En admettant la conjecture de Dickson, nous démontrons que, pour chaque couple d’entiers $q\>0$ et $k\>0$, il existe une partie infinie $L_{q,k}\subset \mathbb{N}$ telle que, pour chacun des entiers $n\in L_{q,k}$ et tout entier $s$ tel que $0\<\left|s\right|\le q$, on ait $n+s=\left|s\right|t_1...t_k$ où $t_1\<...\<t_k$ sont des nombres premiers. De même, pour chaque couple d’entiers $q\>0$ et $k\>0$, il existe une partie infinie $M_{q,k}\subset \mathbb{N}$ telle que, pour chacun des entiers $n\in M_{q,k}$ et tout entier $s$ (nul ou non ) de l’intervalle $\left[-q,q\right]$, on ait $n+s=l{t}_1...t_k$ où $t_1\<...\<t_k$ sont des nombres premiers et l’entier $l$ appartient à l’intervalle $\left[1,2q+1\right]$. La lecture non standard...

Le problème de la primalité est l’un des problèmes les plus simples et les plus anciens de la théorie des nombres. À la fin des années 1970, Adleman, Pomerance et Rumely ont donné le premier algorithme de primalité déterministe, dont le temps de calcul était presque polynomial. Il a fallu 20 années supplémentaires pour qu’Agrawal, Kayal et Saxena donnent un algorithme déterministe de temps de calcul polynomial. L’exposé présentera ces travaux, et il fera également le point sur les différents autres...

Let $F\left(X\right)={\sum }_{n\ge 0}{(-1)}^{\epsilon ₙ}{X}^{-\lambda ₙ}$ be a real lacunary formal power series, where εₙ = 0,1 and ${\lambda }_{n+1}/\lambda ₙ>2$. It is known that the denominators Qₙ(X) of the convergents of its continued fraction expansion are polynomials with coefficients 0, ±1, and that the number of nonzero terms in Qₙ(X) is the nth term of the Stern-Brocot sequence. We show that replacing the index n by any 2-adic integer ω makes sense. We prove that $Q_{\omega }\left(X\right)$ is a polynomial if and only if ω ∈ ℤ. In all the other cases $Q_{\omega }\left(X\right)$ is an infinite formal power series; we discuss its algebraic...

Let pₙ/qₙ = [a₀;a₁,...,aₙ] be the n-th convergent of the continued fraction expansion of [a₀;a₁,a₂,...]. Leaping convergents are those of every r-th convergent $p_{rn+i}/q_{rn+i}$ (n = 0,1,2,...) for fixed integers r and i with r ≥ 2 and i = 0,1,...,r-1. The leaping convergents for the e-type Hurwitz continued fractions have been studied. In special, recurrence relations and explicit forms of such leaping convergents have been treated. In this paper, we consider recurrence relations and explicit forms of the leaping...

Denote the n-th convergent of the continued fraction by pₙ/qₙ = [a₀;a₁,...,aₙ]. We give some explicit forms of leaping convergents of Tasoev continued fractions. For instance, [0;ua,ua²,ua³,...] is one of the typical types of Tasoev continued fractions. Leaping convergents are of the form $p_{rn+i}/q_{rn+i}$ (n=0,1,2,...) for fixed integers r ≥ 2 and 0 ≤ i ≤ r-1.

Let p > 3 be a prime, and let Rₚ be the set of rational numbers whose denominator is not divisible by p. Let Pₙ(x) be the Legendre polynomials. In this paper we mainly show that for m,n,t ∈ Rₚ with m≢ 0 (mod p), $P_{[p/6]}\left(t\right)\equiv -(3/p){\sum }_{x=0}^{p-1}((x³-3x+2t)/p)\left(modp\right)$ and $\left({\sum }_{x=0}^{p-1}((x³+mx+n)/p)\right)²\equiv ((-3m)/p){\sum }_{k=0}^{[p/6]}\left(\genfrac{}{}{0pt}{}{2k}{k}\right)\left(\genfrac{}{}{0pt}{}{3k}{k}\right)\left(\genfrac{}{}{0pt}{}{6k}{3k}\right){((4m³+27n²)/\left(12³·4m³\right))}^k\left(modp\right)$, where (a/p) is the Legendre symbol and [x] is the greatest integer function. As an application we solve some conjectures of Z. W. Sun and the author concerning ${\sum }_{k=0}^{p-1}\left(\genfrac{}{}{0pt}{}{2k}{k}\right)\left(\genfrac{}{}{0pt}{}{3k}{k}\right)\left(\genfrac{}{}{0pt}{}{6k}{3k}\right)/m^k\left(modp²\right)$, where m is an integer not divisible by p.

Let d ≥ 2 be a square-free integer and for all n ≥ 0, let $l\left({\left(\surd d\right)}^{2n+1}\right)$ be the length of the continued fraction expansion of ${\left(\surd d\right)}^{2n+1}$. If ℚ(√d) is a principal quadratic field, then under a condition on the fundamental unit of ℤ[√d] we prove that there exist constants C₁ and C₂ such that $C₁{\left(\surd d\right)}^{2n+1}\ge l\left({\left(\surd d\right)}^{2n+1}\right)\ge C₂{\left(\surd d\right)}^{2n+1}$ for all large n. This is a generalization of a theorem of S. Chowla and S. S. Pillai [2] and an improvement in a particular case of a theorem of [6].

For $m\in$, $(m,6)=1$, it is proved the relations between the sums $$W(m,s)=\sum _{i=1,(i,m)=1}^{m-1}{i}^{-s},s\in ,$$ and Bernoulli numbers. The result supplements the known theorems of C. Leudesdorf, N. Rama Rao and others. As the application it is obtained some connections between the sums $W(m,s)$ and Agoh’s functions, Wilson quotients, the indices irregularity of Bernoulli numbers.