation and can be performed only in special cases.

The formal delta-distribution $\delta(u,z)$ is the ${\mathbb{C}}$ -valued formal distribution

\displaystyle\delta(u,z)=\sum_{i\in{\mathbb{Z}}}\frac{u^{i}}{z^{i+1}}=i_{u/z}\left(\frac{1}{z-u}\right)+i_{z/u}\left(\frac{1}{u-z}\right),

where

\displaystyle i_{u/z}\left(\frac{1}{z-u}\right)=\sum_{i\geq 0}\frac{u^{i}}{z^{i+1}}

is the expansion of the rational function $1/(z-u)$ in the domain $|u|<|z|$ .

One has

	$\displaystyle\delta(u,v)=\delta(v,u)$	$\displaystyle=-\delta(-u,-v),\quad\partial_{u}\delta(u,v)=-\partial_{v}\delta(u,v),$		(2.2)
	$\displaystyle\delta(u,z)a(z)$	$\displaystyle=\delta(u,z)a(u)\quad\text{for any formal distribution}\,a(u).$		(2.3)

Consider the linear space $\hat{a}_{\infty}[[u,u^{-1},w,w^{-1}]]$ of formal distributions in two indeterminates with coefficients in $\hat{a}_{\infty}$ . Combining generators $\{\hat{E}_{ij}\}$ of $\hat{a}_{\infty}$ into a formal distribution $T(u,w)=\sum_{i,j}\hat{E}_{ij}u^{i-1}w^{-j}$ , commutation relations (2.1) can be presented in the form

	$\displaystyle[T(u,w),T(v,z)]$	$\displaystyle=\delta(v,w)\left(T(u,z)+i_{z/u}\left(\frac{1}{u-z}\right)C\right)\,$
		$\displaystyle-\,\delta(u,z)\left(T(v,w)+i_{w/v}\left(\frac{1}{v-w}\right)C\right),$		(2.4)

and $[T(u,w),C]=0$ .

Remark 2.1.

We introduce notation for the central part of this commutation relation:

	$\displaystyle\gamma(u,w,v,z)$	$\displaystyle=\delta(v,w)i_{z/u}\left(\frac{1}{u-z}\right)-\,\delta(u,z)i_{w/v}\left(\frac{1}{v-w}\right)$		(2.5)
		$\displaystyle=i_{z/u}\left(\frac{1}{u-z}\right)i_{v/w}\left(\frac{1}{w-v}\right)-i_{u/z}\left(\frac{1}{z-u}\right)i_{w/v}\left(\frac{1}{v-w}\right).$

Clearly, from (2.2), $\gamma(u,w,v,z)=-\gamma(v,z,u,w).$

The following lemma will be useful for further calculations with $\gamma(u,w,v,z)$ .

Lemma 2.1.

\displaystyle\partial_{u}^{k}\partial_{v}^{l}\gamma(u,w,v,z\,)|_{w=u,z=v}

\displaystyle=\frac{(-1)^{k}k!\,l!}{(k+l+1)!}\partial^{k+l+1}_{v}\delta(u,v),

(2.6)

and

\displaystyle\partial_{u}^{k}\partial_{v}^{l}\gamma(u,w,v,z\,)|_{w=-u,z=-v}=\frac{k!l!}{(k+l+1)!}\partial^{k+l+1}_{v}\delta(u,-v).

(2.7)

In particular,

\displaystyle\gamma(u,u,v,v\,)=\partial_{v}\delta(u,v),\quad\gamma(u,-u,v,-v\,)=\partial_{v}\delta(u,-v).

Proof.

For the proof of (2.6) observe that

	$\displaystyle\partial_{u}^{k}\partial_{v}^{l}\gamma(u,w,v,z))\|_{w=u,z=v}$
	$\displaystyle=\left.\left(i_{z/u}\left(\frac{(-1)^{k}k!}{(u-z)^{k+1}}\right)i_{v/w}\left(\frac{l!}{(w-v)^{l+1}}\right)-\,i_{u/z}\left(\frac{k!}{(z-u)^{k+1}}\right)i_{w/v}\left(\frac{(-1)^{l}l!}{(v-w)^{l+1}}\right)\right)\right\|_{{w=u,z=v}}$
	$\displaystyle=i_{v/u}\left(\frac{(-1)^{k}k!l!}{(u-v)^{k+l+2}}\right)-i_{u/v}\left(\frac{(-1)^{l}k!l!\,}{(v-u)^{k+l+2}}\right)=\frac{(-1)^{k}k!l!}{(k+l+1)!}\partial^{k+l+1}_{v}\delta(u,v).$

The proof of (2.7) follows the same lines. ∎

2.3. Infinite-dimensional Lie algebra $\mathcal{W}_{1+\infty}$

Introduce a family of formal distributions $T^{(k)}(u)\in\hat{a}_{\infty}[[u,u^{-1}]]$ by

\displaystyle T^{(k)}(u)=\left.\partial^{k}_{u}T(u,v)\right|_{v=u}.

Coefficients of the expansion $T^{(k)}(u)=\sum_{r}T^{(k)}_{r}u^{r-k-1}$ are given by

\displaystyle T^{(k)}_{r}=\sum_{j}(r+j-1)_{k}\hat{E}_{r+j,j},

(2.8)

where $(a)_{k}=a(a-1)\dots(a-k+1)$ is the falling factorial. Using that $T^{(k)}(u)$ is a coefficient of the Taylor series expansion

\displaystyle T(u+t,u)=e^{t\partial_{u}}T(u,v)\,|_{v=u}=\sum_{k=0}^{\infty}\frac{\partial^{k}_{u}T(u,v)\,|_{v=u}t^{k}}{k!}=\sum_{k=0}^{\infty}\frac{T^{(k)}(u)}{k!}t^{k},

(2.9)

we deduce the commutation relations of formal distributions $\{T^{(k)}(u)\}$ .

Proposition 2.1.

$\displaystyle[T^{(r)}(u),T^{(k)}(v)]=$	$\displaystyle\sum_{m=0}^{k}{k\choose m}{\partial_{v}^{m}\delta(v,u)\,T^{(r+k-m)}(v)}$
	$\displaystyle-\sum_{m=0}^{r}{r\choose m}{\partial_{u}^{m}}\delta(u,v)\,T^{(r+k-m)}(u)$
	$\displaystyle+\frac{(-1)^{r}r!k!}{(k+r+1)!}\partial^{k+r+1}_{v}\delta(u,v)C.$	(2.10)

Proof.

Using (2.3), (2.9) we write Taylor expansions of the terms of the commutator (2.2):

	$\displaystyle\delta(v+s,u)T(u+t,v)$	$\displaystyle=\delta(v+s,u)T(v+t+s,v)=\delta(v+s,u)\sum_{k=0}^{\infty}\frac{T^{(k)}(v)}{k!}(s+t)^{k}$
		$\displaystyle=\sum_{r=0}^{\infty}\sum_{k=0}^{\infty}\frac{s^{k}t^{r}}{k!r!}\sum_{m=0}^{k}{k\choose m}{\partial_{v}^{m}\delta(v,u)\,T^{(r+k-m)}(v)}.$		(2.11)

Similarly, with

\displaystyle\delta(u+t,v)T(v+s,u)

\displaystyle=\sum_{r=0}^{\infty}\sum_{k=0}^{\infty}\frac{s^{k}t^{r}}{k!r!}\sum_{m=0}^{r}{r\choose m}{\partial_{u}^{m}\delta(u,v)\,T^{(r+k-m)}(u)},

(2.12)

and applying (2.6) in the Taylor expansion

\displaystyle\gamma(u+t,u,v+t,v)

\displaystyle=\sum_{k=0}^{\infty}\sum_{r=0}^{\infty}\frac{s^{k}t^{r}}{k!r!}\partial_{u}^{k}\partial_{v}^{k}\gamma(u,w,v,z)\,|_{w=u,z=v},

(2.13)

we substitute (2.11), (2.12), (2.13) into (2.2) to get commutation relations $[T(u+t,u),T(v+s,v)]$ . Proposition 2.1 follows from (2.9). ∎

Note that Proposition 2.1 implies that the linear span of coefficients of $\{T^{(k)}_{r}\}_{r\in{\mathbb{Z}},k\in{\mathbb{Z}}_{\geq 0}}$ and the central element $C$ form a Lie subalgebra of $\hat{a}_{\infty}$ , commonly denoted as $\mathcal{W}_{1+\infty}$ . Commutation relations in $\mathcal{W}_{1+\infty}$ are described by the commutation relations (2.1) of generating functions $\{T^{(k)}(u)\}_{k\in{\mathbb{Z}}_{\geq 0}}$ .

2.4. $\mathcal{W}_{1+\infty}$ as the central extension of the algebra of differential operators

The Lie algebra $\mathcal{W}_{1+\infty}$ is isomorphic to the unique non-trivial central extension $\hat{D}$ of the Lie algebra $D$ of complex regular differential operators on a circle ${\mathbb{C}}^{\times}$ , see e.g. [12], [28] for more details. Lie algebra $D$ has a natural basis $\{t^{r}(t\partial_{t})^{k}|r\in{\mathbb{Z}},k\in{\mathbb{Z}}_{\geq 0}\}$ , where $t$ is the variable on ${\mathbb{C}}^{\times}$ . The natural action of $D$ on the space $t^{s}{\mathbb{C}}[t,t^{-1}]$ , where $s\in{\mathbb{Z}}$ , defines a homomorphism of Lie algebras $\varphi_{s}:D\to a_{\infty}$ , which can be extended to the homomorphism of central extensions $\varphi_{s}:\hat{D}\to\hat{a}_{\infty}$ . On the basis elements the homomorphism is defined as

\varphi_{s}(t^{r}(t\partial_{t})^{k})=\sum_{j\in{\mathbb{Z}}}(-j+s)^{k}\hat{E}_{j-r,j}.

Then from (2.8), for $l\in{\mathbb{Z}}_{\geq 0}$ , $k\in{\mathbb{Z}}$ , $T^{(k)}_{r}=\varphi_{s}(t^{-r}\,(-t\partial_{t}+s-r-1)_{k})$ . Since the highest degree term of $t^{-r}\,(-t\partial_{t}+s-r-1)_{k}$ is $(-1)^{k}t^{-r+k}\partial_{t}^{k}$ , the set $\{T^{(k)}_{r}\}_{k\in{\mathbb{Z}}_{\geq 0},r\in{\mathbb{Z}}.}$ forms a linear basis of $\mathcal{W}_{1+\infty}\simeq\hat{D}$ .

2.5. Definition of conformal Lie algebra

Proposition 2.1 implies that $\mathcal{W}_{1+\infty}$ defines a conformal Lie algebra. We review the definition of conformal Lie algebra and the $\lambda$ -bracket following [15],[6].

Definition 2.1.

Let $A$ be a ${\mathbb{C}}[\partial]$ -module with the $\lambda$ -bracket operation $[a_{\lambda}b]$ , which is a linear map $A\otimes A\to A[\lambda]={\mathbb{C}}[\lambda]\otimes A$ satisfying the axioms

(1)

(Skewcommutativity) $[a_{\lambda}b]=-[b_{-\lambda-\partial}a]$ ;
(2)

(Sesquilinearity) $[\partial a_{\lambda}b]=-\lambda[a_{\lambda}b]$ , $[a_{\lambda}\partial b]=(\partial+\lambda)[a_{\lambda}b]$
(3)

(Jacobi identity) $[a_{\lambda}[b_{\mu}c]]=[[a_{\lambda}b]_{\lambda+\mu}c]+[b_{\mu}[a_{\lambda}c]]$ .

Then $A$ is called a conformal Lie algebra.

Let $\mathfrak{g}$ be a Lie algebra. Recall [14], [15] that a pair of $\mathfrak{g}$ -valued formal distributions $a(u)=\sum_{n\in{\mathbb{Z}}}a_{n}u^{-n-1}$ and $b(v)=\sum_{n\in{\mathbb{Z}}}b_{n}v^{-n-1}$ is local if

(u-v)^{N}[a(u),b(v)]=0\quad\text{for}\quad N>>0.

Locality property is equivalent to the existence of the Operator Product Expansion (OPE), which means that the commutator of two formal distributions can be expressed through derivatives of formal delta distribution as

\displaystyle[a(u),b(v)]=

\displaystyle\sum_{j=0}^{N-1}c_{j}\,\partial^{j}_{v}\delta(u,v)/j!,

(2.14)

where the coefficient $c_{j}$ is denoted as $(a(v)_{(j)}b(v))$ and is called the $j$ -th product of $a$ and $b$ . One has

\displaystyle(a(v)_{(j)}b(v))=\text{Res}_{u}(u-v)^{j}[a(u),b(v)].

Many examples of conformal Lie algebras arise from families of pairwise local distributions, where the OPE (2.14) defines their lambda-bracket

\displaystyle[a(u),b(v)]=

\displaystyle\sum_{j=0}^{N-1}(a(v)_{(j)}b(v))\frac{\lambda^{j}}{j!}.

2.6. Conformal Lie algebra structure of $\mathcal{W}_{1+\infty}$

An adjusted version of commutation relations (2.1) implies that every pair $(T^{(k)}(u),\,T^{(l)}(v))$ is local, and that the family of formal distributions $\{T^{(k)}(u)\}_{k\in{\mathbb{Z}}_{\geq 0}}$ and their derivatives generates a conformal Lie algebra.

Proposition 2.2.

Let $\mathcal{T}$ be the minimal $\partial_{u}$ -invariant subspace of $\hat{a}_{\infty}[[u,u^{-1}]]$ that contains all $\{T^{(k)}(u)\}$ . Then $\mathcal{T}$ has the structure of a conformal Lie algebra with the $\lambda$ -bracket

	$\displaystyle[T^{(r)}\,_{\lambda}\,T^{(k)}]=$	$\displaystyle\quad\sum_{m=0}^{k}{k\choose m}\lambda^{m}{T^{(r+k-m)}}-\sum_{m=0}^{r}(-1)^{m}{r\choose{m}}(\lambda+\partial)^{m}T^{(r+k-m)}$		(2.15)
		$\displaystyle+\frac{(-1)^{r}r!k!}{(k+r+1)!}\lambda^{k+r+1}\,C.$

Proof.

Note that (2.1) has a nice symmetry in variables $u$ , $v$ , but it is not in the OPE form (2.14) that would immediately define a conformal Lie algebra structure, as it involves not only derivatives $\partial_{v}^{m}\delta(u,v)$ , but also $\partial_{u}^{m}\delta(u,v)$ . We convert (2.1) to the OPE form using the following lemma.

Lemma 2.2.

For any formal distribution $a(u)$ ,

\displaystyle\partial_{u}^{m}\delta(u,v)\,a(u)

\displaystyle=\sum_{p=0}^{m}(-1)^{m}{m\choose p}\partial^{m-p}_{v}\delta(u,v)\,\partial_{v}^{p}a(v).

Proof.

is by induction on $k$ . The base of the induction for $k=0$ is given by (2.3). Assume that for all values less or equal to $k$ and for any formal distribution $a(u)$ the statement holds. Note that from (2.2),

\displaystyle\partial_{u}^{l}\partial_{v}^{m}\delta(v,u)=\partial_{v}^{m}(\partial_{u}^{l}\delta(u,v))=(-1)^{l}\partial_{v}^{m+l}\delta(u,v).

(2.16)

By the induction hypothesis, the Leibniz rule and (2.16),

	$\displaystyle\partial_{u}^{k+1}\delta(u,v)\,a(u)=\partial_{u}\left(\partial_{u}^{k}\delta(u,v)\,a(u)\right)-\partial_{u}^{k}\delta(u,v)\,\partial_{u}a(u)$
	$\displaystyle=\partial_{u}\left(\sum_{p=0}^{k}(-1)^{k}{k\choose p}\partial^{k-p}_{v}\delta(u,v)\,\partial_{v}^{p}a(v)\right)-\sum_{p=0}^{k}(-1)^{k}{k\choose p}\partial^{k-p}_{v}\delta(u,v)\,\partial_{v}^{p+1}a(v)$
	$\displaystyle=\sum_{p=0}^{k}(-1)^{k+1}{k\choose p}\partial^{k-p+1}_{v}\delta(u,v)\partial_{v}^{p}a(v)\ +\sum_{p=1}^{k+1}(-1)^{k+1}{k\choose p-1}\partial^{k-p+1}_{v}\delta(u,v)\,\partial_{v}^{p}a(v)$
	$\displaystyle=\sum_{p=1}^{k}(-1)^{k+1}\left({k\choose p}+{k\choose p-1}\right)\partial^{k-p+1}_{v}\delta(u,v)\partial_{v}^{p}a(v)+(-1)^{k+1}\partial^{k+1}_{v}\delta(u,v)a(u)\,+\,$
	$\displaystyle+(-1)^{k+1}\delta(u,v)\partial_{v}^{k+1}a(v)=\sum_{p=0}^{k+1}(-1)^{k+1}{k+1\choose p}\partial^{k+1-p}_{v}\delta(u,v)\partial_{v}^{p}a(u).$

∎

By Lemma 2.2, the term ${\partial_{u}^{m}}\delta(u,v)\,T^{(r+k-m)}(u)$ in (2.1) contributes to $\lambda$ -bracket as

\sum_{p=0}^{m}(-1)^{m}{m\choose p}\lambda^{m-p}\partial_{v}^{p}T^{(r+k-m)}(v)=(\lambda+\partial)^{m}T^{(r+k-m)}(v),

which proves relation (2.15). Commutation relation (2.15) implies that formal distributions $\{T^{(k)}(u)\}$ are pairwise local, and that the $j$ -th products $T^{(k)}\,_{(j)}T^{(l)}$ are linear combinations of derivatives of $\{T^{(k)}(u)\}$ . Hence, $\mathcal{T}$ is closed under all $j$ -th products, and it is a conformal Lie algebra.

∎

2.7. Heisenberg and Virasoro algebras

One of the advantageous features of the $\lambda$ -bracket form (2.15) of commutation relations is that it effortlessly reveals the presence of several important algebraic structures.

Example 2.1.

For $r=k$ relation (2.15) takes the form

\displaystyle[T^{(k)}\,_{\lambda}\,T^{(k)}]=\quad\sum_{m=0}^{k}{k\choose m}(\lambda^{m}-(-1)^{m}(\lambda+\partial)^{m}){T^{(2k-m)}}+\frac{(-1)^{k}k!k!}{(2k+1)!}\lambda^{2k+1}\,C.

For $r=k=0$ the formula further reduces the defining relations of the Heisenberg algebra:

\displaystyle[T^{(0)}\,_{\lambda}T^{(0)}]=\lambda C,

since it encodes the commutator of formal distributions $[T^{(0)}(u),T^{(0)}(v)]=\partial_{v}\delta(u,v)C$ .

For $r=k=1$ we obtain the $\lambda$ -bracket

\displaystyle[T^{(1)}\,_{\lambda}\,T^{(1)}]=

\displaystyle\partial T^{(1)}+2\lambda T^{(1)}-\frac{1}{6}\lambda^{3}C,

(2.17)

which is the OPE of a Virasoro formal distribution:

\displaystyle[T^{(1)}(u),T^{(1)}(v)]=

\displaystyle\partial T^{(1)}(v)\delta(u,v)+2\partial_{v}\delta(u,v)T^{(1)}(v)-\frac{1}{6}\partial^{3}_{v}\delta(u,v)C.

Remark 2.2.

Note that original relation (2.1) is a symmetric in $u$ and $v$ version of the Virasoro commutation relation:

\displaystyle[T^{(1)}_{m},T^{(1)}_{n}]=\partial_{v}\delta(u,v)T^{(1)}(v)-\partial_{v}\delta(u,v)T^{(1)}(u)-\frac{1}{6}\partial^{3}_{v}\delta(u,v)C.

Similarly, for $r=k=2$ one gets

\displaystyle[T^{(2)}\,_{\lambda}\,T^{(2)}]=

\displaystyle\partial(2T^{(3)}-\partial T^{(2)})+\lambda(2T^{(3)}-\partial T^{(2)})+\frac{1}{30}\lambda^{5}C,

and for $k=r=3$ the $\lambda$ -bracket of the OPE looks like

	$\displaystyle[T^{(3)}\,_{\lambda}\,T^{(3)}]=$	$\displaystyle 3\partial T^{(5)}-3\partial^{2}T^{(4)}+\partial^{3}T^{(3)}+\,\lambda\,(6T^{(5)}-6\partial T^{(4)}+3\partial^{2}T^{(3)})$
		$\displaystyle+\,3\lambda^{2}\,\partial^{2}T^{(3)}+2\lambda^{3}T^{(3)}-\frac{1}{140}\lambda^{7}C.$

Example 2.2.

Substitution $r=1$ in (2.15) and the sesquilinearity property describes the action of the Virasoro algebra on the generators of the algebra $\mathcal{W}_{1+\infty}$ :

\displaystyle[T^{(1)}\,_{\lambda}T^{(k)}]=\partial T^{(k)}+(k+1)\lambda T^{(k)}+\sum_{m=2}^{k}\lambda^{m}{k\choose m}T^{(k+1-m)}-\frac{1}{(k+1)(k+2)}\lambda^{k+2}C.

In particular,

\displaystyle[T^{(1)}\,_{\lambda}T^{(0)}]=(\partial+\lambda)T^{(0)}+\frac{1}{2}\lambda^{2}C,

which illustrates together with (2.17) that coefficients of formal distributions $T^{(0)}$ , $T^{(1)}$ , and of their derivatives generate the Heisenberg-Virasoro subalgebra in $\mathcal{W}_{1+\infty}$ , and that $T^{(0)}$ and $T^{(1)}$ are conformal vectors of the Virasoro algebra action on $\mathcal{W}_{1+\infty}$ with conformal weights $1$ and $2$ accordingly.

3. Presentation of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ through fermions

3.1. Clifford algebra of charged free fermions

It is well-known that the Lie algebra $\hat{a}_{\infty}$ and its subalgebra $\mathcal{W}_{1+\infty}$ can be realized through the Clifford algebra of charged free fermions. We review this presentation in the language of formal distributions, [14],[15].

Definition 3.1.

The Clifford algebra of charged free fermions is generated by elements $\{\psi^{\pm}_{i}\}_{i\in{\mathbb{Z}}+1/2}$ that satisfy relations

\displaystyle\psi_{k}^{\pm}\psi_{l}^{\pm}+\psi_{l}^{\pm}\psi_{k}^{\pm}=0,\quad\psi_{k}^{+}\psi_{l}^{-}+\psi_{l}^{-}\psi_{k}^{+}=\delta_{k,-l},\quad k,l\in{\mathbb{Z}}+1/2.

Collecting generators as coefficients of formal distributions $\psi^{\pm}(u)=\sum_{i\in{\mathbb{Z}}+1/2}\psi^{\pm}_{i}u^{-i-1/2}$ brings commutation relations into the form

	$\displaystyle\psi^{\pm}(u)\psi^{\pm}(v)+\psi^{\pm}(v)\psi^{\pm}(u)$	$\displaystyle=0,$		(3.1)
	$\displaystyle\psi^{+}(u)\psi^{-}(v)+\psi^{-}(v)\psi^{+}(u)$	$\displaystyle=\delta(u,v).$		(3.2)

Given an associative superalgebra $A=A_{0}\oplus A_{1}$ , the parity of a homogeneous element $a\in A_{\alpha}$ , $\alpha\in{\mathbb{Z}}/2{\mathbb{Z}}$ is defined as $p(a)=\alpha$ . Consider a pair of homogeneous formal distributions $a(z)=\sum_{n\in{\mathbb{Z}}}a_{(n)}z^{-n-1}$ and $b(w)=\sum_{n\in{\mathbb{Z}}}b_{(n)}w^{-n-1}$ with coefficients in $A$ . Recall the definition of the normal ordered product $:a(z)b(w):$ (see e.g. [14], [15]).

Definition 3.2.

The normal ordered product $:a(z)b(w):$ is a formal distribution

:a(z)b(w):=a(z)_{+}b(w)+(-1)^{p(a)p(b)}b(w)a(z)_{-},

where $a(z)_{+}=\sum_{n\leq-1}a_{(n)}z^{-n-1}$ , $a(z)_{-}=\sum_{n\geq 0}a_{(n)}z^{-n-1}$ , and $p(a)$ , $p(b)$ are parities of coefficients of $a(z)$ and $b(z)$ respectively.

Relations between normal ordered products and regular products of charged free fermions can be established by direct calculations.

Lemma 3.1.

	$\displaystyle:\psi^{\pm}(u)\psi^{\pm}(v):=\psi^{\pm}(u)\psi^{\pm}(v),$
	$\displaystyle:\psi^{+}(u)\psi^{-}(v):=-\psi^{-}(v)\psi^{+}(u)+i_{u/v}\left(\frac{1}{v-u}\right)=\psi^{+}(u)\psi^{-}(v)-i_{v/u}\left(\frac{1}{u-v}\right).$

3.2. Presentation through charged free fermions

Proposition 3.1.

The algebra $\hat{a}_{\infty}$ can be realized through charged free fermions by

\displaystyle T(u,w)\to:\psi^{+}(u)\psi^{-}(w):\quad\text{ and}\quad C\to Id.

(3.3)

Accordingly, the algebra $\mathcal{W}_{1+\infty}$ can be realized as a subalgebra of the Clifford algebra of charged free fermions by

\displaystyle T^{(k)}(u)\to:\partial^{k}\psi^{+}(u)\,\psi^{-}(u):.

(3.4)

Proof.

Using Lemma 3.1, we substitute the commutator of normal ordered products by commutator of regular products, rearrange terms to compute their anti-commutators and substitute back by normal ordered products. Note that all products of formal distributions in calculations below are well-defined since they involve formal distributions of different variables $u,v,w,z$ .

	$\displaystyle[:\psi^{+}(u)$	$\displaystyle\psi^{-}(w):,:\psi^{+}(v)\psi^{-}(z):]=[\psi^{+}(u)\psi^{-}(w),\psi^{+}(v)\psi^{-}(z)]$
	$\displaystyle=$	$\displaystyle\psi^{+}(u)\psi^{-}(w)\psi^{+}(v)\psi^{-}(z)-\psi^{+}(v)\psi^{-}(z)\psi^{+}(u)\psi^{-}(w)$
	$\displaystyle=$	$\displaystyle-\psi^{+}(u)\psi^{+}(v)\psi^{-}(w)\psi^{-}(z)+\psi^{+}(u)\psi^{-}(z)\delta(w,v)$
		$\displaystyle+\psi^{+}(v)\psi^{+}(u)\psi^{-}(z)\psi^{-}(w)-\psi^{+}(v)\psi^{-}(w)\delta(z,u)$
	$\displaystyle=$	$\displaystyle\quad\psi^{+}(u)\psi^{-}(z)\delta(w,v)-\psi^{+}(v)\psi^{-}(w)\delta(z,u)$
	$\displaystyle=$	$\displaystyle:\psi^{+}(u)\psi^{-}(z):\delta(w,v)-:\psi^{+}(v)\psi^{-}(w):\delta(z,u)+\gamma(u,w,v,z),$

where $\gamma(u,w,v,z)$ is given by (2.5). Hence, commutator $[:\psi^{+}(u)\psi^{-}(w):,:\psi^{+}(v)\psi^{-}(z):]$ matches the commutator (2.2), and we get (3.3). Since derivation $\partial_{u}$ is compatible with the normal ordered product operation, and we obtain (3.4). ∎

Remark 3.1.

Under identification in Proposition 3.1,

\hat{E}_{ij}\to:\psi^{+}_{-i+1/2}\psi^{-}_{j-1/2}:=\begin{cases}\quad\psi^{+}_{-i+1/2}\psi^{-}_{j-1/2},&i\geq 1,\\ -\psi^{-}_{j-1/2}\psi^{+}_{-i+1/2}=\psi^{+}_{-i+1/2}\psi^{-}_{j-1/2}-\delta_{i,j},&i\leq 0,\end{cases}

and

\displaystyle T^{(k)}_{r}

\displaystyle=\sum_{j\geq 1-r+}(r+j-1)_{k}\psi^{+}_{-r-j+1/2}\psi^{-}_{j-1/2}-\sum_{j\leq-r}(r+j-1)_{k}\psi^{-}_{j-1/2}\psi^{+}_{-r-j+1/2}.

Proposition 3.2.

	$\displaystyle[T(u,v),\psi^{+}(z)]$	$\displaystyle=\psi^{+}(u)\delta(v,z),\quad\quad\quad\quad[T(u,v),\psi^{-}(z)]=-\psi^{-}(v)\delta(u,z),$
	$\displaystyle[T^{(k)}(u),\psi^{+}(z)]$	$\displaystyle=\partial^{k}_{u}\psi^{+}(u)\,\delta(u,z),\quad\quad\quad[T^{(k)}(u),\psi^{-}(z)]=-\psi^{-}(u)\,\partial_{u}^{k}\delta(u,z).$

Proof.

These commutation relations immediately follow from Lemma 3.1 with substitution of normal ordered products by regular products. For example,

	$\displaystyle[T(u,v),\psi^{\pm}(z)]$	$\displaystyle=[:\psi^{+}(u)\psi^{-}(v):,\psi^{\pm}(z)]=[\psi^{+}(u)\psi^{-}(v),\psi^{\pm}(z)]$
		$\displaystyle=\psi^{+}(u)\psi^{-}(v)\psi^{\pm}(z)-\psi^{\pm}(z)\psi^{+}(u)\psi^{-}(v),$

and then relations (3.1, 3.2) are applied. ∎

Remark 3.2.

The last two relations of Proposition (3.2) describe the action of conformal Lie algebra of $\mathcal{W}_{1+\infty}$ on the space of charged free fermions:

\displaystyle[T^{(k)}_{\lambda}\psi^{+}]=\partial^{k}\psi^{+},\quad\quad[T^{(k)}_{\lambda}\psi^{-}]=(-1)^{k+1}(\lambda+\partial)^{k}\psi^{-}.

4. Review of properties of symmetric functions

4.1. On the boson-ferimon correspondance

In this section we construct the action of Lie algebras $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ on a vector space that contains countably many copies of the ring of symmetric functions. This construction is based on the discussed in Section 3 presentation of Lie algebras through charged free fermions and on the boson-fermion correspondence. We only briefly outline the main idea of this important isomorphism of modules of infinite-dimensional algebraic structures, referring for more details to many books and papers, such as e.g.[14], [15], [22]. For our purposes we will need only the resulting from the boson-fermion correspondence formulas (5.3), (5.4) of action of charged free fermions on the ring of symmetric functions.

The boson-fermion correspondence establishes an isomorphism of two infinite-dimensional vector spaces as modules over several important algebraic structures. The first of the two vector spaces is the space of semi-infinite wedge forms, usually called the fermionic Fock space. The second one is the space of countably many copies of a polynomial ring of infinitely many variables, called the bosonic Fock space. It turns out that the Clifford algebra of charged free fermions, Heisenberg algebra, Virasoro algebra act on both spaces in an equivalent way, and moreover, the actions of these algebraic structures are closely related to each other.

In particular, the Clifford algebra of charged free fermions has a natural action on the fermionic Fock space of semi-infinte wedge forms [15]. The boson-fermion correspondence transports this action into the action on the boson Fock space. An important step of this transition is the identification of each polynomial ring that constitute graded components of the boson Fock space with the ring of symmetric functions. This way charged free fermions act on the space of countably many copies of the ring of symmetric functions. We recall the explicit formulas of this action below in Section 5.1.

Presentations of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ through charged free fermions provides actions of these two Lie algebras on the space of countably many copies of the ring of symmetric functions. Our goal is to describe properties of these actions in terms of generating functions, which is done in Section 6. In the rest of this section we review the necessary facts about symmetric functions.

4.2. Symmetric functions

We review properties of symmetric functions [20], [27] in the setup and notations similar to [10], [16], [23], [25]. Consider the algebra of formal power series ${\mathbb{C}}[[{\bf x}]]={\mathbb{C}}[[x_{1},x_{2},\dots]]$ . Let $\lambda=(\lambda_{1}\geq\dots\geq\lambda_{l}>0)$ be a partition of length $l$ . The monomial symmetric function is a formal series

m_{\lambda}=\sum_{(i_{1},\dots,i_{l})\in\mathbb{N}^{l}}x^{\lambda_{1}}_{i_{1}}\dots x^{\lambda_{l}}_{i_{l}}.

Let $\Lambda$ be the subalgebra of ${\mathbb{C}}[[{\bf x}]]$ spanned as a vector space by all monomial symmetric functions. It is called the algebra of symmetric functions. Note that elements of $\Lambda$ are invariant with respect to any permutation of a finite number of indeterminates $x_{1},x_{2},\,\dots$ . The following families of symmetric functions play important role in our study.

For a partition $\lambda=(\lambda_{1}\geq\dots\geq\lambda_{l}>0)$ , Schur symmetric function $s_{\lambda}$ is defined as

\displaystyle s_{\lambda}(x_{1},x_{2},\dots)=\sum_{T}{\bf x}^{T},

(4.1)

where the sum is over all semi-standard tableaux of shape $\lambda$ .

Complete symmetric functions $h_{k}=s_{(k)}$ , elementary symmetric functions $e_{k}=s_{(1^{k})}$ , and power sums $p_{k}$ are defined by the formulas

$\displaystyle h_{k}$	$\displaystyle=\sum_{i_{1}\leq i_{2}\leq\dots\leq i_{k}}x_{i_{1}}x_{i_{2}}\dots x_{i_{k}},\quad$	(4.2)
$\displaystyle e_{k}$	$\displaystyle=\sum_{i_{1}<i_{2}<\dots<i_{k}}x_{i_{1}}x_{i_{2}}\dots x_{i_{k}},\quad\quad$	(4.3)
$\displaystyle p_{k}$	$\displaystyle=\sum_{i\geq 1}x_{i}^{k}.$	(4.4)

Set $h_{-k}(x_{1},x_{2}\dots)=e_{-k}(x_{1},x_{2}\dots)=p_{-k}(x_{1},x_{2}\dots)=0$ for $k\in{\mathbb{N}}$ and $h_{0}=e_{0}=p_{0}=1$ .

The algebra $\Lambda$ is a polynomial algebra in each of these three families of generators:

\Lambda={\mathbb{C}}[h_{1},h_{2},\dots]={\mathbb{C}}[e_{1},e_{2},\dots]={\mathbb{C}}[p_{1},p_{2},\dots].

Schur symmetric functions $\{s_{\lambda}\}$ labeled by all partitions form a linear basis of $\Lambda$ . They can be also expressed through complete symmetric functions by the Jacobi - Trudi identity

\displaystyle s_{\lambda}=\det[h_{\lambda_{i}-i+j}]_{1\leq i,j\leq l}.

(4.5)

We will use (4.5) as the extension of the definition of $s_{\lambda}$ assuming that $\lambda$ can be any integer vector. Then for any $\lambda\in{\mathbb{Z}}^{l}$ either $s_{\lambda}$ coincides up to a sign with a Schur symmetric function (4.1) labeled by a partition, or equals zero.

Generating series of the above families of symmetric functions are elements of $\Lambda[[u]]$ that are given by

\displaystyle H(u)=\sum_{k\geq 0}{h_{k}}{u^{k}}

\displaystyle=\prod_{i\in{\mathbb{N}}}\frac{1}{1-x_{i}u},\quad E(u)=\sum_{k\geq 0}{e_{k}}{u^{k}}=\prod_{i\in{\mathbb{N}}}{(1+x_{i}u)}.

(4.6)

We also denote $P(u)=\sum_{k\geq 1}{p_{k}}{u^{k-1}}$ .

Most of the time we will not recall dependence (4.2)-(4.4) of symmetric functions on the original variables of $x_{i}$ ’s, but treat them as a polynomials in one of families of generators, $\{p_{i}\}$ , or $\{e_{i}\}$ , or $\{h_{i}\}$ . In some cases below we still need to specify the set of indeterminates $(x_{1},x_{2}\dots)$ as in definitions (4.2) - (4.4). Then we write $h_{k}=h_{k}(x_{1},x_{2},\dots)$ , $e_{k}=e_{k}(x_{1},x_{2},\dots)$ , $p_{k}=p_{k}(x_{1},x_{2},\dots)$ , $H(u;x_{1},x_{2},\dots)$ , $E(u;x_{1},x_{2},\dots)$ , $P(u;x_{1},x_{2},\dots)$ , etc.

There is a natural scalar product on $\Lambda$ , where the set of Schur symmetric functions $\{s_{\lambda}\}$ labeled by partitions $\lambda$ form an orthonormal basis, $\langle s_{\lambda},s_{\mu}\rangle=\delta_{\lambda,\mu}.$ Then, for any linear operator acting on the vector space $\Lambda$ , one can define the corresponding adjoint operator. In particular, any symmetric function $f\in\Lambda$ defines an operator of multiplication $f:g\mapsto fg$ for any $g\in\Lambda$ . The corresponding adjoint operator $f^{\perp}$ is defined by the standard rule $\langle f^{\perp}g_{1},g_{2}\rangle=\langle g_{1},fg_{2}\rangle$ for all $g_{1},g_{2}\in\Lambda$ .

It is known [20, I.5 Example 3] that $p_{n}^{\perp}=n\frac{\partial}{\partial p_{n}}.$ Since any element $f\in\Lambda$ can be expressed as a polynomial function of power sums

\displaystyle f=F(p_{1},p_{2},p_{3},\dots),

the corresponding adjoint operator $f^{\perp}$ is a polynomial differential operator with constant coefficients

\displaystyle f^{\perp}=F(\partial/\partial p_{1},2\partial/\partial p_{2},3\partial/\partial p_{3},\dots).

In particular, $e_{k}$ and $h_{k}$ are homogeneous polynomials of degree $k$ in $(p_{1},p_{2},p_{3},\dots)$ , so the adjoint operators $e^{\perp}_{k}$ and $h^{\perp}_{k}$ are homogeneous polynomials of degree $k$ in $(\partial/\partial p_{1},2\partial/\partial p_{2},\dots)$ , which implies the following statement.

Lemma 4.1.

For any symmetric function $f\in\Lambda$ there exists a positive integer $N=N(f)$ , such that

e^{\perp}_{l}(f)=0\quad\text{and}\quad h^{\perp}_{l}(f)=0\quad\text{ for all\qquad$l\geq N$}.

Lemma 4.1 guarantees that formal distributions of operators that we use later in this text act locally finitely on the boson Fock space and that the products of these formal distributions of operators are well-defined.

Denote by $\mathcal{D}$ the algebra of differential operators acting on $\Lambda={\mathbb{C}}[p_{1},p_{2},\dots]$ , which consists of finite sums

\sum_{i_{1},\dots,i_{m}}F_{i_{1},\dots i_{m}}(p_{1},p_{2},\dots){\partial_{p_{1}}^{i_{1}}}\dots\partial_{p_{m}}^{i_{m}},

where coefficients $F_{i_{1},\dots i_{m}}(p_{1},p_{2},\dots)$ are polynomials in $(p_{1},p_{2},\dots)$ . Then operators of multiplication $p_{n},h_{n},e_{n}$ , their adjoints $p_{n}^{\perp},h_{n}^{\perp},e_{n}^{\perp}$ along with their products are elements of $\mathcal{D}$ .

We will use the same notation for the generating series of the corresponding multiplication operators: $H(u),E(u),P(u)\in\mathcal{D}[[u]]$ . Similarly, we define $E^{\perp}(u),H^{\perp}(u),P^{\perp}(u)\in\mathcal{D}[[u^{-1}]]$ as

	$\displaystyle E^{\perp}(u)=\sum_{k\geq 0}\frac{e^{\perp}_{k}}{u^{k}},\quad H^{\perp}(u)=\sum_{k\geq 0}\frac{h^{\perp}_{k}}{u^{k}},$		(4.7)
	$\displaystyle\quad P^{\perp}(u)=\sum_{k\geq 1}\frac{p^{\perp}_{k}}{u^{k+1}}=\sum_{k\geq 1}\frac{k{\partial/}{\partial_{p_{k}}}}{u^{k+1}}.$		(4.8)

The properties of operators collected in the next proposition are either well-known or can be easily deduced from well-known properties of symmetric functions [20, I.2].

Proposition 4.1.

In $\mathcal{D}[[u]]$ (resp. in $\mathcal{D}[[u^{-1}]]$ ),

	$\displaystyle H(u)E(-u)=1,\quad H^{\perp}(u)E^{\perp}(-u)=1,$		(4.9)
	$\displaystyle P(-u)={\partial_{u}E(u)}{H(-u)},\quad P(u)={\partial_{u}H(u)}{E(-u)},$		(4.10)
	$\displaystyle\partial_{u}^{k}H(u)=H(u)(\partial_{u}+P(u))^{k-1}(\,P(u)\,),$		(4.11)
	$\displaystyle\partial_{u}^{k}E(u)=E(u)(\partial_{u}+P(-u))^{k-1}(\,P(-u)),$		(4.12)
	$\displaystyle{P^{\perp}(-u)=-\partial_{u}E^{\perp}(u)\,\,H^{\perp}(-u),\quad}P^{\perp}(u)=-\partial_{u}H^{\perp}(u)\,\,E^{\perp}(-u),$		(4.13)
	$\displaystyle H(u)=exp\left(\sum_{k\geq 1}\frac{p_{k}}{k}{u^{k}}\right),\quad E(u)=exp\left(-\sum_{k\geq 1}\frac{(-1)^{k}p_{k}}{k}{u^{k}}\right),$		(4.14)
	$\displaystyle H^{\perp}(u)=exp\left(\sum_{k\geq 1}\frac{\partial}{\partial p_{k}}\frac{1}{u^{k}}\right),\quad E^{\perp}(u)=exp\left(-\sum_{k\geq 1}{(-1)^{k}}\frac{\partial}{\partial p_{k}}\frac{1}{u^{k}}\right).$		(4.15)

In (4.11), (4.12), $(\partial_{u}+P(\pm u))^{k-1}(\,P(\pm u))$ means that differential operator $(\partial_{u}+P(\pm u))^{k-1}$ is applied to formal distribution of operators $P(\pm u)$ .

Lemma 4.2.

We have the following commutation relations in $\mathcal{D}[[u^{-1},v]]$ :

$\displaystyle\left(1-\frac{v}{u}\right)E^{\perp}(u)E(v)$	$\displaystyle=E(v)E^{\perp}(u),$	(4.16)
$\displaystyle\left(1-\frac{v}{u}\right)H^{\perp}(u)H(v)$	$\displaystyle=H(v)H^{\perp}(u),$	(4.17)
$\displaystyle H^{\perp}(u)E(v)$	$\displaystyle=\left(1+\frac{v}{u}\right)E(v)H^{\perp}(u),$	(4.18)
$\displaystyle E^{\perp}(u)H(v)$	$\displaystyle=\left(1+\frac{v}{u}\right)H(v)E^{\perp}(u),$	(4.19)
$\displaystyle P^{\perp}(u)H(v)$	$\displaystyle=H(v)P^{\perp}(u)-\frac{v}{u^{2}}i_{v/u}\left(1-\frac{v}{u}\right)^{-1}H(v),$	(4.20)
$\displaystyle P^{\perp}(u)E(v)$	$\displaystyle=E(v)P^{\perp}(u)+\frac{v}{u^{2}}i_{v/u}\left(1+\frac{v}{u}\right)^{-1}E(v).$	(4.21)

Proof.

Relations (4.16) - (4.19) can be found [20, I.5 Example 29]. For the proof of (4.20), differentiating (4.17), and using (4.13) and (4.9), we get

$\displaystyle P^{\perp}(u)H(v)$	$\displaystyle=$	$\displaystyle-E^{\perp}(-u)\partial_{u}H^{\perp}(u)H(v)$
	$\displaystyle=$	$\displaystyle-E^{\perp}(-u)\left(\frac{v}{u^{2}}i_{v/u}\left(1-\frac{v}{u}\right)^{-2}H(v)H^{\perp}(u)+i_{v/u}\left(1-\frac{v}{u}\right)^{-1}H(v)\partial_{u}H^{\perp}(u)\right)$
	$\displaystyle=$	$\displaystyle-\frac{v}{u^{2}}\left(1-\frac{v}{u}\right)i_{v/u}\left(1-\frac{v}{u}\right)^{-2}H(v)E^{\perp}(-u)H^{\perp}(u)$
		$\displaystyle-\left(1-\frac{v}{u}\right)i_{v/u}\left(1-\frac{v}{u}\right)^{-1}H(v)\partial_{u}H^{\perp}(u)E^{\perp}(-u)$
	$\displaystyle=$	$\displaystyle H(v)\left(-\frac{v}{u^{2}}i_{v/u}\left(1-\frac{v}{u}\right)^{-1}+P^{\perp}(u)\right).$

Relation (4.21) is proved similarly. ∎

Finally, we would like to state a simple corollary of Lemma 4.2 that will be used in Sections 6 and 10.

Corollary 4.1.

In $\mathcal{D}[[u^{-1},u_{1},\dots,u_{l}]]$

	$\displaystyle E^{\perp}(-u)\prod_{i=1}^{l}H(u_{i})$	$\displaystyle=\prod_{s=1}^{l}\left(1-\frac{u_{i}}{u}\right)\prod_{i=1}^{l}H(u_{i})E^{\perp}(-u),$
	$\displaystyle H^{\perp}(u)\prod_{i=1}^{l}H(u_{i})$	$\displaystyle=\prod_{s=1}^{l}i_{u_{i}/u}\left(1-\frac{u_{i}}{u}\right)^{-1}\prod_{i=1}^{l}H(u_{i})H^{\perp}(u),$
	$\displaystyle P^{\perp}(u)\prod_{i=1}^{l}H(u_{i})$	$\displaystyle=\prod_{i=1}^{l}H(u_{i})P^{\perp}(u)-\left(\sum_{i=1}^{l}\frac{u_{i}}{u^{2}}i_{u_{i}/u}\left(1-\frac{u_{i}}{u}\right)^{-1}\right)\prod_{i=1}^{l}H(u_{i}).$

5. Action of charged free fermions on the space of symmetric functions

5.1. Action of charged free fermions on symmetric functions

Consider the ring of symmetric functions $\Lambda$ as the ring of polynomials in power sums $\Lambda={\mathbb{C}}[p_{1},p_{2},\dots]$ . The boson Fock space is the graded space $\mathcal{B}={\mathbb{C}}[z,z^{-1}]\otimes\Lambda$ that consists of countably many copies of $\Lambda$ . The grading $\mathcal{B}=\oplus_{m\in{\mathbb{Z}}}\mathcal{B}^{(m)}$ with $\mathcal{B}^{(m)}=z^{m}\,\Lambda$ is called charge decomposition.

Let $R(u)$ act on the elements of the form $z^{m}f\in\mathcal{B}^{(m)}$ , where $f\in\Lambda$ , $m\in{\mathbb{Z}}$ , by the rule

R(u)(z^{m}f)={z^{m+1}}{u}^{m+1}f.

Then $R^{-1}(u)$ acts as

R^{-1}(u)(z^{m}f)={z}^{m-1}u^{-m}f.

The meaning of the action of $R^{\pm 1}(u)$ is the translation between equivalent symmetric functions as elements in different graded components of $\mathcal{B}$ .

Define formal distributions $\psi^{\pm}(u)$ of operators acting on the space $\mathcal{B}$ through the action of $R^{\pm 1}(u)$ and the $\mathcal{D}$ -valued generating series (4.6), (4.7):

	$\displaystyle\psi^{+}(u)$	$\displaystyle=u^{-1}R(u)H(u)E^{\perp}(-u),$		(5.1)
	$\displaystyle\psi^{-}(u)$	$\displaystyle=R^{-1}(u)E(-u)H^{\perp}(u),$		(5.2)

or, in other words, for any $m\in{\mathbb{Z}}$ and any $f\in\Lambda$ ,

	$\displaystyle\psi^{+}(u)(z^{m}f)$	$\displaystyle=z^{m+1}u^{m}H(u)E^{\perp}(-u)(f),$		(5.3)
	$\displaystyle\psi^{-}(u)(z^{m}f)$	$\displaystyle=z^{m-1}{u^{-m}}E(-u)H^{\perp}(u)(f).$		(5.4)

Let the operators $\{\psi^{\pm}_{i}\}_{i\in{\mathbb{Z}}+1/2}$ be the coefficients of the expansions

\psi^{\pm}(u)=\sum_{i\in{\mathbb{Z}}+1/2}\psi^{\pm}_{i}u^{-i-1/2}.

Proposition 5.1.

Formulas (5.1), (5.2) define quantum fields $\psi^{\pm}(u)$ of operators acting on the space $\mathcal{B}$ that satisfy relations of charged free fermions (3.1), (3.2).

For the original proof we refer to [9]. Alternatively, one can deduce the relations from the definition (5.3, 5.4) and Lemma 4.2 (see [23] for this approach).

Remark 5.1.

From (4.14), (4.15) one immediately gets the bosonic form of $\psi^{\pm}(u)$ :

	$\displaystyle\psi^{+}(u)$	$\displaystyle=u^{-1}R(u)\exp\left(\sum_{n\geq 1}\frac{p_{n}}{n}{u^{n}}\right)\exp\left(-\sum_{n\geq 1}\frac{\partial}{\partial p_{n}}\frac{1}{u^{n}}\right),$
	$\displaystyle\psi^{-}(u)$	$\displaystyle=R^{-1}(u)\exp\left(-\sum_{n\geq 1}\frac{p_{n}}{n}{u^{n}}\right)\exp\left(\sum_{n\geq 1}\frac{\partial}{\partial p_{n}}\frac{1}{u^{n}}\right).$

5.2. Generating function for Schur symmetric functions

Applying commutation relation (4.19) to a product of multiple operators of the form (5.3), we get

\displaystyle\psi^{+}(u_{l})\dots\psi^{+}(u_{1})\,(z^{m}f)=z^{l}\prod_{1\leq i<j\leq l}\left(1-\frac{u_{i}}{u_{j}}\right)\prod_{i=1}^{l}u_{i}^{k+i-1}H(u_{i})\prod_{i=1}^{l}E^{\perp}(-u_{i})(z^{m}f).

In particular, let

S^{(m)}(u_{l},..,u_{1})=\psi^{+}(u_{l})\dots\psi^{+}(u_{1})\,(z^{m})

be the result of action of operators $\psi^{+}(u_{l})\dots\psi^{+}(u_{1})$ on the $m$ -th vacuum vector $z^{m}\in\mathcal{B}^{(m)}$ . Since $E^{\perp}(-u_{i})(z^{m})=z^{m}$ , we can write

\displaystyle S^{(m)}(u_{l},..,u_{1})=z^{m+l}u_{l}^{k}\dots u_{1}^{k}S(u_{l},..,u_{1}),

where ${S}(u_{l},..,u_{1})$ is a formal distribution with coefficients in $\Lambda$ :

\displaystyle{S}(u_{l},..,u_{1})

\displaystyle=\prod_{1\leq i<j\leq l}\left(u_{j}-{u_{i}}\right)\prod_{i=1}^{l}H(u_{i})

(5.5)

In the expansion ${S}(u_{l},..,u_{1})=\sum_{\alpha\in{\mathbb{Z}}}S_{\alpha}u_{1}^{\alpha_{1}}\dots u_{l}^{\alpha_{l}}$ coefficients $S_{\alpha}=\det[h_{\alpha_{i}+1-j}]=0$ if $\alpha_{i}<0$ for any $i$ , which allows us to write this expansion as the sum over $\alpha\in{\mathbb{Z}}^{l}_{\geq 0}$ rather than over all ${\mathbb{Z}}^{l}$ . Also $S_{\alpha}=0$ if $\alpha_{i}=\alpha_{j}$ for some $i\neq j$ . The transition to standard notations of Schur symmetric functions $s_{\lambda}$ is given by formulas

	$\displaystyle s_{(\lambda_{1},\dots,\lambda_{l})}=S_{(\lambda_{1}+l-1,\lambda_{2}+l-2,\dots\lambda_{l})},$
	$\displaystyle e_{k}=s_{(1^{k})}=S_{({\underbrace{l,\,l-1,\,\dots,l-k+1,\,}_{k}}\,\,{\underbrace{l-k-1\,,l-k-2\,,\dots,0\,}_{l-k}})},$
	$\displaystyle h_{k}=s_{(k)}=S_{(k+l-1,\quad{\underbrace{l-k-1,\,l-2,\,\dots,2,\,1,\,0}_{l-1}})}.$

Every coefficient $S_{\alpha}$ is either zero, or coincides up to a sign with one of Schur symmetric functions $s_{\lambda}$ , and every Schur symmetric function $s_{\lambda}$ appears as a coefficient in the expansion of $\mathcal{S}(u_{l},..,u_{1})$ . In that sense $\mathcal{S}(u_{l},..,u_{1})$ can be viewed as the generating function for Schur symmetric functions $\{s_{\lambda}\}_{\lambda\in{\mathbb{Z}}^{l}}$ .

Remark 5.2.

From (5.5) and Cauchy identity (see e.g.[20] I.4 ,(4.3)) in the region $|u_{1}|<|u_{2}|<\dots<|u_{l}|$ ,

\displaystyle{S}(u_{l},..,u_{1})\prod_{1\leq i<j\leq l}i_{u_{i}/u_{j}}\left(u_{j}-{u_{i}}\right)^{-1}=\prod_{i=1}^{l}H(u_{i})=\sum_{\lambda}s_{\lambda}(x_{1},x_{2},\dots)s_{\lambda}(u_{1},u_{2},\dots u_{l}).

5.3. Bosonic normal ordered products for charged free fermions

Following [22], we introduce another version of normal ordered product ** **, defined for polynomial differential operators acting on the boson Fock space. In a product of two formal distributions of such operators this reordering moves all multiplication operators by $p_{i}$ ’s to the left from all differentiations $\partial/\partial p_{i}$ ’s.We we will call it bosonic normal ordered product. In particular, for charged free fermions the bosonic normal ordering is defined as follows:

	$\displaystyle\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\psi^{+}(u)\psi^{-}(v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}(z^{m}f)$	$\displaystyle=z^{m}u^{m}{v^{-m}}H(u)E(-v)E^{\perp}(-u)H^{\perp}(v)(f),$
	$\displaystyle\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\psi^{+}(u)\psi^{+}(v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}(z^{m}f)$	$\displaystyle=z^{m+2}u^{m+1}{v^{m}}H(u)H(v)E^{\perp}(-u)E^{\perp}(-v)(f),$
	$\displaystyle\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\psi^{-}(u)\psi^{-}(v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}(z^{m}f)$	$\displaystyle=z^{m-2}u^{-m+1}{v^{-m}}E(-u)E(-v)H^{\perp}(u)H^{\perp}(v)(f).$

Relations between different products of charged free fermions follow from Lemmas 3.1 and 4.2, and can be summarized as follows:

	$\displaystyle:\psi^{+}(u)\psi^{-}(v):$	$\displaystyle=i_{u/v}\left(\frac{1}{v-u}\right)-\psi^{-}(v)\psi^{+}(u)=i_{u/v}\left(\frac{1}{v-u}\right)(1-\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\psi^{+}(u)\psi^{-}(v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}),$		(5.6)
	$\displaystyle:\psi^{\pm}(u)\psi^{\pm}(v):$	$\displaystyle=\psi^{\pm}(u)\psi^{\pm}(v)=\left(1-\frac{v}{u}\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\psi^{\pm}(u)\psi^{\pm}(v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}.$		(5.7)

We introduce notation $\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,\psi^{+}(u)\psi^{-}(v)\,\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}$ .

Proposition 5.2.

In the bosonic order the generating functions of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ have the form

	$\displaystyle T(u,v)$	$\displaystyle=i_{u/v}\left(\frac{1}{v-u}\right)(1-\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\,T(u,v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}),$		(5.8)
	$\displaystyle T^{(k)}(u)$	$\displaystyle=\frac{1}{k+1}\partial^{k+1}_{u}\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\,T(u,v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\|_{v=u}.$		(5.9)

Proof.

Statement (5.8) immediately follows from (5.6). Note that we cannot evaluate $T^{(k)}(u)$ in the most straight-forward way as a derivative of $-\psi^{-}(v)\psi^{+}(u)+i_{u/v}\left(\frac{1}{v-u}\right)$ , as each of these two formal distributions has a singularity at $u=v$ . However, the bosonic presentation (5.8) allows us to overcome this obstacle. Since the linear polynomial $(v-u)$ is the inverse element for the formal power series $i_{u/v}\left(\frac{1}{v-u}\right)$ , multiplication of both parts of (5.8) by $(v-u)$ results in

\displaystyle(v-u)T(u,v)=1-\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}.

Application of the operator $\partial^{k}_{u}$ to both sides gives the relation

-k\,\partial^{k-1}_{u}T(u,v)+(v-u)\partial^{k}_{u}T(u,v)=-\partial^{k}_{u}\,\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}},

and setting $v=u$ we get (5.9). ∎

6. Action of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ on symmetric functions

6.1. The first main result

In this section we describe the action of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ on symmetric functions through the language of formal distributions those coefficients are operators of multiplication and differentiation on the space of symmetric functions. Remark 6.1 provides a comment on the interesting duality of formula (6.4). To highlight this duality we introduce special notations for products of $\left(1-\frac{u_{i}}{u}\right)$ and $i_{u_{i}/u}\left(1-\frac{u_{i}}{u}\right)^{-1}$ . Set

{\mathcal{E}_{l}}(-u^{-1})={E}(-u^{-1};\,u_{1},\dots,u_{l})=\prod_{s=1}^{l}\left(1-\frac{u_{i}}{u}\right),

and

{\mathcal{H}_{l}(u^{-1})}=\frac{1}{\mathcal{E}_{l}(-u^{-1})}={H}(u^{-1};\,u_{1},\dots,u_{l})=\prod_{i=1}^{l}i_{u_{i}/u}\left(1-\frac{u_{i}}{u}\right)^{-1}.

Observe that ${\mathcal{E}_{l}}(-u^{-1})$ is a polynomial in variables $(u_{1},\dots,u_{l},u^{-1})$ , and ${\mathcal{H}_{l}(u^{-1})}$ is a formal power series in variables $(u_{1},\dots,u_{l},u^{-1})$ . The introduced notations are purposefully compatible with (4.6), providing interpretation of coefficients of ${\mathcal{E}_{l}}(-u^{-1})$ and ${\mathcal{H}_{l}(u^{-1})}$ as symmetric polynomials $e_{k}(u_{1},\dots,u_{l})$ and $h_{k}(u_{1},\dots,u_{l})$ respectively.

Theorem 6.1.

(1)

Let $f\in\Lambda$ . The action of $\hat{a}_{\infty}$ on $z^{m}f\in\mathcal{B}^{(m)}$ is described by

\displaystyle T(u,v)(z^{m}f)=z^{m}i_{u/v}\left(\frac{1}{v-u}\right)\left(1-\frac{u^{m}}{v^{m}}E(-v)H(u)E^{\perp}(-u)H^{\perp}(v)\right)f.

(6.1)

(2)

For any $f\in\mathcal{B}^{(0)}$ the action of $\mathcal{W}_{1+\infty}$ on this element is described by

\displaystyle T^{(k)}(u)\,(f)=\frac{1}{k+1}\sum_{r+s=k+1}{s\choose{k+1}}E(-u)\,\partial^{r}_{u}H(u)\,\,H^{\perp}(u)\,\partial_{u}^{s}(E^{\perp}(-u))\,(f).

(6.2)

(3)

The action of $\hat{a}_{\infty}$ on the basis of Schur symmetric functions is described by the following relation on generating functions.

		$\displaystyle T(u,v)\left(S(u_{l},..,u_{1})^{(m)}\right)$
	$\displaystyle=$	$\displaystyle\quad i_{u/v}\left(\frac{1}{v-u}\right)\left(S(u_{l},..,u_{1})^{(m)}-\frac{1}{zu^{l}v^{m}}{{\mathcal{H}_{l}(v^{-1})}}E(-v)S(u,u_{l},..,u_{1})^{(m)}\right).$		(6.3)

(4)

The action of $\mathcal{W}_{1+\infty}$ on the basis of Schur symmetric functions is described by the following relation on generating functions.

$\displaystyle T^{(k)}(u)\,\left(S(u_{l},..,u_{1})\right).$ (6.4)

Proof.

(1)

Action (6.1) is immediate consequence of (5.8).

(2)

Singularities of (6.1) at $v=u$ prevent us from using that formula directly to evaluate the action of $T^{(k)}(u)$ on symmetric functions. Instead we use (5.9):

	$\displaystyle T^{(k)}(u)(f)$	$\displaystyle=\frac{1}{k+1}\partial^{k+1}_{u}\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\,T(u,v)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\|_{v=u}(f)$
		$\displaystyle=\frac{1}{k+1}\partial^{k+1}_{u}\left(H(u)E(-v)E^{\perp}(-u)H^{\perp}(v)\right)\|_{v=u}(f).$

Then the Leibniz rule implies (6.2).

(3)

By (5.5), Corollary 4.1 and the fact that $H^{\perp}(v)(z^{r})=E^{\perp}(-u)(z^{r})=z^{r}$ , one gets

\displaystyle H^{\perp}(v)S(u_{l},\dots,u_{1})^{(m)}={\mathcal{H}_{l}(v^{-1})}S(u_{l},\dots,u_{1})^{(m)},

(6.5)

and

	$\displaystyle H(u)E^{\perp}(-u)S(u_{l},\dots,u_{1})^{(m)}$
	$\displaystyle=z^{m+l}u_{l}^{m}\dots u_{1}^{m}\prod_{1\leq i<j\leq l}\left(u_{j}-{u_{i}}\right)H(u)\prod_{i=1}^{l}\left(1-\frac{u_{i}}{u}\right)\;\prod_{i=1}^{l}H(u_{i})$
	$\displaystyle={z^{-1}u^{-l-m}}S(u,u_{l},\dots,u_{1})^{(m)}.$

Substitution of

	$\displaystyle\frac{u^{m}}{v^{m}}$	$\displaystyle E(-v)H(u)E^{\perp}(-u)H^{\perp}(v)S(u_{l},\dots,u_{1})^{(m)}$
		$\displaystyle=\frac{\mathcal{H}_{l}(v^{-1})E(-v)}{zu^{l}v^{m}}S(u,u_{l},\dots,u_{1})^{(m)}$

in (6.1) proves (3).

(4)

Differentiation of the first equality in Corollary 4.1 implies

\partial_{u}^{s}(E^{\perp}(-u))S(u_{l},\dots,u_{1})=\partial_{u}^{s}\mathcal{E}_{l}(-u^{-1})S(u_{l},\dots,u_{1}).

Using that $E(-u)=H(u)^{-1}$ , ${\mathcal{E}_{l}}(-u^{-1})={\mathcal{H}_{l}}(u^{-1})$ and (6.6), we get from (3)

	$\displaystyle T^{(k)}(u)$	$\displaystyle\,\left(S(u_{l},..,u_{1})\right)$
		$\displaystyle=\frac{1}{k+1}\sum_{s+r=k+1}{s\choose{k+1}}E(-u)\partial^{r}_{u}H(u){{\mathcal{H}_{l}}(u^{-1})}{\partial_{u}^{s}{\mathcal{E}_{l}}(-u^{-1})}S(u_{l},..,u_{1})$
		$\displaystyle=\frac{\,\partial_{u}^{k+1}({\mathcal{E}_{l}}(-u^{-1})H(u)\,)}{(k+1){\mathcal{E}_{l}}(-u^{-1})H(u)}S(u_{l},..,u_{1}).$

∎

Remark 6.1.

Formula (6.4) carries remarkable properties, including a “duality symmetry”.

First, observe that the action of $\hat{a}_{\infty}$ by (3) and the action of $\mathcal{W}_{1+\infty}$ by (6.4) are expressed through multiplication operators by symmetric functions $h_{k}(x_{1},x_{2},\dots)$ and symmetric polynomials $e_{m}(u_{1},\dots u_{l})$ , and contain no differentiation operators (compare, for example, with (6.2) that involves application of differential operators $h_{k}^{\perp}(x_{1},x_{2},\dots)$ and $e_{k}^{\perp}(x_{1},x_{2},\dots)$ ).

Second, (6.4) interprets formal distribution $S(u_{1},..,u_{l})$ as an algebraic expression

S(u_{1},..,u_{l})\in{\mathbb{C}}[[x_{1},x_{2},\dots]]\otimes{\mathbb{C}}[[u_{1},\dots u_{l}]],

in other words, composed of products of symmetric functions $s_{\alpha}=s_{\lambda}(x_{1},x_{2},\dots)\in\Lambda\subset{\mathbb{C}}[[x_{1},x_{2},\dots]]$ and monomials $u_{1}^{\alpha_{1}}\dots u_{l}^{\alpha_{l}}\in{\mathbb{C}}[u_{1},\dots u_{l}]$ , where both components are treated on equal footing. In this interpretation the action of $T^{(k)}(u)$ reveals certain duality: it is expressed through action on the “ ${\mathbb{C}}[[x_{1},x_{2},\dots]]$ part” by multiplication operators combined in $H(u;x_{1},\dots,x_{l})$ , and on the “ ${\mathbb{C}}[u_{1},\dots u_{l}]$ part” by action of multiplication operators combined in $\mathcal{E}_{l}(u;u_{1},\dots,u_{l})$ . Note also that operators “ $H(u)$ ” and “ $\mathcal{E}_{l}(u)$ ” acting in different components can be considered of “dual nature” due to relations of the kind (4.9). The observed here combinatorial duality are naturally connected to Remark 5.2.

6.2. Actions of Heisenberg and Virasoro algebra

Example 6.1.

Formula (6.4) interprets the action of Heisenberg algebra action on the basis Schur symmetric functions as a result of multiplication of the following formal distributions:

\displaystyle T^{(0)}(u)\,\left(S(u_{1},..,u_{l})\right)=

\displaystyle\left(\sum_{k\geq 1}{p_{k}}u^{k-1}+\sum_{i=1}^{l}i_{u_{i}/u}\left(\frac{u_{i}}{u(u-u_{i})}\right)\right)\,S(u_{1},..,u_{l}).

(6.6)

Indeed, from (6.4) and (4.10) we get

	$\displaystyle T^{(0)}(u)\,\left(S(u_{1},..,u_{l})\right)$	$\displaystyle=\frac{\mathcal{E}_{l}(-u^{-1})\partial_{u}H(u)+\partial_{u}\mathcal{E}_{l}(-u^{-1})H(u)}{\mathcal{E}_{l}(-u^{-1})H(u)}\left(S(u_{1},..,u_{l})\right)$
		$\displaystyle=\left(\frac{\partial_{u}H(u)}{H(u)}+\frac{\partial_{u}\mathcal{E}_{l}(-u^{-1})}{\mathcal{E}_{l}(-u^{-1})}\right)\left(S(u_{1},..,u_{l})\right)$
		$\displaystyle=\left(P(u)+\frac{1}{u^{2}}{\mathcal{P}_{l}}(u^{-1})\right)S(u_{1},..,u_{l})$
		$\displaystyle=\left(\sum_{k\geq 1}{p_{k}}u^{k-1}+\sum_{k\geq 1}\sum_{i=1}^{l}\frac{u_{i}^{k}}{u^{k+1}}\right)S(u_{1},..,u_{l}).$

Example 6.2.

From (6.4), (4.11), and (4.12), we obtain the action of Virasoro algebra on Schur symmetric functions in a form of multiplication of the following formal distributions:

	$\displaystyle T^{(1)}(u)\,\left(S(u_{1},..,u_{l})\right)=\left(\frac{\partial^{2}_{u}H(u)}{2H(u)}+\frac{\partial_{u}\mathcal{E}_{l}(-u^{-1})\,\partial_{u}H(u)}{\mathcal{E}_{l}(-u^{-1})H(u)}+\frac{\partial^{2}_{u}\mathcal{E}_{l}(-u^{-1})}{2\mathcal{E}_{l}(-u^{-1})}\right)S(u_{1},..,u_{l})$
	$\displaystyle=\left(\frac{\partial_{u}P(u)}{2}+\frac{P(u)^{2}}{2}+\,\frac{{\mathcal{P}_{l}}(u^{-1})P(u)}{u^{2}}+\frac{\partial_{u}\mathcal{P}_{l}(u^{-1})}{2u^{4}}+\frac{\mathcal{P}_{l}(u^{-1})^{2}}{2u^{4}}\right)S(u_{1},..,u_{l})$
	$\displaystyle=\Bigl(\frac{1}{2}\partial_{u}P(u)+\frac{1}{2}P(u)^{2}+P(u)\sum_{i=1}^{l}i_{u_{i}/u}\left(\frac{u_{i}}{u(u-u_{i})}\right)-\sum_{i=1}^{l}i_{u_{i}/u}\left(\frac{u_{i}}{2u^{4}(u-u_{i})^{2}}\right)$
	$\displaystyle+\sum_{i,j=1}^{l}i_{u_{i}u_{j}/u}\left(\frac{u_{i}u_{j}}{2u^{2}(u-u_{i})(u-u_{j})}\right)\Bigr)S(u_{1},..,u_{l}).$

In both Examples 6.1 and 6.2 the explicit action of the a particular generator $T^{(0)}_{k}$ or $T^{(0)}_{k}$ on a particular Schur symmetric function can be recovered by application of the Murnaghan–Nakayama rule $p_{r}\cdot s_{\lambda}=\sum_{\mu}(-1)^{ht(\mu/\lambda)+1}s_{\mu},$ where the sum is over all partitions $\mu$ such that $\mu/\lambda$ is a rim-hook of size $r$ and $ht(\mu/\lambda)$ is the number of rows in the diagram $\mu/\lambda$ , ([27], 7.17).

Example 6.3.

Introduce a formal distribution

J(u)=P(u)+P^{\perp}(u)=\sum_{k\geq 1}{p_{k}}{u^{k-1}}+\sum_{k\geq 1}k{\partial/}{\partial_{p_{k}}}{u^{-k-1}}.

Formula (6.7) is often used as the definition of the formal distribution of generators $T^{(k)}(u)$ . We deduce it here as a corollary of (6.2).

Proposition 6.1.

\displaystyle T^{(k)}(u)=\frac{1}{k+1}\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J(u))^{k}(J(u))\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}.

(6.7)

Remark 6.2.

Formula (6.7) should be understood as that the operator $(\partial_{u}+J(u))^{k}$ is applied to the formal distribution of multiplication operators $J(u)$ , and then the resulting formal distribution of operators acting on the space of symmetric functions is put in the bosonic normal order. For example,

\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J(u))(J(u))\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=\partial_{u}P(u)+(P(u))^{2}+2P(u)P^{\perp}(u)+\partial_{u}P^{\perp}(u)+(P^{\perp}(u))^{2}.

Proof.

We will prove (6.7) by induction. Base of induction:

T^{(0)}(u)=P(u)+P^{\perp}(u)=J(u).

Assume that (6.7) holds for some $k$ . Then, by induction assumption,

\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J)^{k+1}(J)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J)(\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J)^{k}J\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}})\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=(k+1)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J)\left(T^{(k)}(u)\right)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}.

By (6.2), (4.11) and (4.12),

\displaystyle T^{(k)}(u)=\frac{1}{k+1}\sum_{\begin{subarray}{c}r+s=k+1\\ r,s\geq 0\end{subarray}}{{k+1}\choose{s}}\,(\partial_{u}+P(u))^{r-1}(P(u))\,(\partial_{u}+P^{\perp}(u))^{s-1}(P^{\perp}(u)),

where for the uniform formula with some minor abuse of notations we assume $(\partial_{u}+P^{\perp}(u))^{-1}(P^{\perp}(u))=(\partial_{u}+P(u))^{-1}(P(u))=1$ . By Leibniz rule,

	$\displaystyle\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}(\partial_{u}+P+P^{\perp})\,\left((\partial_{u}+P)^{r-1}(P)\,(\partial_{u}+P^{\perp})^{s-1}(P^{\perp})\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}$
	$\displaystyle=\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\left((\partial_{u}+P)^{r}(P)\,(\partial_{u}+P^{\perp})^{s-1}(P^{\perp})\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}+\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\left((\partial_{u}+P^{r-1}(P)\,(\partial_{u}+P^{\perp})^{s}(P^{\perp})\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}.$

Hence, the coefficient of the term $(\partial_{u}+P)^{r-1}(P)\,(\partial_{u}+P^{\perp})^{s-1}(P^{\perp})$ in
$(k+1)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}(\partial_{u}+J)\,\left(T^{(k)}(u)\right)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}$ is ${{k+1}\choose{s}}+{{k+1}\choose{s-1}}={{k+2}\choose{s}}$ , and

	**	$\displaystyle(\partial_{u}+J)^{k+1}(J)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}=(k+1)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}(\partial_{u}+J)\left(T^{(k)}(u)\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}$
		$\displaystyle=\sum_{\begin{subarray}{c}r+s=k+2,\\ r,s\geq 0\end{subarray}}{{k+2}\choose{s}}\,(\partial_{u}+P(u))^{r-1}(P(u))\,(\partial_{u}+P^{\perp}(u))^{s-1}(P^{\perp}(u))=(k+2)T^{(k+1)}(u).$

∎

6.3. Some connections to existing results

As it is mentioned in the Introduction, the explicit formulas for the action of operators from $\mathcal{W}_{1+\infty}$ were studied by several authors. In particular, in [19] explicit formulas for the action of generators $T^{(k)}_{r}$ on Schur symmetric functions $S_{\mu}$ were proved. The authors of [19] also provided a list of references to the previous works that computed particular cases of actions of such operators. From those we would like to mention that combinatorial description [5] of the action of the Heisenberg algebra on symmetric functions is the most close to our approach, c.f. Example 6.1. Also we would like to outline the framework of formal distributions behind the explicit formulas obtained in [19] connecting it with the set up of this note. Namely, by (5.6), $\partial_{u}^{k}T(u,v)=-\psi^{-}(v)\partial_{u}^{k}\psi^{+}(u)+\partial_{u}^{k}i_{u/v}\left(\frac{1}{v-u}\right),$ and a variation of calculation in the proof of (6.4), followed by the expansion of formal distributions, gives

		$\displaystyle\partial_{u}^{k}T(u,v)S(u_{l},\dots,u_{1})$
	$\displaystyle=$	$\displaystyle-\psi^{-}(v)\partial_{u}^{k}\psi^{+}(u)\psi^{+}(u_{l})\dots\psi^{+}(u_{1})(1)+\partial_{u}^{k}i_{u/v}\left(\frac{1}{v-u}\right)\psi^{+}(u_{l})\dots\psi^{+}(u_{1})(1)$
	$\displaystyle=$	$\displaystyle-\psi^{-}(v)\partial_{u}^{k}S(u,u_{l},\dots,u_{1})+\partial_{u}^{k}i_{u/v}\left(\frac{1}{v-u}\right)S(u_{l},\dots,u_{1})$
	$\displaystyle=$	$\displaystyle-\sum_{s,a\in{\mathbb{Z}},\,\mu\in{\mathbb{Z}}^{l}}(a)_{k}\psi^{-}_{s-1/2}(S_{(\mu,a)})v^{-s}u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}u^{a-k}$
		$\displaystyle+\sum_{a=0}^{\infty}\sum_{\mu\in{\mathbb{Z}}^{l}}({a)}_{k}S_{\mu}u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}{u^{a-k}}{v^{-a-1}}$
	$\displaystyle=$	$\displaystyle\sum_{s,a\in{\mathbb{Z}},\,\mu\in{\mathbb{Z}}^{l}}(a)_{k}\left(-\psi^{-}_{s-1/2}(S_{\mu,a})+\delta_{a\geq 0}\delta_{a,s+1}S_{\mu}\right)u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}u^{a-k}v^{-s}.$

Evaluating at $v=u$ we obtain

	$\displaystyle T(u)^{(k)}S(u_{l},\dots,u_{1})$
	$\displaystyle=\sum_{r\in{\mathbb{Z}},\,\mu\in{\mathbb{Z}}^{l}}\sum_{s\in{\mathbb{Z}}}(r+s)_{k}\left(-\psi^{-}_{s-1/2}(S_{\mu,r+s})+\delta_{r+s\geq 0}\delta_{r,1}S_{\mu}\right)u^{r-k}u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}$
	$\displaystyle=\sum_{r\in{\mathbb{Z}},\,\mu\in{\mathbb{Z}}^{l}}\left(-\sum_{s\in{\mathbb{Z}}}(r+s)_{k}\psi^{-}_{s-1/2}(S_{\mu,r+s})+\delta_{r,1}\sum_{s\geq-1}(s+1)_{k}S_{\mu}\right)u^{r-k}u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}.$

Substitution of the combinatorial expression for the value of the action $\psi^{-}_{s-1/2}(S_{\mu,r+s})$ of charged free fermions on the Schur symmetric function in this formula gives the value of the coefficient of the monomial $u^{r-k}u_{1}^{\mu_{1}}\dots u_{l}^{\mu_{l}}$ and describes the action of $T^{(k)}_{r}$ on $S_{\mu}$ by a combinatorial formula equivalent to [19], where all calculations are carefully performed and the final result is provided.

7. Lie algebra $\mathcal{W}^{B}_{1+\infty}$ as a subalgebra of $\hat{a}_{\infty}$

7.1. Lie algebra $\hat{o}_{\infty}$

The B-type analogue of $\mathcal{W}_{1+\infty}$ was studied in [2, 19, 28], and we follow these papers for definitions. Define an anti-involution $i$ on $a_{\infty}$ by its action on generators

i(E_{kl})=(-1)^{k+l}E_{-l,-k},

or, equivalently, by the transformation of the matrix of generators

i(T(u,w))=-\frac{w}{u}T(-w,-u).

Let

\displaystyle T^{B}(u,w)=uT(u,-w)-wT(w,-u)=u(T(u,-w)-i(T(u,-w)).

(7.1)

Then

\displaystyle i(T^{B}(u,w))=-T^{B}(u,w)=T^{B}(w,u),\quad\text{and}\quad T^{B}(u,u)=0.

Coefficients of expansion of $T^{B}(u,w)$ are given by

T^{B}(u,w)=\sum_{i,j\in{\mathbb{Z}}}(-1)^{j}\hat{F}_{ij}u^{i}w^{-j},\quad\text{where}\quad\hat{F}_{ij}=\hat{E}_{ij}-(-1)^{i+j}\hat{E}_{-j\,-i}.

Proposition 7.1.

$[T^{B}(u,w),C]=0$ , and

	$\displaystyle[T^{B}(u,w),T^{B}(v,z)]$	$\displaystyle=2v\delta(v,-w)\left(T^{B}(u,z)+i_{z/u}\left(\frac{2u}{u+z}\right)C\right)$
		$\displaystyle-2v\delta(v,-u)\left(T^{B}(w,z)+i_{z/w}\left(\frac{2w}{w+z}\right)C\right)$
		$\displaystyle+\,2z\delta(z,-w)\left(T^{B}(v,u)+i_{u/v}\left(\frac{2v}{v+u}\right)C\right)$
		$\displaystyle-2z\delta(z,-u)\left(T^{B}(v,w)+i_{w/v}\left(\frac{2v}{v+w}\right)C\right).$

Proof.

Statement follows from (7.1) and commutation relations (2.2) in $\hat{a}_{\infty}$ . ∎

Remark 7.1.

The central part of this commutation relation can be expressed through $\gamma(u,w,v,z)$ defined by (2.5) as

	$\displaystyle 2v\delta(v,-w)i_{z/u}\left(\frac{2u}{u+z}\right)-2v\delta(v,-u)i_{z/w}\left(\frac{2w}{w+z}\right)$
	$\displaystyle+\,2z\delta(z,-w)i_{u/v}\left(\frac{2v}{v+u}\right)-2z\delta(z,-u)i_{w/v}\left(\frac{2v}{v+w}\right)$
	$\displaystyle\quad=4vu\gamma(u,-w,v-z)+4vw\gamma(-w,-u,v,z).$

Let $o_{\infty}=\{a\in a_{\infty}|i(a)=-a\}$ , and let $\hat{o}_{\infty}=o_{\infty}\oplus{\mathbb{C}}C$ . Together with commutativity of the central element $C$ , Proposition 7.1 expresses commutation relations between generators $\{\hat{F}_{ij}\}$ of $\hat{o}_{\infty}$ in terms of relations of generating functions $T^{B}(u,w)$ .

7.2. Inifinite-dimensional Lie algebra $\mathcal{W}^{B}_{1+\infty}$

Let $\mathcal{W}_{1+\infty}^{B}=\mathcal{W}_{1+\infty}\cap\hat{o}_{\infty}$ . As it is observed in [28], every element of $\mathcal{W}_{1+\infty}^{B}$ is a linear span of expressions of the form $\sum_{j\in{\mathbb{Z}}}f(-j)F_{k-j,j}$ with polynomial coefficients $f(-j)$ . Since

\displaystyle i(T^{(k)}_{r})=(-1)^{r}\sum_{j\in{\mathbb{Z}}}(r+j-1)_{k}E_{-j,-r-j}=(-1)^{r}\sum_{j\in{\mathbb{Z}}}(-j-1)_{k}E_{j+r,-j},

this implies that $i(T^{(k)}_{r})\in\mathcal{W}_{1+\infty}$ , and $T^{(k)}_{r}-i(T^{(k)}_{r})\in\mathcal{W}_{1+\infty}^{B}$ . Note also

T^{(k)}(u)-i(T^{(k)}(u))=\sum_{r}\left(T^{(k)}_{r}-i(T^{(k)}_{r})\right)u^{k-r}=\sum_{r}(r+j-1)_{k}F_{r+j,j}u^{k-r}.

On the basis elements of the algebra $\hat{D}$ , the defined above anti-involution has the form

i(t^{k}D^{l})=(-D)^{l}(-t)^{k}=(-t)^{k}(-D-k)^{l}.

Following [2, 19], define

T^{B(k)}(u)=\partial^{k}_{x}T^{B}(x,y)|_{x=u,y=-u}.

In the expansion $T^{B(k)}(u)=\sum_{r\in{\mathbb{Z}}}T^{B(k)}_{r}u^{r-k}$ the coefficients have the form
$T^{B(k)}_{r}=\sum_{j\in{\mathbb{Z}}}(r+j)_{k}F_{j+r,j}$ , and using that

(a+1)_{k}=(a)_{k}+k(a)_{k-1},\quad(a-1)_{k}=(a)_{k}-k(a)_{k-1},

we get relation between $T^{B(k)}(u)$ and $T^{(k)}(u)$ :

	$\displaystyle T^{B(k)}(u)$	$\displaystyle=T^{(k)}(u)-i(T^{(k)}(u))+k(T^{(k-1)}(u)-i(T^{(k-1)}(u))),\quad$
	$\displaystyle T^{(k)}(u)$	$\displaystyle-i(T^{(k)}(u))=T^{B(k)}(u)-kT^{B(k-1)}(u).\quad$

In particular, these relations show that $T^{B(k)}_{r}\in\mathcal{W}_{1+\infty}^{B}$ .

Example 7.1.

For $k=0$ one gets $T^{B(0)}(u)=u\,T^{(0)}(u)-uT^{(0)}(-u),\,$ and from Proposition 7.1,

[T^{B(0)}(u),T^{B(0)}(v)]=4uv\left(\partial_{v}\delta(u,v)-\partial_{v}\delta(-u,v)\right).

8. Presentation of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ through neutral fermions

Similarly to Section 3, we will use the action of a Clifford algebra to construct the action of generators of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ on symmetric functions.

8.1. Clifford algebra of neutral fermions

Definition 8.1.

The Clifford algebra of neutral fermions is generated by $\{\varphi_{i}\}_{i\in{\mathbb{Z}}}$ with relations

\displaystyle\varphi_{m}\varphi_{n}+\varphi_{n}\varphi_{m}=2(-1)^{m}\delta_{m+n,0}\quad\text{for}\quad m,n\in{\mathbb{Z}}.

(8.1)

Collecting generators as coefficients of a formal distribution $\varphi(u)=\sum_{j\in{\mathbb{Z}}}\varphi_{j}u^{-j}$ , relations (8.1) can be written as

\varphi(u)\varphi(v)+\varphi(v)\varphi(u)=2v\delta(v,-u).

With the same Definition 3.2 of normal ordered product, by direct calculation we get the following relations.

Lemma 8.1.

\displaystyle:\varphi(u)\varphi(v):

\displaystyle=\varphi(u)\varphi(v)-2i_{v/u}\left(\frac{u}{u+v}\right)=-\varphi(v)\varphi(u)+2i_{u/v}\left(\frac{v}{v+u}\right).

8.2. Presentation through neutral fermions

Proposition 8.1.

The algebra $\hat{o}_{\infty}$ can be realized through neutral fermions by the identification

\displaystyle T^{B}(u,v)\to:\varphi(u)\varphi(v):\quad\text{ and $C\to Id$.}

Accordingly,

\displaystyle T^{B(k)}(u,v)\to:\partial^{k}_{u}\varphi(u)\varphi(v):.

Proof.

Using Lemma 8.1, we express commutators of normal order products through commutators of regular products, rearrange the order of terms.

	$\displaystyle[:{\varphi}(u){\varphi}(w):,:{\varphi}(v){\varphi}(z):]=[{\varphi}(u){\varphi}(w),{\varphi}(v){\varphi}(z)]$
	$\displaystyle={\varphi}(u){\varphi}(w){\varphi}(v){\varphi}(z)-{\varphi}(v){\varphi}(z){\varphi}(u){\varphi}(w)$
	$\displaystyle=2v\delta(v,-w){\varphi}(u){\varphi}(z)-2v\delta(v,-u){\varphi}(w){\varphi}(z)+2z\delta(z,-w){\varphi}(v){\varphi}(u)-2z\delta(z,-u){\varphi}(v){\varphi}(w).$

Again with the help of Lemma 8.1, we express regular products through normal order products to conclude that $:\varphi(u)\varphi(v):$ satisfies exactly the same commutation relation as $T^{B}(u,v)$ in Proposition 7.1. ∎

Proposition 8.2.

	$\displaystyle[T^{B}(u,v),\varphi(z)]$	$\displaystyle=2z\delta(z,-u)\varphi(v)-2z\delta(z,-v)\varphi(u),$
	$\displaystyle[T^{B(k)}(u),\varphi(z)]$	$\displaystyle=-2z\partial_{u}^{k}\delta(u,-z)\varphi(-u)-2z\delta(z,u)\partial_{u}^{k}\varphi(u).$

Proof.

By Proposition 8.1 and Lemma 8.1,

\displaystyle[T^{B}(u,v),\varphi(z)]

\displaystyle=[\varphi(u)\varphi(v),\varphi(z)]=2z\delta(z,-u)\varphi(v)-2z\delta(z,-v)\varphi(u).

Using (2.2) and differentiating both sides of the line above, we get the second statement. ∎

9. Review of Properties of Schur $Q$ -functions

The analogue of Theorem 6.1 that would describe the action of generators of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ on particular families of symmetric functions can be obtained by direct calculations. However, below, whenever it is possible, we aim to reduce our argument to the work already done for type $A$ , instead of repeating it in the complete analogy to type $A$ . With this goal in mind, in this section we review how Schur $Q$ -functions are connected to Schur symmetric functions.

9.1. Symmetric Schur $Q$ -functions

Let $q_{k}=q_{k}(x_{1},x_{2},\dots)$ be symmetric functions in variables $x_{1},x_{2},\dots$ , defined as coefficients of the expansion of formal series $Q(u)\in\Lambda[[u]]$ , where

\displaystyle Q(u)=\sum_{k\in{\mathbb{Z}}}q_{k}u^{k}=E(u)H(u)=\prod_{i\in{\mathbb{N}}}\frac{1+x_{i}u}{1-x_{i}u}.

(9.1)

Then $q_{k}=\sum_{i=0}^{k}e_{i}h_{k-i}$ for $k>0$ , $q_{0}=1$ , and $q_{k}=0$ for $k<0$ .

Proposition 9.1.

[20, III.8] We have in $\mathcal{D}[[u]]$ (respectively, in $\mathcal{D}[[u^{-1}]]$ )

\displaystyle Q(u)=S_{odd}(u)^{2},\quad\text{where}\quad S_{odd}(u)=exp\left(\sum_{n\in{\mathbb{N}}_{odd}}\frac{p_{n}}{n}{u^{n}}\right),

\displaystyle S_{odd}^{\perp}(u)=exp\left(\sum_{n\in{\mathbb{N}}_{odd}}\frac{\partial}{\partial p_{n}}\frac{1}{u^{n}}\right),

where ${\mathbb{N}}_{odd}=\{1,3,5,\dots\}$ .

Consider formal expansion of the rational function

\displaystyle i_{v/u}\left(\frac{u-v}{u+v}\right)=\left(1-\frac{v}{u}\right)\sum_{k\in{\mathbb{Z}}_{+}}(-1)^{k}\frac{v^{k}}{u^{k}}=1+2\sum_{k\in{\mathbb{N}}}(-1)^{k}\frac{v^{k}}{u^{k}}\in{\mathbb{C}}[[v/u]].

Note that

\displaystyle i_{v/u}\left(\frac{u-v}{u+v}\right)+i_{u/v}\left(\frac{v-u}{v+u}\right)=(u-v)\delta(u,-v)=2\sum_{k\in{\mathbb{Z}}}(-1)^{k}\frac{u^{k}}{v^{k}}.

(9.2)

Then Schur Q-function $Q_{\lambda}=Q_{\lambda}(x_{1},x_{2},\dots)$ can be defined as coefficients of ${u^{\lambda_{1}}\dots u^{\lambda_{l}}}$ of the formal distribution

\displaystyle Q(u_{l},\dots,u_{1})=\sum_{\lambda_{1},\dots,\lambda_{l}\in{\mathbb{Z}}}{Q_{\lambda}}{u_{1}^{\lambda_{1}}\dots u_{l}^{\lambda_{l}}}=\prod_{1\leq i<j\leq l}i_{u_{j}/u_{i}}\left(\frac{u_{i}-u_{j}}{u_{i}+u_{j}}\right)\prod_{i=1}^{l}Q(u_{i}),

(9.3)

see [20, III, (8.8)]. The values of these coefficients as symmetric polynomials are given by the formula

\displaystyle Q_{\lambda}(x_{1},\dots,x_{N})=\frac{2^{l}}{(N-l)!}\sum_{\sigma\in S_{N}}\prod_{i=1}^{l}x_{\sigma(i)}^{\lambda_{i}}\prod_{i<j}\frac{x_{\sigma(i)}+x_{\sigma(j)}}{x_{\sigma(i)}-x_{\sigma(j)}},

(9.4)

for strict partitions $\lambda$ , and are zero otherwise [20, III, (8.7)]. Schur $Q$ -polynomials have a stabilization property [20, III, (2.5)], hence one can omit the number $N$ of variables $x_{i}$ ’s, as long as it is not less than the length of the partition $\lambda$ , considering $Q_{\lambda}$ as a function of infinitely many variables $(x_{1},x_{2}\dots)$ .

9.2. Action of neutral fermions on the bosonic space $\mathcal{B}_{odd}$

Let $\mathcal{B}_{odd}$ be the subspace of the ring of symmetric functions generated by odd power sums:

\mathcal{B}_{odd}={\mathbb{C}}[p_{1},p_{3},p_{5},\dots]\subset\Lambda={\mathbb{C}}[p_{1},p_{2},p_{3},p_{4}\dots].

Then $q_{k}\in\mathcal{B}_{odd}$ , and that $\mathcal{B}_{odd}={\mathbb{C}}[q_{1},q_{3},\dots]$ , [20, III.8 (8.3)].

It is not difficult to prove (see e.g. [25]) that the space $\mathcal{B}_{odd}$ is invariant with respect to action of $e^{\perp}_{k}$ , and $h^{\perp}_{k}$ , and that the restrictions to ${\mathcal{B}_{odd}}$ of action of coefficients of the following formal distributions coincide:

E^{\perp}(u)|_{\mathcal{B}_{odd}}=H^{\perp}(u)|_{\mathcal{B}_{odd}}=S_{odd}^{\perp}(u).

Denote as $E_{odd}^{\perp}(u)=E^{\perp}(u)|_{\mathcal{B}_{odd}}$ and $H_{odd}^{\perp}(u)=H^{\perp}(u)|_{\mathcal{B}_{odd}}$ .

Define a formal distribution $\varphi(u)$ of operators acting on $\mathcal{B}_{odd}$ :

\displaystyle\varphi(u)=E(u)H(u)E^{\perp}_{odd}(-u)=Q(u)S_{odd}^{\perp}(-u).

(9.5)

Let $\{\varphi_{i}\}_{i\in{\mathbb{Z}}}$ be coefficients of the expansion $\varphi(u)=\sum_{j\in{\mathbb{Z}}}\varphi_{-j}u^{j}$ . One can prove (see e.g. [25]) that $\varphi(u)$ is a quantum field that acts exactly as the Clifford algebra of neutral fermions on the space $\mathcal{B}_{odd}$ , satisfying relations

\varphi(u)\varphi(v)+\varphi(v)\varphi(u)=2v\delta(v,-u),

or, in terms of coefficients,

\displaystyle\varphi_{m}\varphi_{n}+\varphi_{n}\varphi_{m}=2(-1)^{m}\delta_{m+n,0}\quad\text{for}\quad m,n\in{\mathbb{Z}}.

(9.6)

This quantum field can be also expressed in the bosonic form

\displaystyle\varphi(u)=Q(u)S_{odd}(-u)^{\perp}=exp\left(\sum_{n\in{\mathbb{N}}_{odd}}\frac{2p_{n}}{n}{u^{n}}\right)exp\left(-\sum_{n\in{\mathbb{N}}_{odd}}\frac{\partial}{\partial p_{n}}\frac{1}{u^{n}}\right).

Note that (9.5) implies that the action of neutral fermions on $\mathcal{B}_{odd}$ can be expressed through the restritions of the action of charged free fermions on $\mathcal{B}_{odd}$ :

\displaystyle\varphi(u)=H(u)\psi^{+}(u)|_{\mathcal{B}_{odd}}=E(u)\psi^{-}(-u)|_{\mathcal{B}_{odd}}.

By (9.5) and Lemma 4.2 we get variations of vertex operator presentations of Schur $Q$ -functions, first considered in [8]:

	$\displaystyle\varphi(u_{l})\dots\varphi(u_{1})(1)$	$\displaystyle=Q(u_{l},\dots,u_{1}),$		(9.7)
	$\displaystyle\varphi_{-\lambda_{l}}\dots\varphi_{-\lambda_{1}}(1)$	$\displaystyle=Q_{\lambda}.$

9.3. Bosonic normal ordered products for neutral fermions

Similarly to Section 5.3, define the bosonic normal ordered product of neutral fermions as

\displaystyle\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\varphi(u)\varphi(v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=Q(u)Q(v)E_{odd}^{\perp}(-u)E_{odd}^{\perp}(-v).

(9.8)

Relations between different products of neutral free fermions are collected in the following lemma that follows from the definition of $\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\varphi(u)\varphi(v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}$ and Lemma 4.2.

Lemma 9.1.

\displaystyle:\varphi(u)\varphi(v):

\displaystyle=\varphi(u)\varphi(v)-2i_{v/u}\left(\frac{u}{u+v}\right)=i_{v/u}\left(\frac{u-v}{u+v}\right)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\varphi(u)\varphi(v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}-2i_{v/u}\left(\frac{u}{u+v}\right).

(9.9)

We introduce notation

\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\varphi(u)\varphi(v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}.

Note that this formal distribution can be expressed through restriction of formal distribution $\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}$ to $\mathcal{B}_{odd}$ :

\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=E(u)H(v)\,\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T(u,-v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}|_{\mathcal{B}_{odd}},

and that $\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}T^{B}(u,-u)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}=1.$

Define a differential operator

D^{(k)}_{u}=\partial_{u}^{k}+\frac{2u}{k+1}\partial_{u}^{k+1}.

Lemma 9.2.

Operator $D^{(k)}_{u}$ satisfies the analogue of Leibniz rule:

D^{(k)}_{u}(fg)=\sum_{r=0}^{k}{k\choose r}\partial_{u}^{r}(f)\,\,D^{(k-r)}_{u}(g)+\frac{2u}{k+1}\partial^{k+1}_{u}(f)\,g.

Proposition 9.2.

We can express the generating functions of $\hat{o}_{\infty}$ and of $\mathcal{W}^{B}_{1+\infty}$ as

T^{B}(u,v)=i_{v/u}\left(\frac{1}{u+v}\right)\left((u-v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}-2u\right),

and

\displaystyle T^{B\,(k)}(u)

\displaystyle=D^{(k)}_{u}\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,T^{B}(u,v))\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}|_{v=-u}-2\delta_{k,0}.

(9.10)

Proof.

The first statement immediately follows from Proposition 8.1 and Lemma 9.1. For the second statement, observe that

(u+v)T^{B}(u,v)=(u-v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}-2u.

Taking the derivative $\partial_{u}^{k+1}$ of both sides of this equality and evaluating $v=-u$ gives

\displaystyle(k+1)T^{B(k)}(u)=(k+1)\partial_{u}^{k}\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,|_{v=-u}+2u\partial_{u}^{k+1}\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}\,T^{B}(u,v)\raisebox{-0.21529pt}{*}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{*}}|_{v=-u}\,-2\delta_{k,0},

which proves (9.10). ∎

10. Action of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ on symmetric functions

10.1. The second main result

Introduce notation

\displaystyle\mathcal{Q}_{l}(-u^{-1})=\prod_{s=1}^{l}i_{u_{i}/u}\left(\frac{1+u_{i}/u}{1-u_{i}/u}\right)=\mathcal{E}_{l}(-u^{-1})\mathcal{H}_{l}(-u^{-1}).

Theorem 10.1.

(1)

The action of $\hat{o}_{\infty}$ on any $f\in\mathcal{B}_{odd}$ is described by

\displaystyle T^{B}(u,v)(f)=\left(i_{v/u}\left(\frac{u-v}{u+v}\right)Q(u)Q(v)E^{\perp}_{odd}(-u)E^{\perp}_{odd}(-v)-i_{v/u}\left(\frac{2u}{u+v}\right)\right)(f).

(10.1)

(2)

The action of $\hat{o}_{\infty}$ on the basis of Schur symmetric $Q$ -functions in terms of generating functions is given by

\displaystyle T^{B}(u,v)

\displaystyle\left(Q(u_{l},..,u_{1})\right)=\quad Q(u,v,u_{l},\dots,u_{1})-i_{v/u}\left(\frac{2u}{u+v}\right)Q(u_{l},\dots,u_{1}).

(10.2)

(3)

The action of $\mathcal{W}^{B}_{1+\infty}$ on any $f\in\mathcal{B}_{odd}$ is described by

	$\displaystyle T^{B\,(k)}(u)\,(f)=$	$\displaystyle\left(\sum_{r=0}^{k}{k\choose r}\partial_{u}^{r}(Q(u))Q(-u)\,\,E_{odd}^{\perp}(u)D^{(k-r)}_{u}(E_{odd}^{\perp}(-u))\right)(f)$		(10.3)
		$\displaystyle+\left(\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)-2\delta_{k,0}\right)(f).$

(4)

The action of $\mathcal{W}^{B}_{1+\infty}$ on the basis of Schur symmetric $Q$ -functions is defined by

\displaystyle T^{B\,(k)}(u)\,\left(Q(u_{l},..,u_{1})\right)=D^{(k)}_{u}(Q(u)\mathcal{Q}_{l}(u))\,Q(-u,u_{l},..,u_{1})-2\delta_{k,0}Q(u_{l},..,u_{1}).

(10.4)

Proof.

(1)

Follows immediately from (9.8) and Proposition 9.2.

(2)

By Lemma 9.1 and (9.7),

	$\displaystyle T^{B}(u,v)$	$\displaystyle\left(Q(u_{l},..,u_{1})\right)=\left(\varphi(u)\varphi(v)-i_{v/u}\left(\frac{2u}{u+v}\right)\right)\left(Q(u_{l},..,u_{1})\right)$
	$\displaystyle=$	$\displaystyle Q(u,v,u_{l},..,u_{1})-i_{v/u}\left(\frac{2u}{u+v}\right)Q(u_{l},..,u_{1}).$

(3)

Using Lemma 9.2, and that $E_{odd}^{\perp}(-u)E_{odd}^{\perp}(u)=1$ , for $k>0$ we get

	$\displaystyle T^{B\,(k)}(u)=D^{(k)}_{u}\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\,T^{B}(u,v))\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\|_{v=-u}-2\delta_{k,0}$
	$\displaystyle=D^{(k)}_{u}\left(Q(u)Q(v)E_{odd}^{\perp}(-v)E_{odd}^{\perp}(-u)\right)\raisebox{-0.21529pt}{}\hbox to0.0pt{\hss\raisebox{-5.16663pt}{}}\|_{v=-u}-2\delta_{k,0}$
	$\displaystyle=\sum_{r=0}^{k}{k\choose r}\partial_{u}^{r}(Q(u))Q(-u)\,\,E_{odd}^{\perp}(u)D^{(k-r)}_{u}(E_{odd}^{\perp}(-u))+\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)-2\delta_{k,0}.$

(4)

Note that

E^{\perp}_{odd}(-u)Q(u_{l},\dots,u_{1})=\mathcal{Q}_{l}(-u^{-1})Q(u_{l},\dots,u_{1}),

and therefore

D^{(k-r)}_{u}(E^{\perp}_{odd}(-u)Q(u_{l},\dots,u_{1}))=D^{(k-r)}_{u}(\mathcal{Q}_{l}(-u^{-1}))Q(u_{l},\dots,u_{1}).

From (10.3), (9.5), and Lemma 9.2, for $k>0$ ,

	$\displaystyle T^{B\,(k)}(u)Q(u_{l},\dots,u_{1})$
	$\displaystyle=\sum_{r=0}^{k}{k\choose r}\partial_{u}^{r}(Q(u))Q(-u)\,\,E_{odd}^{\perp}(u)D^{(k-r)}_{u}(\mathcal{Q}_{l}(-u^{-1}))Q(u_{l},\dots,u_{1})$
	$\displaystyle+\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)Q(u_{l},\dots,u_{1}).$
	$\displaystyle=\sum_{r=0}^{k}{k\choose r}\partial_{u}^{r}(Q(u))D^{(k-r)}_{u}(\mathcal{Q}_{l}(u))\,Q(-u,u_{l},\dots,u_{1})$
	$\displaystyle+\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)Q(u_{l},\dots,u_{1}).$
	$\displaystyle=D^{(k)}_{u}(Q(u)\mathcal{Q}_{l}(u))Q(-u,u_{l},\dots,u_{1})-\frac{2u}{k+1}\partial^{k+1}_{u}(Q(u))\,\mathcal{Q}_{l}(u)Q(-u,u_{l},\dots,u_{1})$
	$\displaystyle+\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)Q(u_{l},\dots,u_{1}).$

Observe that, by (9.3), $Q(-u,u_{l},\dots,u_{1})=Q(-u)\mathcal{Q}_{l}(-u)Q(u_{l},\dots,u_{1}),$ and that $\mathcal{Q}_{l}(u)\mathcal{Q}_{l}(-u)=1$ . This implies that

\displaystyle-\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)\mathcal{Q}_{l}(u)Q(-u,u_{l},\dots,u_{1})+\frac{2u}{k+1}\partial^{k+1}_{u}Q(u)Q(-u)Q(u_{l},\dots,u_{1})=0.

For $k=0$ one more term of the form $2\delta_{k,0}Q(u_{l},..,u_{1})$ is subtracted, which completes the proof of proposition.

∎

Remark 10.1.

In analogy to Remark 6.1, we observe that the action (10.3) of $\hat{o}_{\infty}$ and the action (10.4) of $\mathcal{W}^{B}_{1+\infty}$ are again described by multiplication operators, with a duality property naturally connected to the Cauchy identity for Schur $Q$ -functions

\displaystyle Q(u_{l},\dots,u_{1})\prod_{1\leq i<j\leq l}i_{u_{j}/u_{i}}\left(\frac{u_{i}+u_{j}}{u_{i}-u_{j}}\right)=\prod_{i=1}^{l}Q(u_{i})=\sum_{\lambda\in{\mathbb{Z}}}{Q_{\lambda}(x_{1},x_{2},\dots)}Q_{\lambda}({u_{1},\dots,u_{l}}),

for $|u_{1}|<|u_{2}|<\dots<|u_{l}|$ , [20, III.8 (8.13)].

Acknowledgments

The research of the second author is supported by the Simons Foundation Travel Support for Mathematicians grant MP-TSM-00002544.

Statements and Declarations

There is no data associated with this paper. The authors have no competing interests to declare that are relevant to the content of this article.

References

[1] Alexandrov, A.: Intersection numbers on $M_{g,n}$ and BKP hierarchy. J. High Energy Phys. 2021, 13 (2021). https://doi.org/10.1007/jhep09(2021)013
[2] Alexandrov, A.: Generalized Brézin–Gross–Witten tau function as a hypergeometric solution of the BKP hierarchy. Adv. Math. 412, 36 (2023). https://doi.org/10.1016/j.aim.2022.108809
[3] Awata, H., Fukuma, M., Matsuo, Y., Odake, S.: Representation theory of the $W_{1+\infty}$ algebra. Prog. Theor. Phys. Suppl. 118, pp. 343–373 (1995). https://doi.org/10.1143/PTPS.118.343
[4] Awata, H., Fukuma, M., Matsuo, Y., Odake, S.: Representation theory of $W_{1+\infty}$ algebra. In: Group Theoretical Methods in Physics (Toyonaka, 1994), pp. 335–338. World Scientific, River Edge, NJ (1995). https://doi.org/10.1143/PTPS.118.343
[5] Baker, T.H.: Symmetric function products and plethysms and the boson–fermion correspondence. J. Phys. A 28, 589-606 (1995). https://doi.org/10.1088/0305-4470/28/3/014
[6] Bakalov, B., Kac, V.G., Voronov, A.A.: Cohomology of conformal algebras. Commun. Math. Phys. 200, 561–598 (1999). https://doi.org/10.1007/s002200050541
[7] Frenkel, E., Kac, V., Radul, A., Wang, W.: $W_{1+\infty}$ and $W(\mathfrak{gl}_{N})$ with central charge $N$ . Comm. Math. Phys. 170, 337–357 (1995). https://doi.org/10.1142/S0217751X92003781
[8] Jing, N.: Vertex operators, symmetric functions, and the spin group $\Gamma_{n}$ . J. Algebra 138, 340–398 (1991). https://doi.org/10.1016/0021-8693(91)90177-A
[9] Jing, N.: Vertex operators and Hall–Littlewood symmetric functions. Adv. Math. 87, 226–248 (1991). https://doi.org/10.1016/0001-8708(91)90072-F
[10] Jing, N., Rozhkovskaya, N.: Generating functions for symmetric and shifted symmetric functions. J. Comb. 10, 111–127 (2019). https://doi.org/10.4310/joc.2019.v10.n1.a5
[11] Kac, V.G., Peterson, D.H.: Spin and wedge representations of infinite-dimensional Lie algebras and groups. Proc. Natl. Acad. Sci. U.S.A. 78, 3308–3312 (1981). https://doi.org/10.1073/pnas.78.6.3308
[12] Kac, V.G., Radul, A.O.: Quasifinite highest weight modules over the Lie algebra of differential operators on the circle. Comm. Math. Phys. 157, 429–457 (1993). https://doi.org/10.1007/BF02096878
[13] Kac, V.G., Radul, A.O.: Representation theory of the vertex algebra $W_{1+\infty}$ . Transform. Groups 1, 41–70 (1996). https://doi.org/10.1007/BF02587735
[14] Kac, V.G.: Vertex Algebras for Beginners, 2nd edn. University Lecture Series, vol. 10. Amer. Math. Soc., Providence, RI (1998)
[15] Kac, V.G., Raina, A.K.: Bombay Lectures on Highest Weight Representations of Infinite-Dimensional Lie Algebras, 1st edn. Adv. Ser. Math. Phys., vol. 2. World Scientific, NJ (1987)
Kac, V.G., Raina, A.K., Rozhkovskaya, N.: Bombay Lectures on Highest Weight Representations of Infinite-Dimensional Lie Algebras, 2nd edn. World Scientific, NJ (2013)
[16] Kac, V.G., Van de Leur, J.W., Rozhkovskaya, N.: Polynomial tau-functions of the KP, BKP, and the s-component KP hierarchies. J. Math. Phys. 62, 021702 (2021). https://doi.org/10.1063/5.0013017
[17] Liu, X., Yang, C.: $Q$ -polynomial expansion for Brézin–Gross–Witten tau function. Adv. Math. 404, 22 (2022). https://doi.org/10.1016/j.aim.2022.108456
[18] Liu, X., Yang, C.: Schur $Q$ -polynomials and Kontsevich–Witten tau function. Peking Math. J. 7 713-758 (2024). https://doi.org/10.1007/s42543-023-00064-6
[19] Liu, X., Yang, C.: Action of $W$ -type operators on Schur functions and Schur $Q$ -functions. J. Lond. Math. Soc. (2) 111 29 (2025). https://doi.org/10.1112/jlms.70080
[20] Macdonald, I.G.: Symmetric Functions and Hall Polynomials, 2nd edn. Oxford Univ. Press, New York (1995)
[21] Mironov, A., Morozov, A.: Superintegrability of Kontsevich matrix model. Eur. Phys. J. C 81, 270 (2021). https://doi.org/10.1140/epjc/s10052-021-09030-x
[22] Miwa, T., Jimbo, M., Date, E.: Solitons: Differential Equations, Symmetries and Infinite Dimensional Algebras. Cambridge Univ. Press, Cambridge (2000)
[23] Necoechea, G., Rozhkovskaya, N.: Generalized vertex operators of Hall–Littlewood polynomials as twists of charged free fermions. J. Math. Sci. (N.Y.) 247, 926–938 (2020). https://doi.org/10.1007/s10958-020-04847-5
[24] Pope, C.N., Shen, X., Romans, L.J.: $\mathcal{W}_{\infty}$ ) and the Racah–Wigner algebra. Nuclear Phys. B 339, 191–122 (1990). https://doi.org/10.1016/0550-3213(90)90539-P
[25] Rozhkovskaya, N.: Multiparameter Schur $Q$ -functions are solutions of the BKP hierarchy. SIGMA Symmetry Integrability Geom. Methods Appl. 15, 14 (2019). https://doi.org/10.3842/SIGMA.2019.065
[26] M. Sato, Soliton equations as dynamical systems on infinite-dimensional Grassmann manifold, RIMS Kokyuroku, 439, 30–46 (1981)
[27] Stanley, R.P.: Enumerative Combinatorics, Vol. 2. Cambridge Univ. Press, Cambridge (1999)
[28] Van de Leur, J.: The Adler–Shiota–van Moerbeke formula for the BKP hierarchy. J. Math. Phys. 36, 4940–4951 (1995). https://doi.org/10.1063/1.531352
[29] You, Y.: Polynomial solutions of the BKP hierarchy and projective representations of symmetric groups. In: Infinite-Dimensional Lie Algebras and Groups (Luminy–Marseille, 1988). Adv. Ser. Math. Phys., vol. 7, 449–464. World Scientific, Singapore (1989)

Generating functions of 𝒲1+∞\mathcal{W}_{1+\infty} action on symmetric functions

Abstract.

2010 Mathematics Subject Classification:

1. Introduction

2. Lie subalgebra 𝒲1+∞\mathcal{W}_{1+\infty} of a^∞\hat{a}_{\infty}

2.1. Lie algebra a^∞\hat{a}_{\infty}

2.2. Formal distributions

Remark 2.1.

Lemma 2.1.

Proof.

2.3. Infinite-dimensional Lie algebra 𝒲1+∞\mathcal{W}_{1+\infty}

Proposition 2.1.

Proof.

2.4. 𝒲1+∞\mathcal{W}_{1+\infty} as the central extension of the algebra of differential operators

2.5. Definition of conformal Lie algebra

Definition 2.1.

2.6. Conformal Lie algebra structure of 𝒲1+∞\mathcal{W}_{1+\infty}

Proposition 2.2.

Proof.

Lemma 2.2.

Proof.

2.7. Heisenberg and Virasoro algebras

Example 2.1.

Remark 2.2.

Example 2.2.

3. Presentation of a^∞\hat{a}_{\infty} and 𝒲1+∞\mathcal{W}_{1+\infty} through fermions

3.1. Clifford algebra of charged free fermions

Definition 3.1.

Definition 3.2.

Lemma 3.1.

3.2. Presentation through charged free fermions

Proposition 3.1.

Proof.

Remark 3.1.

Proposition 3.2.

Proof.

Remark 3.2.

4. Review of properties of symmetric functions

4.1. On the boson-ferimon correspondance

4.2. Symmetric functions

Lemma 4.1.

Proposition 4.1.

Lemma 4.2.

Proof.

Corollary 4.1.

5. Action of charged free fermions on the space of symmetric functions

5.1. Action of charged free fermions on symmetric functions

Proposition 5.1.

Remark 5.1.

5.2. Generating function for Schur symmetric functions

Remark 5.2.

5.3. Bosonic normal ordered products for charged free fermions

Proposition 5.2.

Proof.

6. Action of a^∞\hat{a}_{\infty} and 𝒲1+∞\mathcal{W}_{1+\infty} on symmetric functions

6.1. The first main result

Theorem 6.1.

Proof.

Remark 6.1.

6.2. Actions of Heisenberg and Virasoro algebra

Example 6.1.

Example 6.2.

Example 6.3.

Proposition 6.1.

Remark 6.2.

Proof.

6.3. Some connections to existing results

7. Lie algebra 𝒲1+∞B\mathcal{W}^{B}_{1+\infty} as a subalgebra of a^∞\hat{a}_{\infty}

7.1. Lie algebra o^∞\hat{o}_{\infty}

Proposition 7.1.

Proof.

Remark 7.1.

7.2. Inifinite-dimensional Lie algebra 𝒲1+∞B\mathcal{W}^{B}_{1+\infty}

Example 7.1.

8. Presentation of o^∞\hat{o}_{\infty} and 𝒲1+∞B\mathcal{W}^{B}_{1+\infty} through neutral fermions

8.1. Clifford algebra of neutral fermions

Definition 8.1.

Lemma 8.1.

8.2. Presentation through neutral fermions

Proposition 8.1.

Generating functions of $\mathcal{W}_{1+\infty}$ action on symmetric functions

2. Lie subalgebra $\mathcal{W}_{1+\infty}$ of $\hat{a}_{\infty}$

2.1. Lie algebra $\hat{a}_{\infty}$

2.3. Infinite-dimensional Lie algebra $\mathcal{W}_{1+\infty}$

2.4. $\mathcal{W}_{1+\infty}$ as the central extension of the algebra of differential operators

2.6. Conformal Lie algebra structure of $\mathcal{W}_{1+\infty}$

3. Presentation of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ through fermions

6. Action of $\hat{a}_{\infty}$ and $\mathcal{W}_{1+\infty}$ on symmetric functions

7. Lie algebra $\mathcal{W}^{B}_{1+\infty}$ as a subalgebra of $\hat{a}_{\infty}$

7.1. Lie algebra $\hat{o}_{\infty}$

7.2. Inifinite-dimensional Lie algebra $\mathcal{W}^{B}_{1+\infty}$

8. Presentation of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ through neutral fermions

9. Review of Properties of Schur $Q$ -functions

9.1. Symmetric Schur $Q$ -functions

9.2. Action of neutral fermions on the bosonic space $\mathcal{B}_{odd}$

10. Action of $\hat{o}_{\infty}$ and $\mathcal{W}^{B}_{1+\infty}$ on symmetric functions