Equivariant cohomology of infinite-dimensional Grassmannian and shifted Schur functions (Q1951834)

Denote by \(\mathcal{H}:=L^2(S^1)\) the Hilbert space and let \(\mathcal{H}^+\) and \(\mathcal{H}^-\) be closed subspaces of \(\mathcal{H}\) spanned by \(\{z^j:j\geq0\}\) and \(\{z^j:j<0\}\). The Segel-Wilson version Sato Grassmannian \(\mathrm{Gr}(\mathcal{H})\) is defined to be a space of all closed linear subspaces \(W\subset \mathcal{H}\) such that the projection \(\pi_{\_}: W\to \mathcal{H}_{\_}\) is a Fredholm operator and the projection \(\pi_+: W\to \mathcal{H}_{+}\) is a compact operator. The group \(S^1=\{\alpha\in \mathbb{C}: |\alpha|=1\}\) acts on \(\mathcal{H}\) by assigning \(f\) to \(\tilde{f}\), where \(\tilde{f}(z)=f(\alpha z)\). The action generates an action of \(S^1\) on Grassmannian. If we represent a function on \(S^1\) as a Fourier series \(f(z)=\sum a_n z^n\), then this action sends \(a_k\to \alpha^k a_k\). More general action of \(S^1\) on \(\mathcal{H}\) is given by sending \(a_k\to \alpha^{n_k}a_k\) where \(n_k\) is an arbitrary doubly infinite sequence of integers. Similarly, this action also generates an action of \(S^1\) on Grassmannian \(\mathrm{Gr}(\mathcal{H})\), which is a disjoint union of connected components \(\mathrm{Gr}_d(\mathcal{H})\) labeled by the index of the projection \(\pi_{\_}: W\to \mathcal{H}_{\_}\). All components are homeomorphic. Moreover, every component is homotopy equivalent to a subspace having a cell decomposition \(K=\cup \sigma_{\lambda}\) consisting of even-dimensional cells. The cells are labeled by partitions. This decomposition is \(S^1\)-invariant and so the equivariant cohomology \(H_{S^1}(\mathrm{Gr}_d(\mathcal{H}))\) has a free system of generators \(\Omega^T_{\lambda}:=[\overline{\Sigma}_{\lambda}]\) as a module over \(H_{S^1}(pt)\). These generators is interpreted as cohomology classes dual to Schubert cycles \(\overline{\Sigma}_{\lambda}\) having finite codimension. In this paper, the main result is the calculation on the multiplication table in the basis \(\Omega^T_{\lambda}\). The explicit formula is too complicated to state here.