Let Y be a smooth complex projective Calabi{Yau threefold. Donaldson-Thomas invariants [Tho00] are integer invariants that virtually enumerate curves on Y. They are organised in a generating series DT(Y) that is interesting from a variety of perspectives. For example, well-known series in mathematics and physics appear in explicit computations. Furthermore, closer to the topic of this thesis, the generating series of birational Calabi-Yau threefolds determine one another [Cal16a]. The crepant resolution conjecture for Donaldson-Thomas invariants [BCY12] conjectures another such comparison result. It relates the Donaldson{Thomas generating series of a certain type of three-dimensional Calabi-Yau orbifold to that of a particular resolution of singularities of its coarse moduli space. The conjectured relation is an equality of generating series. In this thesis, I first provide a counterexample showing that this conjecture cannot hold as an equality of generating series. I then verify that both generating series are the Laurent expansion about different points of the same rational function. This suggests a reinterpretation of the crepant resolution conjecture as an equality of rational functions. Second, following a strategy of Bridgeland [Bri11] and Toda [Tod10a, Tod13, Tod16a], I prove a wall-crossing formula in a motivic Hall algebra relating the Hilbert scheme of curves on the orbifold to that on the resolution. I introduce the notion of pair object associated to a torsion pair, putting ideal sheaves and stable pairs on the same footing, and generalise the wall-crossing formula to this setting, essentially breaking the former in many pieces. Pairs, and their wall-crossing formula, are fundamentally objects of the bounded derived category of the Calabi-Yau orbifold. Finally, I present joint work with J. Calabrese and J. Rennemo [BCR] in which we use the wall-crossing formula and Joyce's integration map to prove the crepant resolution conjecture for Donaldson-Thomas invariants as an equality of rational functions. A crucial ingredient is a result of J. Rennemo that detects when two generating functions related by a wall-crossing are expansions of the same rational function.

Gromov-Witten invariants play a crucial role in symplectic- and enumerative Geometry as well as topological String Theory. Essentially, theseinvariants are a count of (pseudo)holomorphic curves of a given genus,going through n-marked points on a symplectic manifold. In the last fewdecades, this has been a huge research topic for both physicists as well asmathematicians, and breakthroughs in calculation techniques have beenmade using Mirror Symmetry. We investigate and explicitly calculateclosed genus zero Gromov-Witten invariants of toric Calabi-Yau threefolds, namely O(−3) → P2 and the resolved conifold. This will be doneby using localization techniques, mirror symmetry and the so called diskpartition function.

Diese Arbeit widmet sich der Untersuchung von zwei Problemen der abzählenden Geometrie im Zusammenhang mit linearen Systemen auf algebraischen Kurven. Das erste Problem besteht darin, die Frage der Gültigkeit der Jonquières-Formeln zu klären. Diese Formeln berechnen die Anzahl von Divisoren mit vorgeschriebener Multiplizität, genannt de Jonquières-Divisoren, die in einem linearen System auf einer glatten projektiven Kurve enthalten sind. Um dies zu tun, konstruieren wir den Raum der de Jonquières-Divisoren als einen Determinantenzyklus des symmetrischen Produkts der Kurve und beweisen, dass er für eine allgemeine Kurve die erwartete Dimension hat. Dabei beschreiben wir die Degenerationen der Jonquières-Divisoren zu den Knotenkurven sowohl mit linearen Systemen als auch mit kompaktifizierten Picard-Schemata. Das zweite Problem behandelt Zyklen von Untergeordneten-, oder allgemeiner, Sekanten-Divisoren zu einem gegebenen linearen System auf einer Kurve. Wir betrachten den Durchschnitt zweier solcher Zyklen, die Sekanten-Divisoren von zwei verschiedenen linearen Systemen auf der gleichen Kurve entsprechen, und untersuchen die Gültigkeit der enumerativen Formeln, die die Anzahl der Teiler im Durchschnitt zählen. Wir untersuchen einige interessante Fälle mit unerwarteten Transversalitätseigenschaften und etablieren eine allgemeine Methode, um zu überprüfen, wann dieser Durchschnitt leer ist. / This thesis is dedicated to the study of two enumerative geometry problems in the context of linear series on algebraic curves. The first problem is that of settling the issue of the validity of the de Jonquières formulas. These formulas compute the number of divisors with prescribed multiplicity, or de Jonquières divisors, contained in a linear series on a smooth projective curve. To do so, we construct the space of de Jonquières divisors as a determinantal cycle of the symmetric product of the curve and prove that, for a general curve with a general linear series, it is of expected dimension. In doing so, we describe the degenerations of de Jonquières divisors to nodal curves using both limit linear series and compactified Picard schemes. The second problem deals with cycles of subordinate or, more generally, secant divisors to a given linear series on a curve. We consider the intersection of two such cycles corresponding to secant divisors of two different linear series on the same curve and investigate the validity of the enumerative formulas counting the number of divisors in the intersection. We study some interesting cases, with unexpected transversality properties, and establish a general method to verify when this intersection is empty.

We establish the existence of a symmetry within the Gromov-Witten theory of CPn and its blowup along points. The nature of this symmetry is encoded in the Cremona transform and its resolution, which lives on the toric variety of the permutohedron. This symmetry expresses some difficult to compute invariants in terms of others less difficult to compute. We focus on enumerative implications; in particular this technique yields a one line proof of the uniqueness of the rational normal curve. Our method involves a study of the toric geometry of the permutohedron, and degeneration of Gromov-Witten invariants. / En este trabajo establecemos la existencia de una simetra en el marco de la teora de Gromov-Witten para CPn y su explosion a lo largo de puntos. La naturaleza de esta simetra queda codicada en la transformacion de Cremona y su resolucion en una variedad torica del permutoedro. Esta simetra expresa algunos invariantes difciles de calcular junto con otros que no lo son tanto. Nos centramos en implicaciones enumerativas; en particular esta tecnica ofrece una prueba enuna lnea de la unicidad de la curva racional normal. Nuestro metodo involucra un estudio de la geometra torica del permutoedro, as como el de la degeneracion de los invariantes de Gromov-Witten.