2
$\begingroup$

I would like to know the definition of Givental $J$-function of cotangent bundle of flag variety. To state my question more precisely, let us briefly recall the definition of the Givental $J$-function of a compact Kahler variety $X$. Let $T_0=1,T_1, \cdots, T_m$ be the basis of the cohomology group $H^*(X,\mathbb{Z})$, and $T_1, \cdots, T_r$ be the basis of the second cohomology group $H^2(X,\mathbb{Z})$. We define the matrix $g_{ij}=\int_{X}T_i\cup T_j$, and its inverse matrix $g^{ij}=(g_{ij})^{-1}$, which provide the dual basis \begin{equation} T^a=\sum_{b=1}^mg^{ab}T_b~, \end{equation} so that $\int_X T^i\cup T_j=\delta^i_j$. We denote by $\overline{\mathcal{M}}_{g,n}(X,\beta)$ the moduli space of stable maps from connected genus $g$ curves with $n$-marked points to $X$ representing the class $\beta\in H_2(X,\mathbb{Z})$. Let $\mathcal{L}_1,\cdots,\mathcal{L}_n$ be the corresponding tautological line bundles over $\overline{\mathcal{M}}_{g,n}(X,\beta)$. For $\gamma_1,\cdots,\gamma_n\in H^*(X,\mathbb{Z})$ and non-negative integers $d_i$, the gravitational correlation function is defined \begin{equation} \left\langle \tau_{d_1} \gamma_1, \cdots, \tau_{d_n} \gamma_n \right\rangle_{g,\beta}=\int_{[\overline{\mathcal{M}}_{g,n}(X,\beta)]^{\textrm{vir}}} \prod_{i=1}^n c_1(\mathcal{L}_i) ^{d_i}\cup\textrm{ev}^*(\gamma_i)~. \end{equation} The $J$-function of $X$ is defined by using the psi class $\psi=c_1\mathcal{L}_1)$ \begin{equation} J[X]=e^{\delta/\hbar}\left(1+\sum_{\beta\in H_2(X,\mathbb{Z})}\sum_{a=1}^m q^\beta \left\langle\frac{T_a}{\hbar-\psi},1 \right\rangle_{0,\beta}~T^a\right)~, \end{equation} where $\delta=\sum_{i=1}^r t_i T_i$ and $q^\beta=e^{\int_\beta \delta}$.

Recently, I have conjectured in this paper the expression of hypergeometric type for $J$-function of cotangent bundle of complete flag variety $\mathrm{Fl}_N=SL(N,\mathbb{C})/B$ by using the supersymmetric partition function on $S^2$: \begin{eqnarray} J[T^*\mathrm{Fl}_N]&=&\sum_{\vec{k}^{(I)}} \prod_{I=1}^{N-1}z_I^{\vert k^{(I)}\vert } \prod_{I=2}^{N-1}\prod_{s\neq t}^{I} \tfrac{(1+\hbar^{-1} H_{st}^{(I)}+\hbar^{-1}m)_{k_s^{(I)}-k_t^{(I)}}}{(\hbar^{-1} H_{st}^{(I)})_{k_s^{(I)}-k_t^{(I)}}} \cr && \prod_{I=1}^{N-2} \prod_{s=1}^{I}\prod_{t=1}^{{I+1}} \tfrac{(\hbar^{-1} H_{s}^{(I)} -\hbar^{-1} H_{t}^{(I+1)}-\hbar^{-1}m)_{k^{(I)}_s - k^{(I+1)}_t}}{(1+\hbar^{-1} H_{s}^{(I)} -\hbar^{-1} H_{t}^{(I+1)})_{k^{(I)}_s - k^{(I+1)}_t}}\cr && \prod_{s=1}^{N-1}\prod_{t=1}^{N} \tfrac{(\hbar^{-1} H_{s}^{(N-1)}-\hbar^{-1} H_{t}^{(N)}-\hbar^{-1}m)_{k^{(N-1)}_s}}{(1+\hbar^{-1} H_{s}^{(N-1)}-\hbar^{-1} H_{t}^{(N)})_{k^{(N-1)}_s}}~, \end{eqnarray} where we identify $H_{s}^{(I)}$ $(s = 1,...,I)$ with Chern roots to the duals of the universal bundles $\mathcal{S}_I$: \begin{equation} 0\subset \mathcal{S}_1 \subset \mathcal{S}_2 \subset \dots \subset \mathcal{S}_{N-1} \subset \mathcal{S}_{N} = {\mathbb{C}}^N \otimes {\cal O}_{\mathrm{Fl}_N}~. \end{equation} I have checked that, after multiplying an appropriate factor, it becomes an eigenfunction of trigonometric Calgero-Moser Hamiltonian \begin{equation} H=\hbar^2\sum_{i}\frac{\partial^2}{\partial t_i^2}-2m(m+\hbar)\sum_{\alpha\in \Delta^+} \frac 1{(e^{\langle t,\alpha\rangle/2}-e^{-\langle t,\alpha\rangle/2})^2}~. \end{equation} This is consistent with theorem 3.2 of Braverman-Maulik-Okounkov.

Question: The formula of $J[T^*\mathrm{Fl}_N]$ certainly contains the parameter $m$ that does not appear in the definition of $J$-function of a compact Kahler variety. It appears to me that one has to introduce the parameter $m$ since the cotangent bundle of the complete flag variety is non-compact. I would like to know the definition of $J[T^*\mathrm{Fl}_N]$ and how it incorporates the parameter $m$.

$\endgroup$

1 Answer 1

2
$\begingroup$

The parameter $m$ appears if you study everything equivariantly with respect to the ${\mathbb C}^*$-action which acts in the standard way on the fibers of the cotangent bundle.

$\endgroup$

Your Answer

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

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