On the cohomology of stable map spaces (Q1417456)

In the algebro-geometric approach to Gromov-Witten theory, the study of \(n\)-pointed stable maps \(f:C\to X\) from a connected stable curve \(C\) of genus \(g\) with \(n\) marked smooth points to a projective manifold \(X\) has turned out to be an essential ingredient. The notion of a family of \(n\)-pointed stable maps to \(X\), parametrized by a scheme \(S\) over \(\mathbb{C}\), leads to various contravariant functors. In particular, for any fixed homology class \(\beta\in H_0(X,\mathbb{Z})\), there is the natural contravariant functor \[ \overline{\mathcal M}_{g,n}(X,\beta): (\text{Schemes}/\mathbb{C})\to\text{(Sets)} \] of \(n\)-pointed stable maps from stable curves of genus \(g\) to \(X\) with prescribed homology class \(\beta\). A well-known theorem of M. Kontsevich states that \(\overline{\mathcal M}_{g,n}(X,\beta)\) is actually a proper stack over \(\mathbb{C}\), and that \(\overline{\mathcal M}_{0,n} (\mathbb{P}^r,\beta)\) is even a smooth stack. Moreover, by a result of V. Alexeev it is also known that the stack \(\overline{\mathcal M}_{g,n}(X,\beta)\) has a coarse moduli space \(\overline M_{g,n}(X,\beta)\) which is a projective scheme over \(\mathbb{C}\). In this context, the paper under review is devoted to a closer study of the (smooth) stack \(\overline{\mathcal M}_{0,0}(\mathbb{P}^n,d)\) and its moduli space \(\overline M_{0,0}(\mathbb{P}^n,d)\), where, generically, rational curves of degree \(d\) in \(\mathbb{P}^n\) are the main objects of investigation. More precisely, the authors develop an approach to computing the cohomology ring of the stack \(\overline{\mathcal M} (\mathbb{P}^n,d)\) by generators and relations. In view of the fact that the only previous result in this direction is the computation of the Betti numbers of \(\overline{\mathcal M}_{0,n}(\mathbb{P}^n,d)\) due to \textit{E. Getzler} and \textit{R. Pandharipande} [The Poincaré polynomial of Kontsevich's space of stable maps \(\overline{\mathcal M}_{0,n}(\mathbb P^n,d)\); to appear], the authors' contribution provided by the present paper is highly novel and pioneering. Namely, extending the method of using \(\mathbb{C}^*\)-actions for studying integrals over spaces of stable maps, as initiated by \textit{E. Akyildiz} and \textit{J. B. Carrell} [Manuscr. Math. 58, 473--486 (1987; Zbl 0626.14017)], to the stack-theoretic setup of the case under consideration, and applying some suitable hypercohomology with respect to Koszul complexes, the authors obtain an explicit description of the ``big cell'' à la \textit{A. Bialynicki-Birula} [Ann. Math. (2) 98, 480--497 (1973; Zbl 0275.14007)] in \(\overline{\mathcal M}_{0,0}(\mathbb{P}^n,d)\). Actually, the big cell in \(\overline{\mathcal M}_{0,0} (\mathbb{P}^n,d)\) turns out to be a vector bundle over the moduli space \(\overline M_{0,d}\). As the space \(\overline M_{0,d}\) of stable rational curves with \(d\) marked points is comparatively well understood, the authors' result provides a promising method to tackle the cohomology ring of the stack \(\overline{\mathcal M}_{0,0}(\mathbb{P}^n,d)\) by studying this special fibration. In the latter part of their paper, the authors focus on the particularly simple case \(d=3\), where \(\overline M_{0,d}\) is just a point. Their computations, in this special case, lead to a conjectural description of the cohomology ring of \(\overline{\mathcal M}_{0,0}(\mathbb{P}^n,3)\) by generators and relations, which has been verified for \(n\geq 5\) by computer algebra. On the other hand, as the cohomology ring of \(\overline{\mathcal M}_{0,0}(\mathbb{P}^n,d)\) stabilizes as \(n\) increases, the limit ring is a just as interesting object, and the main concrete result of the present paper is a complete description of this limit denoted by \(H^* (\overline {\mathcal M}_{0,0}(\mathbb{P}^\infty,3),\mathbb{C})\). In fact, the generators of this ring appear as expressions in the Chern classes of certain canonical vector bundles, and the ring itself is reduced and four-dimensional. Finally, the much easier case \(d=2\) is discussed with complete results. In the course of their work, the authors develop both the conceptual and the technical framework in great detail, including the theory of equivariant vector fields on varieties with \(\mathbb{C}^*\)-action, the generalization of the results of Akyildiz and Carrell to stacks, Bialynicki-Birula's construction of the big cell with respect to a \(\mathbb{C}^*\)-action, and the relations to de Rham cohomology. Thus, despite its highly advanced character, this comprehensive paper is written in a very lucid, intelligible and enlightening style.