Equivariant quantum cohomology of cotangent bundle of \(G/P\) (Q904019): Difference between revisions

A subgroup \(P\) of a complex semisimple linear algebraic group \(G\) is said to be parabolic if \({\mathcal P}:=G/P\) is a projective variety. Like for all kind of manifolds, the cotangent bundle \(T^*{\mathcal P}\) carries a canonical symplectic form which, in addition, makes it into a symplectic resolution, i.e. the map \(T^*{\mathcal P}\rightarrow H^0(T^*{\mathcal P},O_{T^*{\mathcal P}})\) is projective and birational. The obvious transitive action of \(G\) on \({\mathcal P}\) induces an action on \(T^*{\mathcal P}\) and is thus meaningful to speak of its equivariant cohomology. Let \(A\) be a maximal subtorus contained in a Borel subgroup \(B\) of \(G\): the paper under review is concerned with the quantum product in the equivariant cohomology of \(T^*{\mathcal P}\), with respect to the action of \(A\times {\mathbb C}\) and to the action of \(G\times {\mathbb C^*}\). Recall that the quantum multiplication is a deformation of the usual cohomology product over a ring of formal power series that encodes enumerative informations regarding parameterized rational curves in the variety of a given fixed homology class. In the framework of the paper under review, it encodes the \(3\)-points, genus \(0\), degree \(\beta\) equivariant Gromov-Witten invariant of \(X\), where \(\beta\in H_2(X,{\mathbb Z})\). To each element \(y\) of the Weil group \(W\), \textit{D. Maulik} and \textit{A. Okounkov} taught us how to associate an element of the stable basis introduced and described in the important preprint [``Quantum groups and quantum cohomology'', Preprint, \url{arXiv:1211.1287}]. The basis element \(\mathrm{stab}_+(y)\) associated to a given \(y\in W\) is constructed via a delicate procedure involving the subgroup of the Weil group \(W\) generated by the simple reflections associated to the roots and the \(A\)-equivariant Euler class of the cotangent space \(T_{\overline{y}}{\mathcal P}\), where \(\overline{y}\) is the coset \(yW_P\), which the author describe carefully in the second section of the article. A first main result of the paper, already stated in the introduction and numbered Theorem 1.1, concerns the quantum equivariant product in \(H_{ A\times{\mathbb C}^*}(T^*{\mathcal P})\). It is a sort of Chevalley rule describing the expansion \(D_\lambda* \mathrm{stab}_+(y)\) of the product of a distinguished divisor class on \(T^*{\mathcal P}\) and the stable basis element associated to \(y\). The divisor class is the first Chern class \(D_\lambda\) of a line bundle over \({\mathcal P}\) associated to a weight \(\lambda\) vanishing at a co-root corresponding to a simple root of \(G\). The second main result, Theorem 1.2, is obtained as a consequence of both Theorem 1.1. and the restriction formula for stable bases detailedly studied by \textit{C. Su}, [``Restriction formula for stable basis of Springer resolution'', Preprint, \url{arXiv:1501.04214}]: it describes the quantum intersection product between \(D_\lambda\) and any \(G\times {\mathbb C^*}\)-equivariant cohomology classes of \(T^*{\mathcal P}\). The author remarks that the formula proven in theorem 1.2. is conjugated, in a suitable sense, to one previously conjectured by A.~Braverman in a private communication. The proof of the two main results is contained respectively in Section 3 and 4, while Section 2 is devoted to summarize the essential preliminaries about quantum cohomology and the construction of the stable basis alluded to above. The essential, but complete, reference list is extremely helpful to the less expert reader wishing to find his own way to follow and appreciate the results and techniques employed in this truly intriguing article.