Linear inhomogeneous congruences in continued fractions on finite alphabets (Q2090542)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Linear inhomogeneous congruences in continued fractions on finite alphabets
scientific article

    Statements

    Linear inhomogeneous congruences in continued fractions on finite alphabets (English)
    0 references
    0 references
    0 references
    25 October 2022
    0 references
    The authors study words in a fixed finite alphabet \(\mathbf{A}\subseteq\mathbb{N}\) and let to words in \(\mathbf{A}\) uniquely correspond finite continued fractions by \[ [d_1,d_2,\ldots, d_k]=\frac{1}{d_1+\frac{1}{d_2+..+\frac{1}{d_k}}}. \] Let \(\mathcal{R}_{\mathbf{A}}\) denote the set off irreducible rational fractions \(y/Y\) expressed in terms of finite continued fractions on \(\mathbf{A}\) \[ \mathcal{R}_{\mathbf{A}}=\Big\{\frac{y}{Y}\Big|\begin{array}{l}y,Y\in\mathbb{N},y/Y=[d_1,d_2,\ldots,d_k], \hbox{ gcd}(y,Y)=1,y\leq Y,\cr d_j\in\mathbf{A}\hbox{ for }j=1,2,\ldots,k\end{array}\Big\}. \] Let natural numbers \(a,b,q,t,P^{(1)}\) and \(P^{(2)}\), as well as a real number \(p>0\) be given (\(a,b,q\) are co-prime), then \(T^{\mathbf{A}}_{p,b,t}\) is the number of solutions of the congruence \[ ay-bY\equiv t \pmod{q}\tag{1} \] in integer variables \(Y\) and \(y\) from the intervals \([p,P^{(1)}]\) and \([1,P^{(2)}]\) respectively, provided that \(y/Y\) belongs to \(\mathcal{R}_{\mathbf{A}}\). The main results of the paper concern an upper bound for \(T^{\mathbf{A}}_{p,b,t}\) for certain values of \(p\), extending results (for \(t=0,p=b=1,\mathbf{A}=\mathbb{N}\) and slightly different condiitons on \(a,q\) and \(y,Y\)) due to \textit{N. M. Korobov} [Trigonometric sums and their applications (Russian). Moskva: Nauka (1989; Zbl 0665.10026)]. They also studies the case where \(y,Y\) satisfy the additional constraint \[ \left|\frac{y}{Y}-\psi\right|\leq\varepsilon,\eqno{(2)} \] wherer \(\psi\) is also a fraction from \(\mathcal{R}_{\mathbf{A}}\) and \(\varepsilon\in (0,1]\) is an arbitrary small number. The number of solutions of \((1)\) in integer variables \(Y\in [p,P]\) and \(y\in [1,P]\) for \(y/Y\in \mathcal{R}_{\mathbf{A}}\) is denoted by \(T^{\mathbf{A},\varepsilon}_{p,b,t}\). The main results are now: Theorem 1.1. Let the inequalities \(P\geq 2\sqrt{q/\varepsilon}\) and \((2)\) be valid. Then the following estimate holds: \[ T^{\mathbf{A},\varepsilon}_{2\sqrt{q/\varepsilon},b,0}\ll \left(\frac{P^2\varepsilon}{q}\right)^{\Delta_{\mathbf{A}}}. \] Theorem 1.2. Let the inequalities \(P\geq 2\sqrt{q/\varepsilon}\) and \((2)\) be valid. Then the following estimate holds: \[ T^{\mathbf{A},\varepsilon}_{2\sqrt{q/\varepsilon},b,t}\ll \frac{(P^2\varepsilon)^{\Delta_{\mathbf{A}}}}{q^{2\Delta_{\mathbf{A}-1}}}. \] Theorem 3.1. If the inequality \(P\geq 2\sqrt{q/\varepsilon}\) holds, then, for \(M_1=\max\left\{1,\frac{C'}{\sqrt{\varepsilon}}\right\}\) and a suitable choice of the constant \(C'\), and for \(M_3=\max\{ \sqrt{q}/2,1\}\), the following estimate holds: \[ |S_{q,\varepsilon}|\ll \frac{|G_{\mathbf{A}}^{P,2A}|}{(q/\varepsilon)^{\Delta_{\mathbf{A}}}}. \] There is another result proven in the paper, but it would take too much space to reproduce all intermediate definitions.; nevertheless, it is feproduced here: Theorem 4.1. If the inequality \(P\geq 2\sqrt{q/\varepsilon}\) holds, then, for \(M_1=\max\left\{1,\frac{C'}{\sqrt{\varepsilon}}\right\}\) and a suitable choice of the constant \(C'\), and for \(M_3=\max\{ \sqrt{q}/2,1\}\), the following estimate holds: \[ \sum_{\left(\begin{array}{l}y\cr Y\end{array}\right)\in G_{\mathbf{A}}^{P,2A}}\ \mathbf{1}_{ay-bY\equiv t \pmod{q}}\mathbf{1}_{\{|y/Y-\psi|\leq\varepsilon\}}\ll \frac{|G_{\mathbf{A}}^{P,2}|\varepsilon^{\Delta_{\mathbf{A}}}}{q^{2\Delta_{\mathbf{A}}-1}}\ll \frac{P^{2\Delta_{mathbf{A}}}\varepsilon^{\Delta_{\mathbf{A}}}}{q^{2\Delta_{\mathbf{A}}-1}} \] The layout of the paper is as follows: \S1. Introduction (\(1\frac{1}{2}\) page). Statement of the problem, historical background, main results. \S2. Continued fractions and matrices (\(4\frac{1}{2}\) pages). \S3. Sums of Korobov \(\delta\)-symbols (\(2\) pages). \S4. Replacing an inhomogeneous congruence by two congruences expressed in terms of a matrix (\(1\frac{1}{2}\) page). \S5. Proof of the main results (\(\frac{1}{2}\) page). References (\(20\) items). The interested reader is invited to peruse the paper meticulously: it is worth it!
    0 references
    linear inhomogeneous congruence
    0 references
    linear homogeneous congruence
    0 references
    continued fraction
    0 references
    finite alphabet
    0 references

    Identifiers