Title: Solving diophantine equations by factorization in number fields Author: Bc. Maroš Hrnčiar Department: Department of Algebra Supervisor: Mgr. Vítězslav Kala, Ph.D., Mathematical Institute, University of Göttingen Abstract: The question of solvability of diophantine equations is one of the oldest mathematical problems in the history of mankind. While different approaches have been developed for solving certain types of equations, this thesis predo- minantly deals with the method of factorization over algebraic number fields. The idea behind this method is to express the equation in the form L = yn where L equals a product of typically linear factors with coefficients in a particular number field. Provided that several assumptions are met, it follows that each of the factors must be the n-th power of an element of the field. The structure of number fields plays a key role in the application of this method, hence a crucial part of the thesis presents an overview of algebraic number theory. In addition to the general theoretical part, the thesis contains all the necessary computations in specific quadratic and cubic number fields describing their basic characteristics. However, the main objective of this thesis is solving specific examples of equati- ons. For instance, in the case of equation x2 + y2 = z3 we...

Within-population behavioural variation can greatly affect the ecology of a species and the outcome of evolutionary processes. This study aimed to determine how variable sperm whale social and vocal behaviour is between both individuals and their social units. The population of whales off Dominica is small and isolated from communities in neighbouring waters. Female and immature whales live together in social units containing about 7 animals. I analysed their social relationships and their ‘coda’ communication signals using an unparalleled dataset of social and vocal interactions at the level of the individual. Within units, calves were significant nodes in their social unit’s network, and thus I provide quantitative support for the hypothesis that communal calf care acts as the primary evolutionary driver for group formation in this species. Social relationships within and between units were diverse, while the spatial spread of individuals within units and their travel speeds were similar among all of the units. I identified long-term patterns of association between units consistent over decadal time scales. Social units had characteristic vocal repertoires, but all were dominated by the ‘1+1+3’ and ‘5R’ coda types. Differences between units resulted from some units using specific 4-click coda types. Units and individuals used different accents on their ‘5R’ codas, but the ‘1+1+3’ coda was stereotyped across all individuals and units studied. The repertoires of different units were as similar as units within vocal clans in the Pacific. My results support the hypothesis that the ‘5R’ coda may function in individual identification. The stability of the ‘1+1+3’ coda may be the result of selection for a marker of clan membership. Individual repertoires differed consistently across years; and contrary to an existing hypothesis, new mothers did not vary their repertoire to be more distinct after giving birth. However, calves did use a class-specific ‘3+1’ coda. In summary, sperm whale social and vocal behaviour vary between individuals and among units. Variation in the social and vocal behaviour of female sperm whales results from a trade-off between individuality and conformity within units and clans.

Artins Vermutung über Primitivwurzeln besagt, dass es zu jeder ganzen Zahl a, die weder 0, ±1 noch eine Quadratzahl ist, unendlich viele Primzahlen p gibt, sodass a eine Primitivwurzel modulo p ist, d.h. a erzeugt eine multiplikative Untergruppe von Q*, dessen Reduktion modulo p Index 1 in (Z/pZ)* hat. Dies wirft die Frage nach Verteilung von Index und Ordnung dieser Reduktion in (Z/pZ)* auf, wenn man p variiert. Diese Arbeit widmet sich verallgemeinerten Fragestellungen in Zahlkörpern: Ist K ein Zahlkörper und Gamma eine endlich erzeugte unendliche Untergruppe von K*, so werden Momente von Index und Ordnung der Reduktion von Gamma sowohl modulo bestimmter Familien von Primidealen von K als auch modulo aller Ideale von K untersucht. Ist Gamma die Gruppe der Einheiten von K, so steht diese Fragestellung in engem Zusammenhang mit der Ramanujan Vermutung in Zahlkörpern. Des Weiteren werden analoge Probleme für rationale elliptische Kurven E betrachtet: Bezeichnet Gamma die von einem rationalen Punkt von E erzeugte Gruppe, so wird untersucht, wie sich Index und Ordnung der Reduktion von Gamma modulo Primzahlen verhalten. Teilweise unter Voraussetzung gängiger zahlentheoretischer Vermutungen werden jeweils asymptotische Formeln in manchen Fällen bewiesen und generelle Schwierigkeiten geschildert, die solche in anderen Fällen verhindern. Darüber hinaus wird eine weitere verwandte Fragestellung betrachtet und bewiesen, dass zu jeder hinreichend großen Primzahl p stets eine Primitivwurzel modulo p existiert, die sich als Summe von zwei Quadraten darstellen lässt und nach oben im Wesentlichen durch die Quadratwurzel von p beschränkt ist.

Les algèbres centrales simples ont de nombreuses applications en théorie des nombres, mais leur algorithmique est encore peu développée. Dans cette thèse, j’apporte une contribution dans deux directions. Premièrement, je présente des algorithmes de complexité prouvée, ce qui est nouveau dans la plupart des cas. D’autre part, je développe des algorithmes heuristiques mais très efficaces dans la pratique pour les exemples qui nous intéressent le plus, comme en témoignent mes implantations. Les algorithmes sont à la fois plus rapides et plus généraux que les algorithmes existants. Plus spécifiquement, je m’intéresse aux problèmes suivants : calcul du groupe des unités d’un ordre et problème de l’idéal principal. Je commence par étudier le diamètre du domaine fondamental de certains groupes d’unités grâce à la théorie des représentations. Je décris ensuite un algorithme prouvé pour calculer des générateurs et une présentation du groupe des unités d’un ordre maximal dans une algèbre à division, puis un algorithme efficace qui calcule également un domaine fondamental dans le cas où le groupe des unités est un groupe kleinéen. Je donne en outre un algorithme de complexité prouvée qui détermine si un idéal d’un tel ordre est principal, et qui en calcule un générateur le cas échéant, puis je décris un algorithme heuristiquement sous-exponentiel pour résoudre le même problème dans le cas d’une algèbre de quaternions indéfinie. / Central simple algebras have many applications in number theory, but their algorithmic theory is not yet fully developed. I present algorithms to compute effectively with central simple algebras that are both faster and more general than existing ones. Some of these algorithms have proven complexity estimates, a new contribution in this area; others rely on heuristic assumptions but perform very efficiently in practice.Precisely, I consider the following problems: computation of the unit group of an order and principal ideal problem. I start by studying the diameter of fundamental domains of some unit groups using representation theory. Then I describe an algorithm with proved complexity for computing generators and a presentation of the unit group of a maximal order in a division algebra, and then an efficient algorithm that also computes a fundamental domain in the case where the unit group is a Kleinian group. Similarly, I present an algorithm with proved complexity that decides whether an ideal of such an order is principal and that computes a generator when it is. Then I describe a heuristically subexponential algorithm that solves the same problem in indefinite quaternion algebras.