Soit $p_1 = 2, p_2 = 3, p_3 = 5,\ldots$ la suite des nombres premiers, et soient $q \ge 3$ et $a$ des entiers premiers entre eux. R\'ecemment, Daniel Shiu a d\'emontr\'e une ancienne conjecture de Sarvadaman Chowla. Ce dernier a conjectur\'e qu'il existe une infinit\'e de couples $p_n,p_$ de premiers cons\'ecutifs tels que $p_n \equiv p_{n+1} \equiv a \bmod q$. Fixons $\epsilon > 0$. Une r\'ecente perc\'ee majeure, de Daniel Goldston, J\`anos Pintz et Cem Y{\i}ld{\i}r{\i}m, a \'et\'e de d\'emontrer qu'il existe une suite de nombres r\'eels $x$ tendant vers l'infini, tels que l'intervalle $(x,x+\epsilon\log x]$ contienne au moins deux nombres premiers $\equiv a \bmod q$. \'Etant donn\'e un couple de nombres premiers $\equiv a \bmod q$ dans un tel intervalle, il pourrait exister un nombre premier compris entre les deux qui n'est pas $\equiv a \bmod q$. On peut d\'eduire que soit il existe une suite de r\'eels $x$ tendant vers l'infini, telle que $(x,x+\epsilon\log x]$ contienne un triplet $p_n,p_{n+1},p_{n+2}$ de nombres premiers cons\'ecutifs, soit il existe une suite de r\'eels $x$, tendant vers l'infini telle que l'intervalle $(x,x+\epsilon\log x]$ contienne un couple $p_n,p_{n+1}$ de nombres premiers tel que $p_n \equiv p_{n+1} \equiv a \bmod q$. On pense que les deux \'enonc\'es sont vrais, toutefois on peut seulement d\'eduire que l'un d'entre eux est vrai, sans savoir lequel. Dans la premi\`ere partie de cette th\`ese, nous d\'emontrons que le deuxi\`eme \'enonc\'e est vrai, ce qui fournit une nouvelle d\'emonstration de la conjecture de Chowla. La preuve combine des id\'ees de Shiu et de Goldston-Pintz-Y{\i}ld{\i}r{\i}m, donc on peut consid\'erer que ce r\'esultat est une application de leurs m\'thodes. Ensuite, nous fournirons des bornes inf\'erieures pour le nombre de couples $p_n,p_{n+1}$ tels que $p_n \equiv p_{n+1} \equiv a \bmod q$, $p_{n+1} - p_n < \epsilon\log p_n$, avec $p_{n+1} \le Y$. Sous l'hypoth\`ese que $\theta$, le \og niveau de distribution \fg{} des nombres premiers, est plus grand que $1/2$, Goldston-Pintz-Y{\i}ld{\i}r{\i}m ont r\'eussi \`a d\'emontrer que $p_{n+1} - p_n \ll_{\theta} 1$ pour une infinit\'e de couples $p_n,p_$. Sous la meme hypoth\`ese, nous d\'emontrerons que $p_{n+1} - p_n \ll_{q,\theta} 1$ et $p_n \equiv p_{n+1} \equiv a \bmod q$ pour une infinit\'e de couples $p_n,p_$, et nous prouverons \'egalement un r\'esultat quantitatif. Dans la deuxi\`eme partie, nous allons utiliser les techniques de Goldston-Pintz-Yldrm pour d\'emontrer qu'il existe une infinit\'e de couples de nombres premiers $p,p'$ tels que $(p-1)(p'-1)$ est une carr\'e parfait. Ce resultat est une version approximative d'une ancienne conjecture qui stipule qu'il existe une infinit\'e de nombres premiers $p$ tels que $p-1$ est une carr\'e parfait. En effet, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n \le Y$ tels que $n = \ell_1\cdots \ell_r$, avec $\ell_1,\ldots,\ell_r$ des premiers distincts, et tels que $(\ell_1-1)\cdots (\ell_r-1)$ est une puissance $r$-i\`eme, avec $r \ge 2$ quelconque. \'Egalement, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n = \ell_1\cdots \ell_r \le Y$ tels que $(\ell_1+1)\cdots (\ell_r+1)$ est une puissance $r$-i\`eme. Finalement, \'etant donn\'e $A$ un ensemble fini d'entiers non-nuls, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n \le Y$ tels que $\prod_ (p+a)$ est une puissance $r$-i\`eme, simultan\'ement pour chaque $a \in A$. / Let $p_1 = 2, p_2 = 3, p_3 = 5,\ldots$ be the sequence of all primes, and let $q \ge 3$ and $a$ be coprime integers. Recently, and very remarkably, Daniel Shiu proved an old conjecture of Sarvadaman Chowla, which asserts that there are infinitely many pairs of consecutive primes $p_n,p_{n+1}$ for which $p_n \equiv p_{n+1} \equiv a \bmod q$. Now fix a number $\epsilon > 0$, arbitrarily small. In their recent groundbreaking work, Daniel Goldston, J\`anos Pintz and Cem Y{\i}ld{\i}r{\i}m proved that there are arbitrarily large $x$ for which the short interval $(x, x + \epsilon\log x]$ contains at least two primes congruent to $a \bmod q$. Given a pair of primes $\equiv a \bmod q$ in such an interval, there might be a prime in-between them that is not $\equiv a \bmod q$. One can deduce that \emph{either} there are arbitrarily large $x$ for which $(x, x + \epsilon\log x]$ contains a prime pair $p_n \equiv p_{n+1} \equiv a \bmod q$, \emph{or} that there are arbitrarily large $x$ for which the $(x, x + \epsilon\log x]$ contains a triple of consecutive primes $p_n,p_{n+1},p_{n+2}$. Both statements are believed to be true, but one can only deduce that one of them is true, and one does not know which one, from the result of Goldston-Pintz-Y{\i}ld{\i}r{\i}m. In Part I of this thesis, we prove that the first of these alternatives is true, thus obtaining a new proof of Chowla's conjecture. The proof combines some of Shiu's ideas with those of Goldston-Pintz-Y{\i}ld{\i}r{\i}m, and so this result may be regarded as an application of their method. We then establish lower bounds for the number of prime pairs $p_n \equiv p_{n+1} \equiv a \bmod q$ with $p_{n+1} - p_n < \epsilon\log p_n$ and $p_{n+1} \le Y$. Assuming a certain unproven hypothesis concerning what is referred to as the `level of distribution', $\theta$, of the primes, Goldston-Pintz-Y{\i}ld{\i}r{\i}m were able to prove that $p_{n+1} - p_n \ll_{\theta} 1$ for infinitely many $n$. On the same hypothesis, we prove that there are infinitely many prime pairs $p_n \equiv p_{n+1} \equiv a \bmod q$ with $p_{n+1} - p_n \ll_{q,\theta} 1$. This conditional result is also proved in a quantitative form. In Part II we apply the techniques of Goldston-Pintz-Y{\i}ld{\i}r{\i}m to prove another result, namely that there are infinitely many pairs of distinct primes $p,p'$ such that $(p-1)(p'-1)$ is a perfect square. This is, in a sense, an `approximation' to the old conjecture that there are infinitely many primes $p$ such that $p-1$ is a perfect square. In fact we obtain a lower bound for the number of integers $n$, up to $Y$, such that $n = \ell_1\cdots \ell_r$, the $\ell_i$ distinct primes, and $(\ell_1 - 1)\cdots (\ell_r - 1)$ is a perfect $r$th power, for any given $r \ge 2$. We likewise obtain a lower bound for the number of such $n \le Y$ for which $(\ell_1 + 1)\cdots (\ell_r + 1)$ is a perfect $r$th power. Finally, given a finite set $A$ of nonzero integers, we obtain a lower bound for the number of $n \le Y$ for which $\prod_{p \mid n}(p+a)$ is a perfect $r$th power, simultaneously for every $a \in A$.

Le sujet de cette thèse est l'étude des progressions arithmétiques dans les nombres entiers. Plus précisément, nous nous intéressons à borner inférieurement v(N), la taille du plus grand sous-ensemble des nombres entiers de 1 à N qui ne contient pas de progressions arithmétiques de 3 termes. Nous allons donc construire de grands sous-ensembles de nombres entiers qui ne contiennent pas de telles progressions, ce qui nous donne une borne inférieure sur v(N). Nous allons d'abord étudier les preuves de toutes les bornes inférieures obtenues jusqu'à présent, pour ensuite donner une autre preuve de la meilleure borne. Nous allons considérer les points à coordonnés entières dans un anneau à d dimensions, et compter le nombre de progressions arithmétiques qu'il contient. Pour obtenir des bornes sur ces quantités, nous allons étudier les méthodes pour compter le nombre de points de réseau dans des sphères à plusieurs dimensions, ce qui est le sujet de la dernière section. / The subject of this thesis is the study of arithmetic progressions in the integers. Precisely, we are interested in the size v(N) of the largest subset of the integers from 1 to N that contains no 3 term arithmetic progressions. Therefore, we will construct a large subset of integers with no such progressions, thus giving us a lower bound on v(N). We will begin by looking at the proofs of all the significant lower bounds obtained on v(N), then we will show another proof of the best lower bound known today. For the proof, we will consider points on a large d-dimensional annulus, and count the number of integer points inside that annulus and the number of arithmetic progressions it contains. To obtain bounds on those quantities, it will be interesting to look at the theory behind counting lattice points in high dimensional spheres, which is the subject of the last section.

Cette thèse concerne quelques problèmes liés à la répartition des entiers sans facteur carré dansles progressions arithmétiques. Ces problèmes s’expriment en termes de majorations du terme d’erreurassocié à cette répartition.Les premier, deuxième et quatrième chapitres sont concentrés sur l’étude statistique des termesd’erreur quand on fait varier la progression arithmétique modulo q. En particulier on obtient une formuleasymptotique pour la variance et des majorations non triviales pour les moments d’ordre supérieur. Onfait appel à plusieurs techniques de théorie analytique des nombres comme les méthodes de crible et lessommes d’exponentielles, notamment une majoration récente pour les sommes d’exponentielles courtesdue à Bourgain dans le deuxième chapitre.Dans le troisième chapitre on s’intéresse à estimer le terme d’erreur pour une progression fixée. Onaméliore un résultat de Hooley de 1975 dans deux directions différentes. On utilise ici des majorationsrécentes de sommes d’exponentielles courtes de Bourgain-Garaev et de sommes d’exponentielles torduespar la fonction de Möbius dues à Bourgain et Fouvry-Kowalski-Michel. / This thesis concerns a few problems linked with the distribution of squarefree integers in arithmeticprogressions. Such problems are usually phrased in terms of upper bounds for the error term relatedto this distribution.The first, second and fourth chapter focus on the satistical study of the error terms as the progres-sions varies modulo q. In particular we obtain an asymptotic formula for the variance and non-trivialupper bounds for the higher moments. We make use of many technics from analytic number theorysuch as sieve methods and exponential sums. In particular, in the second chapter we make use of arecent upper bound for short exponential sums by Bourgain.In the third chapter we give estimates for the error term for a fixed arithmetic progression. Weimprove on a result of Hooley from 1975 in two different directions. Here we use recent upper boundsfor short exponential sums by Bourgain-Garaev and exponential sums twisted by the Möbius functionby Bourgain et Fouvry-Kowalski-Michel.

Made available in DSpace on 2016-04-27T16:58:41Z (GMT). No. of bitstreams: 1 Sebastiao Archilia.pdf: 501341 bytes, checksum: 7ee1e1e3e26a096d63298c4185b60432 (MD5) Previous issue date: 2008-06-30 / Secretaria da Educação do Estado de São Paulo / Today it has been disclosed the poor performance of High School students in learning the Algebra. On the other hand, the results of researches, such as Vale e Pimentel (2005) and Machado (2006) among others, emphasized the importance of working with the observation and generalization of patterns to develop the algebraic thinking, which can help to overcome this problem. This situation and the suggestion help me to decide to investigate if high school students in a situation of patterns observation and generalization could construct an algebraic formulation of a general term of an arithmetic progression. To collect data drafted a didactic sequence based on the assumptions of Didactic Engineering as described by Machado (2008). The didactic sequence occurred in tree sessions with the participation of some of my students, all volunteers. For the conclusion I took into account only the results of the data analysis of 11 students present at all sessions. The results led me to conclude that, although students have expressed in natural language a formula for the general term, it was not enough to convert this result for the symbolic algebraic way / Tem sido amplamente divulgada o mau desempenho dos alunos do Ensino Médio em relação a questões de Matemática e especialmente da Álgebra. Por outro lado, os resultados de pesquisas, como os de Vale e Pimentel (2005) e de Machado (2006), entre outros, enfatizam a importância do trabalho com a observação e generalização de padrões para o desenvolvimento do pensamento algébrico, o que pode auxiliar na superação desse problema. Essa situação e a sugestão me levou a investigar se alunos da segunda série do Ensino Médio frente a atividades de observação e generalização de padrões de seqüências constroem uma fórmula para o termo genérico de uma Progressão Aritmética. Para a coleta de dados, elaborei uma seqüência didática embasada nos pressupostos da Engenharia Didática, conforme descrita por Machado (2008). Realizei três sessões com a participação de alguns de meus alunos, todos voluntários. Para a conclusão levei em conta somente os resultados das análises do desempenho de 11 alunos que estiveram presentes em todas as 3 sessões. Os resultados me levaram a concluir que, embora os alunos tenham expressado em linguagem natural uma fórmula para o termo geral, isso não foi suficiente para converterem esse resultado para uma forma simbólica algébrica

Soit $p_1 = 2, p_2 = 3, p_3 = 5,\ldots$ la suite des nombres premiers, et soient $q \ge 3$ et $a$ des entiers premiers entre eux. R\'ecemment, Daniel Shiu a d\'emontr\'e une ancienne conjecture de Sarvadaman Chowla. Ce dernier a conjectur\'e qu'il existe une infinit\'e de couples $p_n,p_$ de premiers cons\'ecutifs tels que $p_n \equiv p_{n+1} \equiv a \bmod q$. Fixons $\epsilon > 0$. Une r\'ecente perc\'ee majeure, de Daniel Goldston, J\`anos Pintz et Cem Y{\i}ld{\i}r{\i}m, a \'et\'e de d\'emontrer qu'il existe une suite de nombres r\'eels $x$ tendant vers l'infini, tels que l'intervalle $(x,x+\epsilon\log x]$ contienne au moins deux nombres premiers $\equiv a \bmod q$. \'Etant donn\'e un couple de nombres premiers $\equiv a \bmod q$ dans un tel intervalle, il pourrait exister un nombre premier compris entre les deux qui n'est pas $\equiv a \bmod q$. On peut d\'eduire que soit il existe une suite de r\'eels $x$ tendant vers l'infini, telle que $(x,x+\epsilon\log x]$ contienne un triplet $p_n,p_{n+1},p_{n+2}$ de nombres premiers cons\'ecutifs, soit il existe une suite de r\'eels $x$, tendant vers l'infini telle que l'intervalle $(x,x+\epsilon\log x]$ contienne un couple $p_n,p_{n+1}$ de nombres premiers tel que $p_n \equiv p_{n+1} \equiv a \bmod q$. On pense que les deux \'enonc\'es sont vrais, toutefois on peut seulement d\'eduire que l'un d'entre eux est vrai, sans savoir lequel. Dans la premi\`ere partie de cette th\`ese, nous d\'emontrons que le deuxi\`eme \'enonc\'e est vrai, ce qui fournit une nouvelle d\'emonstration de la conjecture de Chowla. La preuve combine des id\'ees de Shiu et de Goldston-Pintz-Y{\i}ld{\i}r{\i}m, donc on peut consid\'erer que ce r\'esultat est une application de leurs m\'thodes. Ensuite, nous fournirons des bornes inf\'erieures pour le nombre de couples $p_n,p_{n+1}$ tels que $p_n \equiv p_{n+1} \equiv a \bmod q$, $p_{n+1} - p_n < \epsilon\log p_n$, avec $p_{n+1} \le Y$. Sous l'hypoth\`ese que $\theta$, le \og niveau de distribution \fg{} des nombres premiers, est plus grand que $1/2$, Goldston-Pintz-Y{\i}ld{\i}r{\i}m ont r\'eussi \`a d\'emontrer que $p_{n+1} - p_n \ll_{\theta} 1$ pour une infinit\'e de couples $p_n,p_$. Sous la meme hypoth\`ese, nous d\'emontrerons que $p_{n+1} - p_n \ll_{q,\theta} 1$ et $p_n \equiv p_{n+1} \equiv a \bmod q$ pour une infinit\'e de couples $p_n,p_$, et nous prouverons \'egalement un r\'esultat quantitatif. Dans la deuxi\`eme partie, nous allons utiliser les techniques de Goldston-Pintz-Yldrm pour d\'emontrer qu'il existe une infinit\'e de couples de nombres premiers $p,p'$ tels que $(p-1)(p'-1)$ est une carr\'e parfait. Ce resultat est une version approximative d'une ancienne conjecture qui stipule qu'il existe une infinit\'e de nombres premiers $p$ tels que $p-1$ est une carr\'e parfait. En effet, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n \le Y$ tels que $n = \ell_1\cdots \ell_r$, avec $\ell_1,\ldots,\ell_r$ des premiers distincts, et tels que $(\ell_1-1)\cdots (\ell_r-1)$ est une puissance $r$-i\`eme, avec $r \ge 2$ quelconque. \'Egalement, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n = \ell_1\cdots \ell_r \le Y$ tels que $(\ell_1+1)\cdots (\ell_r+1)$ est une puissance $r$-i\`eme. Finalement, \'etant donn\'e $A$ un ensemble fini d'entiers non-nuls, nous d\'emontrerons une borne inf\'erieure sur le nombre d'entiers naturels $n \le Y$ tels que $\prod_ (p+a)$ est une puissance $r$-i\`eme, simultan\'ement pour chaque $a \in A$. / Let $p_1 = 2, p_2 = 3, p_3 = 5,\ldots$ be the sequence of all primes, and let $q \ge 3$ and $a$ be coprime integers. Recently, and very remarkably, Daniel Shiu proved an old conjecture of Sarvadaman Chowla, which asserts that there are infinitely many pairs of consecutive primes $p_n,p_{n+1}$ for which $p_n \equiv p_{n+1} \equiv a \bmod q$. Now fix a number $\epsilon > 0$, arbitrarily small. In their recent groundbreaking work, Daniel Goldston, J\`anos Pintz and Cem Y{\i}ld{\i}r{\i}m proved that there are arbitrarily large $x$ for which the short interval $(x, x + \epsilon\log x]$ contains at least two primes congruent to $a \bmod q$. Given a pair of primes $\equiv a \bmod q$ in such an interval, there might be a prime in-between them that is not $\equiv a \bmod q$. One can deduce that \emph{either} there are arbitrarily large $x$ for which $(x, x + \epsilon\log x]$ contains a prime pair $p_n \equiv p_{n+1} \equiv a \bmod q$, \emph{or} that there are arbitrarily large $x$ for which the $(x, x + \epsilon\log x]$ contains a triple of consecutive primes $p_n,p_{n+1},p_{n+2}$. Both statements are believed to be true, but one can only deduce that one of them is true, and one does not know which one, from the result of Goldston-Pintz-Y{\i}ld{\i}r{\i}m. In Part I of this thesis, we prove that the first of these alternatives is true, thus obtaining a new proof of Chowla's conjecture. The proof combines some of Shiu's ideas with those of Goldston-Pintz-Y{\i}ld{\i}r{\i}m, and so this result may be regarded as an application of their method. We then establish lower bounds for the number of prime pairs $p_n \equiv p_{n+1} \equiv a \bmod q$ with $p_{n+1} - p_n < \epsilon\log p_n$ and $p_{n+1} \le Y$. Assuming a certain unproven hypothesis concerning what is referred to as the `level of distribution', $\theta$, of the primes, Goldston-Pintz-Y{\i}ld{\i}r{\i}m were able to prove that $p_{n+1} - p_n \ll_{\theta} 1$ for infinitely many $n$. On the same hypothesis, we prove that there are infinitely many prime pairs $p_n \equiv p_{n+1} \equiv a \bmod q$ with $p_{n+1} - p_n \ll_{q,\theta} 1$. This conditional result is also proved in a quantitative form. In Part II we apply the techniques of Goldston-Pintz-Y{\i}ld{\i}r{\i}m to prove another result, namely that there are infinitely many pairs of distinct primes $p,p'$ such that $(p-1)(p'-1)$ is a perfect square. This is, in a sense, an `approximation' to the old conjecture that there are infinitely many primes $p$ such that $p-1$ is a perfect square. In fact we obtain a lower bound for the number of integers $n$, up to $Y$, such that $n = \ell_1\cdots \ell_r$, the $\ell_i$ distinct primes, and $(\ell_1 - 1)\cdots (\ell_r - 1)$ is a perfect $r$th power, for any given $r \ge 2$. We likewise obtain a lower bound for the number of such $n \le Y$ for which $(\ell_1 + 1)\cdots (\ell_r + 1)$ is a perfect $r$th power. Finally, given a finite set $A$ of nonzero integers, we obtain a lower bound for the number of $n \le Y$ for which $\prod_{p \mid n}(p+a)$ is a perfect $r$th power, simultaneously for every $a \in A$.

Le sujet de cette thèse est l'étude des progressions arithmétiques dans les nombres entiers. Plus précisément, nous nous intéressons à borner inférieurement v(N), la taille du plus grand sous-ensemble des nombres entiers de 1 à N qui ne contient pas de progressions arithmétiques de 3 termes. Nous allons donc construire de grands sous-ensembles de nombres entiers qui ne contiennent pas de telles progressions, ce qui nous donne une borne inférieure sur v(N). Nous allons d'abord étudier les preuves de toutes les bornes inférieures obtenues jusqu'à présent, pour ensuite donner une autre preuve de la meilleure borne. Nous allons considérer les points à coordonnés entières dans un anneau à d dimensions, et compter le nombre de progressions arithmétiques qu'il contient. Pour obtenir des bornes sur ces quantités, nous allons étudier les méthodes pour compter le nombre de points de réseau dans des sphères à plusieurs dimensions, ce qui est le sujet de la dernière section. / The subject of this thesis is the study of arithmetic progressions in the integers. Precisely, we are interested in the size v(N) of the largest subset of the integers from 1 to N that contains no 3 term arithmetic progressions. Therefore, we will construct a large subset of integers with no such progressions, thus giving us a lower bound on v(N). We will begin by looking at the proofs of all the significant lower bounds obtained on v(N), then we will show another proof of the best lower bound known today. For the proof, we will consider points on a large d-dimensional annulus, and count the number of integer points inside that annulus and the number of arithmetic progressions it contains. To obtain bounds on those quantities, it will be interesting to look at the theory behind counting lattice points in high dimensional spheres, which is the subject of the last section.

Submitted by Luciana Sebin (lusebin@ufscar.br) on 2016-09-12T17:46:24Z No. of bitstreams: 1 DissHM.pdf: 7239574 bytes, checksum: ae362c35b36e27d4096f64d4e04665e5 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-13T18:27:31Z (GMT) No. of bitstreams: 1 DissHM.pdf: 7239574 bytes, checksum: ae362c35b36e27d4096f64d4e04665e5 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-13T18:27:48Z (GMT) No. of bitstreams: 1 DissHM.pdf: 7239574 bytes, checksum: ae362c35b36e27d4096f64d4e04665e5 (MD5) / Made available in DSpace on 2016-09-13T18:28:00Z (GMT). No. of bitstreams: 1 DissHM.pdf: 7239574 bytes, checksum: ae362c35b36e27d4096f64d4e04665e5 (MD5) Previous issue date: 2015-09-19 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / The importance of the development of sequences that are arithmetic progressions in the high school was observed in learning situations which the students could investigate and identify patterns in numerical and geometric sequences, building the algebraic language to describe them. This work contributes for the student to build the idea of algebra as language to express irregularities, that is one of the proposed issues by the National Curricular Parameters (PCNs) for the math teaching in the fourth cycle of high school, which contrasts, nowadays, with the scarcity of activities involving observed arithmetic progressions in at least twelve years of experience as a math teacher in these cycles. The elaboration of a teaching product, in the way of activity sheets that, through the recognizing of numerical and geometric patterns, takes the student to the comprehension of the concept of arithmetic progression which could be tested through the application of these activity sheets in two classrooms of the ninth year of high school in a public municipally school. The obtained results of these applications were analyzed and compared to the previous analyzes in raised hypothesis during the elaboration of the activity sheets, using, as investigation methodology, the Didactic Engineering. The students did the activities in groups of two or three, were well motivated and participated as principal character during the application of all steps proposed in the paper, which guaranteed the good development of the activity. In according to the evaluation learning, the students reaching the proposed goals and noting that the produced teaching material works. It is believed that the elaborated material can be useful for other teachers who want to develop, in their classes, arithmetic progressions through the recognizing of patterns, adapting it to the reality of their classrooms. This work contributes hugely to the author, bringing a big professional evolution that starts with the issue choice, continued in the elaboration of the didactic sequence and the application of the activity sheets and finished with the reflection of what have been done and is registered here. / A importância do desenvolvimento de progressões aritméticas que são sequências no ensino fundamental foi observada em situações de aprendizagem que os alunos puderam investigar e identificar padrões em sequências numéricas e geométricas, construindo a linguagem algébrica para descrevê-las. Esse trabalho contribui para que o aluno construa a ideia de álgebra como uma linguagem para expressar regularidades, que é um dos conteúdos propostos pelos Parâmetros Curriculares Nacionais (PCNs) para o ensino de Matemática no quarto ciclo do ensino fundamental, o qual contrasta, atualmente, com a escassez de atividades envolvendo progressões aritméticas constatada em pelo menos doze anos de experiência como professor de matemática nesses ciclos. A elaboração de um produto de ensino, na forma de folhas de atividades que, através do reconhecimento de padrões numéricos ou geométricos levam o estudante à compreensão do conceito de progressão aritmética pôde ser conferida através da aplicação dessas folhas de atividades em duas salas de 9º ano do ensino fundamental de uma escola municipal. Os resultados obtidos dessas aplicações foram analisados e comparados com as análises prévias em hipóteses levantadas durante a elaboração das folhas de atividades, usando, como metodologia de investigação, a Engenharia Didática. Os alunos realizaram as atividades em duplas ou em trios, se sentiram bem motivados e participaram como protagonistas durante a aplicação de todas as etapas propostas nas folhas, o que garantiu o bom desenvolvimento das atividades. De acordo com a avaliação do aprendizado, os alunos atingiram os objetivos propostos e constatou-se que o material de ensino produzido e aplicado funciona. Acredita-se que o material elaborado possa ser útil a outros professores que desejarem desenvolver, em suas aulas, progressões aritméticas através do reconhecimento de padrões, podendo adaptá-lo à realidade de suas turmas. Este trabalho contribuiu enormemente ao autor, trazendo uma grande evolução profissional que se iniciou na escolha do tema, permeou pela elaboração da sequência didática e pela aplicação das folhas de atividades e terminou pela reflexão sobre o que foi feito e se encontra registrado aqui.

[pt] Nesta tese de Doutorado, nós estudamos cotas para as probabilidades de desvio de uma variável aleatória X que conta o número de arestas de um hipergrafo induzido por um subconjunto aleatório de m elementos do seu conjunto de vértices. Nós consideramos dois contextos: o primeiro corresponde a hipergrafos que possuem certo tipo de regularidade, ao passo que o segundo lida com hipergrafos que são, em algum sentido, longe de serem regulares. É possível aplicar estes resultados a estruturas discretas, como o conjunto de progressões aritméticas de tamanho k no grupo aditivo de inteiros módulo um primo e também no conjunto dos N primeiros inteiros positivos. Além disso, também deduzimos resultados para o caso em que o subconjunto aleatório é gerado incluindo cada vértice do hipergrafo independentemente com probabilidade p. / [en] In this Ph.D. thesis, we study bounds for the deviation probabilities of a random variable X that counts the number of edges of a hypergraph induced by a random m–element subset of its vertex set. We consider two contexts: the first corresponds to hypergraphs with some kind of regularity, whereas the second addresses hypergraphs that are in some sense far from being regular. It is possible to apply these results to discrete structures such as the set of k–term arithmetic progressions in the additive group of integers modulo a prime and in the set of the first N positive integers. Furthermore, we also deduce results for the case when the random subset is generated by including each vertex of the hypergraph independently with probability p.

Réalisé en cotutelle avec l'Université Paris-Diderot. / Nous présentons trois résultats en combinatoire additive, un domaine récent à la croisée de la combinatoire, l'analyse harmonique et la théorie analytique des nombres. Le thème unificateur de notre thèse est la détection de structures additives dans les ensembles arithmétiques à faible densité, avec un intérêt particulier pour les aspects quantitatifs. Notre première contribution est une estimation de densité améliorée pour le problème, initié entre autres par Bourgain, de trouver une longue progression arithmétique dans un ensemble somme triple. Notre deuxième résultat consiste en une généralisation des bornes de Sanders pour le théorème de Roth, du cas d'un ensemble dense dans les entiers à celui d'un ensemble à faible croissance additive dans un groupe abélien arbitraire. Finalement, nous étendons les meilleures bornes quantitatives connues pour le théorème de Roth dans les premiers, à tous les systèmes d'équations linéaires invariants par translation et de complexité un. / We present three results in additive combinatorics, a recent field at the interface of combinatorics, harmonic analysis and analytic number theory. The unifying theme in our thesis is the detection of additive structure in arithmetic sets of low density, with an emphasis on quantitative aspects. Our first contribution is an improved density estimate for the problem, initiated by Bourgain and others, of finding a long arithmetic progression in a triple sumset. Our second result is a generalization of Sanders' bounds for Roth's theorem from the dense setting, to the setting of small doubling in an arbitrary abelian group. Finally, we extend the best known quantitative results for Roth's theorem in the primes, to all translation-invariant systems of equations of complexity one.

Le sujet principal de cette thèse est la distribution des nombres premiers dans les progressions arithmétiques, c'est-à-dire des nombres premiers de la forme $qn+a$, avec $a$ et $q$ des entiers fixés et $n=1,2,3,\dots$ La thèse porte aussi sur la comparaison de différentes suites arithmétiques par rapport à leur comportement dans les progressions arithmétiques. Elle est divisée en quatre chapitres et contient trois articles. Le premier chapitre est une invitation à la théorie analytique des nombres, suivie d'une revue des outils qui seront utilisés plus tard. Cette introduction comporte aussi certains résultats de recherche, que nous avons cru bon d'inclure au fil du texte. Le deuxième chapitre contient l'article \emph{Inequities in the Shanks-Rényi prime number race: an asymptotic formula for the densities}, qui est le fruit de recherche conjointe avec le professeur Greg Martin. Le but de cet article est d'étudier un phénomène appelé le <<Biais de Chebyshev>>, qui s'observe dans les <<courses de nombres premiers>>. Chebyshev a observé qu'il semble y avoir plus de premiers de la forme $4n+3$ que de la forme $4n+1$. De manière plus générale, Rubinstein et Sarnak ont montré l'existence d'une quantité $\delta(q;a,b)$, qui désigne la probabilité d'avoir plus de premiers de la forme $qn+a$ que de la forme $qn+b$. Dans cet article nous prouvons une formule asymptotique pour $\delta(q;a,b)$ qui peut être d'un ordre de précision arbitraire (en terme de puissance négative de $q$). Nous présentons aussi des résultats numériques qui supportent nos formules. Le troisième chapitre contient l'article \emph{Residue classes containing an unexpected number of primes}. Le but est de fixer un entier $a\neq 0$ et ensuite d'étudier la répartition des premiers de la forme $qn+a$, en moyenne sur $q$. Nous montrons que l'entier $a$ fixé au départ a une grande influence sur cette répartition, et qu'il existe en fait certaines progressions arithmétiques contenant moins de premiers que d'autres. Ce phénomène est plutôt surprenant, compte tenu du théorème des premiers dans les progressions arithmétiques qui stipule que les premiers sont équidistribués dans les classes d'équivalence $\bmod q$. Le quatrième chapitre contient l'article \emph{The influence of the first term of an arithmetic progression}. Dans cet article on s'intéresse à des irrégularités similaires à celles observées au troisième chapitre, mais pour des suites arithmétiques plus générales. En effet, nous étudions des suites telles que les entiers s'exprimant comme la somme de deux carrés, les valeurs d'une forme quadratique binaire, les $k$-tuplets de premiers et les entiers sans petit facteur premier. Nous démontrons que dans chacun de ces exemples, ainsi que dans une grande classe de suites arithmétiques, il existe des irrégularités dans les progressions arithmétiques $a\bmod q$, avec $a$ fixé et en moyenne sur $q$. / The main subject of this thesis is the distribution of primes in arithmetic progressions, that is of primes of the form $qn+a$, with $a$ and $q$ fixed, and $n=1,2,3,\dots$ The thesis also compares different arithmetic sequences, according to their behaviour over arithmetic progressions. It is divided in four chapters and contains three articles. The first chapter is an invitation to the subject of analytic number theory, which is followed by a review of the various number-theoretic tools to be used in the following chapters. This introduction also contains some research results, which we found adequate to include. The second chapter consists of the article \emph{Inequities in the Shanks-Rényi prime number race: an asymptotic formula for the densities}, which is joint work with Professor Greg Martin. The goal of this article is to study <<Chebyshev's Bias>>, a phenomenon appearing in <<prime number races>>. Chebyshev was the first to observe that there tends to be more primes of the form $4n+3$ than of the form $4n+1$. More generally, Rubinstein and Sarnak showed the existence of the quantity $\delta(q;a,b)$, which stands for the probability of having more primes of the form $qn+a$ than of the form $qn+b$. In this paper, we establish an asymptotic series for $\delta(q;a,b)$ which is precise to an arbitrary order of precision (in terms of negative powers of $q$). %(it can be instantiated with an error term smaller than any negative power of $q$). We also provide many numerical results supporting our formulas. The third chapter consists of the article \emph{Residue classes containing an unexpected number of primes}. We fix an integer $a \neq 0$ and study the distribution of the primes of the form $qn+a$, on average over $q$. We show that the choice of $a$ has a significant influence on this distribution, and that some arithmetic progressions contain, on average over q, fewer primes than typical arithmetic progressions. This phenomenon is quite surprising since in light of the prime number theorem for arithmetic progressions, the primes are equidistributed in the residue classes $\bmod q$. The fourth chapter consists of the article \emph{The influence of the first term of an arithmetic progression}. In this article we are interested in studying more general arithmetic sequences and finding irregularities similar to those observed in chapter three. Examples of such sequences are the integers which can be written as the sum of two squares, values of binary quadratic forms, prime $k$-tuples and integers free of small prime factors. We show that a broad class of arithmetic sequences exhibits such irregularities over the arithmetic progressions $a\bmod q$, with $a$ fixed and on average over $q$.