Logarithmic Gromov-Witten invariants

By Mark Gross, Bernd S Siebert. 2011

267 citations recorded by the source corpus.

ing, now let π : Y → W be a proper morphism of schemes and let αi : Mi → OY , i = 1, 2, be two fine saturated log structures on Y . Consider the functor (2.1) LMorY/W (M1, M2) : (Sch/W ) −→ (Sets) that on objects is defined by V %−→ { φ : (M1)V → (M2)V morphism of log structures } . Then the statement that W ×M×M(X) M (X) is an algebraic space10 essentially is a special case of the following proposition. Proposition 2.9. LMorY/W (M1, M2) is represented by an algebraic space locally of finite type over W . The proof is provided, after some preparations, at the end of this subsection. This then finishes the proof of Proposition 2.3 in the case S = Spec k with the trivial log structure. In the general case we have in addition two morphisms of log structures ψi : πMS → Mi, π : C → W , the projection, and we need to restrict to those φ : (M1)V → (M2)V compatible with ψi. But by Proposition 2.9 composition with ψ1 defines a morphism of algebraic spaces LMorY/W (M1, M2) −→ LMorY/W (πMS , M2). Now W ×M×M(X) M (X) arises as the fibre product with the morphism W −→ LMorY/W (πMS , M2) defined by ψ2, and is hence represented by an algebraic space. This finishes the proof of Proposition 2.3 also in the general case. Remark 2.10. One problem in showing representability of LMorY/W (M1, M2) is that it is non-separated, essentially because the induced map M1 → M2 cannot be determined by its restriction to an open dense subset. As an example (cf. [Ol2], Remark 3.12) consider the log structure M on A1 = Spec k[x] with chart N −→ k[x], (a, b) %−→ x. The map N2 → N2, (a, b) %→ (b, a) induces a non-trivial automorphism of M that restricts to the identity on A1 \ {0}. To find an etale cover of the algebraic space representing LMorY/W (M1, M2) in Proposition 2.9 we thus first restrict the map φ. To this end let w → W be a geometric point and let φw : (M1)w → (M2)w be a choice of φ over one geometric fibre Yw of Y → W . Now since Mi are fine sheaves the choice at a geometric point x → Y determines φ at any generization y of x. Moreover, if y specializes to some other point z such that the generization map Mi,z → Mi,y is an isomorphism, then φ is also determined at z. Iterating the generization-specialization process we are lead to the following definition. 10As for algebraic stacks we have to drop the condition of quasi-separatedness from the definition of algebraic spaces ([Kt], Ch.II, Definition 1.1). Licensed to Brown Univ. Prepared on Fri Oct 4 09:27:34 EDT 2013 for download from IP 128.148.231.12. License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use 474 MARK GROSS AND BERND SIEBERT Definition 2.11. Let M be a fine sheaf on a scheme Y and let A ⊂ |Y | be a set of geometric points. We say x ∈ |Y | has property (Agen) with respect to M if there exists a sequence y1, z1, . . . , yr, zr ∈ |Y | for some r with the following properties: y1 ∈ A, zr = x, yi ∈ cl(zi), i = 1, . . . , r, yi ∈ cl(zi−1) and Myi → Mzi−1 is an isomorphism, i = 2, . . . , r. Thus the giving of φ on a closed subset A ⊂ |Y | then determines φ also on the subset UA := { x ∈ |Y | ∣x fulfills (Agen) } of |Y |. Note that by definition UA is closed under generization. Since M is a fine sheaf it is also immediate that UA is a constructible subset of |Y |, and hence UA ⊂ |Y | is open. Since the statement of Proposition 2.9 is local in W and by properness of Y → W we may assume any point of |Y | fulfills (Agen) for A = Yw with respect both to M1 and to M2, that is, UYw = Y . Then for any V → W there is at most one φ : (M1)V → (M2)V compatible with φw under sequences of generization maps. Let us call such φ (or a lift φ to a morphism of log structures) compatible with φw, and similarly for any A ⊂ |Y |. Note that φw may not extend to Y , but it may do so after certain base changes. We first treat the representability problem locally on Y , that is, for Y = W . Lemma 2.12. Let Y = W and suppose that there exists a closed subset A ⊂ Y such that any x ∈ |Y | fulfills (Agen) with respect to both Mi (Definition 2.11). Let φA : (M1)A −→ (M2)A be a homomorphism of sheaves of monoids. Then the functor LMor φA Y : (Y ′ f −→Y ) %−→ { φ : (Y ′, fM2)→(Y ′, fM1) ∣φ is compatible with φA } is represented by a scheme LMorA Y of finite type and affine over Y . Proof. It is sufficient to prove the statement on an etale open cover of Y , since we can then use descent for affine morphisms ([SGA1], VIII, Theorem 2.1) to obtain a scheme over Y . Thus we can assume that we in fact have charts ψi : Pi → Γ(Y, Mi) for the two log structures. We can also assume that φA is induced by a homomorphism of monoids φA : P1 → P2. Let p1, . . . , pn ∈ P1 be a generating set for P1 as a monoid. Consider the sheaf of finitely generated OY -algebras FY := OY [P gp 1 ]/〈α1(ψ1(pi)) − zα2(ψ2(φA(pi))) | 1 ≤ i ≤ n〉. Then the desired scheme is LMorA Y := SpecFY . To see that this is the correct scheme, suppose f : Y ′ → Y is given. We wish to show that giving a commutative diagram of schemes

Subfields: Algebra and Number Theory, Geometry and Topology

View Paper: https://www.semanticscholar.org/paper/ea2bf1cded6fdb7bf8a050ac05543b95a8ae39cf

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Algebra and Number Theory

In Geometry and Topology

Citations in the corpus

Relative and orbifold Gromov-Witten invariants (2010, background): “sor, both based on the theory of logarithmic geometry. The ﬁrst of these was deﬁned by B. Kim by introducing logarithmic structures into J. Li’s degenerations [Kim08]. Later work of Gross and Siebert [GS13] and Abramovich and Chen [Che14, AC14] made it possible to study logarithmic stable maps witho”
Brill-Noether theory for curves of a fixed gonality (2017, methodology): “The proof of the main lifting theorems, as well as the analysis of generic chains of cycles in the following section, will use the geometry of logarithmic prestable maps to toric varieties [2, 17, 28] and their relationship to Berkovich geometry [45, 46, 50].”
Logarithmic Gromov–Witten theory with expansions (2019, methodology): “can be applied to construct the spaces in Theorem A in a manner that is logically independent from the foundational papers in logarithmic Gromov–Witten theory [2, 20, 30], and this is explained by Carocci–Nabijou [15].”
Descendant log Gromov-Witten invariants for toric varieties and tropical curves (2016, methodology): “[GS13], [AC14]), although for the cases we consider this will turn out to coincide with a naive algebraic count (possibly with some fractional multiplicities when g > 0, cf.”
Mirror symmetry for log Calabi-Yau surfaces I (2011, methodology): “many of the results in this paper will extend to Looijenga pairs of all dimensions. This generalization will require the further development of the technology of logarithmic Gromow-Witten invariants, [GS11],[AC11]. We should obtain in complete generality a mirror family X→Speckd[P] for suitable mono”
Decomposition of degenerate Gromov–Witten invariants (2017, methodology): “ith a special emphasis on the relationship with tropical geometry. We develop tropical geometry in the setup of generalized cone complexes, introduced in §2.1. While this point of view was present in [GS1], we make it more explicit here, and in particular discuss tropicalization in a suﬃcient degree”
The tropicalization of the moduli space of curves (2012, unknown): “The cone complex of [KKMSD73] is further generalized to logarithmic schemes in [GS11, Appendix B], where the complex associated to a logarithmic scheme X is called the tropi-calization of X .”
Comparison theorems for Gromov–Witten invariants of smooth pairs and of degenerations (2012, methodology): “The impact of the degeneration formula has been profound, but efforts to generalize the theory to other degenerations have until recently met with limited success (see [34, 33, 16, 20, 3] for recent progress).”
Orbifold techniques in degeneration formulas (2011, methodology): “heory by one induced by the natural log structures, with a degeneration formula worked out by Q. Chen in [Chen1] and virtual localization worked out by Molcho and Routis [MR]. M. Gross and B. Siebert [GS], Chen [Chen] and Abramovich and Chen [AC] have developed a logarithmic theory without expansion”
Skeletons and Fans of Logarithmic Structures (2015, methodology): “hed when the characteristic monoid M X is globally generated in [AC14,Che14], and more generally when the associated group Mgp X is globally generated 32 ABRAMOVICH, CHEN, MARCUS, ULIRSCH AND WISE in [GS13]. The strategy used in [ACMW] for the general setting is to reduce to the case of a globally g”