Sous la direction de Italie) Università degli studi? (Pise, Yuri Bilu Thèse soutenue le 04 décembre 2009: Bordeaux 1 Le but de cette thèse est d'obtenir des versions totalement explicite de deux résultats fondamentales sur les revêtements de courbes algébriques: le Théorème d'existence de Riemann et le théorème de Chevalley-Weil. La motivation de notre travail sur le Théorème d'existence de Riemann réside dans le domaine de l'analyse diophantienne effective, lorsque la technique des revêtements est largement utilisé: trés souvent il arrive qu'on ne connait que le degré du revêtement et les points de ramification, et pour travailler avec le revêtement il faut en avoir une description efficace. Le théorème de Chevalley-Weil est également indispensable dans l'analyse diophantienne, car il permet de réduire un problème diophantien sur la variété V à celui sur le revêtement W, ce qui peut être plus simple à étudier. Dans la thèse on obtient une version du théorème de Chevalley-Weil en dimension 1, explicite en tous les paramètres et nettement meilleur que les versions précédentes. -Théorème de Chevalley-Weil explicite -Théorème d'existence de Riemann explicite -Analyse diophantienne effective The purpose of this thesis is to obtain totally explicit versions for two fundamental results about coverings of algebraic curves: the Riemann Existence Theorem and the Chevalley-Weil Theorem. The motivation behind our work about Riemann Existence Theorem lies in the field of effective Diophantine analysis, where the covering technique is widely used: it happens quite often that only the degree of the covering and the ramification points are known, and to work with the covering curve, one needs to have an effective description of it. The Chevalley-Weil theorem is also indispensable in the Diophantine analysis because it reduces a Diophantine problem on the variety V to that on the covering variety W, which can often be simpler to deal. In the thesis we obtain a version of the Chevalley-Weil theorem in dimension 1, explicit in all parameters and considerably sharper than the previous versions. -Explicit Riemann Existence Theorem -Explicit Chevalley-Weil Theorem -Diophantine analysis Source: http://www.theses.fr/2009BOR13895/document
Introduction The purpose of this thesis is to obtain totally explicit versions for two fun-damental results about coverings of algebraic curves: theRiemann Existence Theoremand theChevalley-Weil Theorem the introduction we briefly. In recall basic facts on these two theorems. The detailed statements, which require certain amount of notation, can be found in the introductions of the corresponding chapters of the thesis. The Riemann Existence Theorem TheRiemann Existence Theoremasserts that every compact Riemann sur-face is (analytically isomorphic to) a complex algebraic curve. In other words, the field of meromorphic functions on a compact Riemann surfaceSis finitely generated and of transcendence degree 1 overC. One of the most common ways of defining Riemann surfaces is realizing them as finite ramified coverings of the Riemann sphereP1(C). Moreover, even if the covering is purely topological, theC-analytic structure on the Riemann sphere lifts, in a unique way, to the covering surface. Thus, the Riemann Existence Theorem can be restated as follows. Theorem ALetMbe a finite subset ofP1(C). Then for any finite covering ofP1(C)closed oriented surface, unramified outside the setby a M, there exists a complex algebraic curveCand a rational functionx∈C(C)such that our covering is isomorphic1toC(C)x→P1(C), the covering defined byx. Moreover, the couple(C x)unique up to a naturally defined isomorphismis 2. We refer to [6] for several more precise statements, and for the connection of the Riemann Existence Theorem and the Inverse Galois Problem. One of the purposes of this thesis is to give an effective description of the curveC, or, more precisely, of the couple (C x), in terms of the degree of the initial topological covering and the set M of the ramification points, ¯ provided the points from that set are defined over the fieldQof all algebraic ¯ numbers. In this case the curveCis also defined overQ(this is the “easy” 1Two coveringsS1π→1SandS2π→2Sof topological spaces areisomorphicif there exists a homeomorphismS1ϕ→S2such thatπ1=π2◦ϕ. 2If (C0, x0) is another such couple, then the field isomorphismC(x)→C(x0) given by x7→x0, extends to a field isomorphismC(C)→C(C0).
4
¯ direction of Belyi’s Theorem). We produce a plane model ofCoverQ, such that one of the coordinates isx, and we give explicit bounds for the degree and the height of the defining equation of this model, and of the degree and discriminant of the number field over which this model is defined. Notice that we do not produce a new proof of the Riemann Existence The-orem. In fact, we do use both the existence and the uniqueness statements of Theorem A. The principal motivation of this work lies in the field of effective Diophan-tine analysis, where the covering technique is widely used. It happens quite often that only the degree of the covering and the ramification points are known, and to work with the covering curve, one needs to have an effective description of it. In particular, in Chapter 3 we use our explicit version of the Riemann ex-istence theorem to get a user-friendly version of the Chevalley-Weil theorem, see Theorem 3.1.5.
The Chevalley-Weil Theorem The Chevalley-Weil theorem is one of the most basic principles of the Dio-phantine analysis. Already Diophantus of Alexandria routinely used reason-ing of the kind “ifaandbare ‘almost’ co-prime integers andabis a square, then each ofaandb Chevalley-Weil theoremis ‘almost’ a square”. The provides a general set-up for this kind of arguments. Theorem B (Chevalley-Weil)LetVe→φVvecongri´eteletaebinfiafo normal projective varieties, defined over a number fieldK there exists. Then e e ¯ a non-zero integerTsuch that for anyP∈V(K)andP∈V(K)such that e e φ(P) =P, the relative discriminant ofK(P)/K(P)dividesT. There is also a similar statement for coverings of affine varieties and inte-gral points. See[17, Section 2.8] for more details. The Chevalley-Weil theorem is indispensable in the Diophantine analysis, because it reduces a Diophantine problem on the varietyVto that on the e covering varietyV particular, the, which can often be simpler to deal. In Chevalley-Weil theorem is used, sometimes implicitly, in the proofs of the great finiteness theorems of Mordell-Weil, Siegel and Faltings. In view of all this, a quantitative version of the Chevalley-Weil theorem, at least in dimension 1, would be useful to have. One such version appears 5
in Chapter 4 of [1], but it is not explicit in all parameters; neither is the version recently suggested by Draziotis and Poulakis [9, 10], who also make some other restrictive assumptions (see Remark 3.1.3 in Chapter 3 for more on this). In the thesis we obtain a version of the Chevalley-Weil theorem in dimen-sion 1, which is explicit in all parameters and considerably sharper than the previous versions. Our approach is different from that of [9, 10], and goes back to [1, 2]. For the precise statement of our results see the introduction of Chapter 3.
Plan of the thesis In Chapter 1 we collect auxiliary facts of diverse nature, which are used throughout the thesis. In Chapter 2 we obtain an explicit version of the Riemann Existence Theorem over a number field. This chapter is based on the article [3], joint with Yu. Bilu. InChapter 3 we obtain several explicit versions of the Chevalley-Weil Theorem for curves. This chapter is based on the article [4], joint with Yu. Bilu and A. Surroca.