On character sums with determinants (Q6146120)

From MaRDI portal
scientific article; zbMATH DE number 7786154
Language Label Description Also known as
English
On character sums with determinants
scientific article; zbMATH DE number 7786154

    Statements

    On character sums with determinants (English)
    0 references
    0 references
    0 references
    10 January 2024
    0 references
    For a prime \(p\) and an integer \(N\geq 2\), let \(\mathbb F_p\) be the set of integers \(\{0, \,1, \ldots, p-1\}\), \(\mathbb F_p^*=\mathbb F_p \backslash\{0\}\), and consider the following notations, \begin{itemize} \item \(\mathcal M_2 (p) \) is the set of \(2\times 2\)-matrices, with coefficients in \(\mathbb F_p\), \item \( \mathcal M_2^* (p)\, (= \mathrm{GL}(2, \mathbb F_p))\) is the subset of \(\mathcal M_2 (p)\) containing all the non-singular matrices, \item \(\mathcal M_2^{\ne \square}(p)\) is the subset of \(\mathcal M_2^* (p)\) defined by \[ \mathcal M_2^{\ne \square}(p) := \left\{A \in \mathcal M_2^* (p) :~A\ne B^2 \text{ for all } B \in \mathcal M_2^* (p) \right\}, \] the elements of \(\mathcal M_2^{\ne \square}(p)\) are called textit{matrices without square root.} \item \({\mathbf M} (N,p)\) is the subset of \(\mathcal M_2 (p)\) defined by \[ {\mathbf M }(N,p) : =\left\{ \begin{pmatrix} a \bmod p & b \bmod p\\ c \bmod p & d \bmod p \end{pmatrix}:~0\leq a,\, b,\, c,\, d \leq N \right\}. \] \end{itemize} Also, let \(\mathfrak z_p\) be the least integer \(n \geq 2\) which is not a quadratic residue modulo \(p\). In the paper under review, the authors first prove that if \(0<\kappa_1<1\) and \(C\) such that, for all \(p\geq 3\), one has the inequality \(2\leq \mathfrak z_p \leq C p^{\kappa_1}\), then for all \(p\geq 3\) the intersection set \[ {\mathbf M }\left(\sqrt C p^{\kappa_1/2} +1 ,p\right)\, \bigcap\, \mathcal M_2^{\ne \square}(p) \] is not empty. Concerning with the number of solutions to the determinant equation \(\Delta = ad-bc\) they prove that for every \(\varepsilon >0\) there exists some \(C> 0\) and \(p_0\), which depend only on \(\varepsilon\), such that for every \(p\geq p_0\), for every \(N\geq p^{1/8\sqrt{\mathrm{e}}+\varepsilon}\) one has \[ \sharp \, \left({\mathbf{M}} (N,p)\, \bigcap \, \mathcal M_2^{\ne \square} (p)\right) \geq C N^4. \] Moreover, they consider several sums concerning the determinant equation \(\Delta = ad-bc\), and provide approximation for these sums. The first sum that the authors study appealing to the properties of the determinant equation is \[ U_\chi (\boldsymbol \alpha, \boldsymbol \beta, N ) := \sum_{1\leq a\leq N}\alpha_a \sum_{1\leq b \leq N} \beta_b \sum_{1\leq c \leq N} \sum_{1\leq d \leq N} \chi (ad-bc), \] where \(\chi\) is a non-principal multiplicative character modulo \(p\), and \(\boldsymbol \alpha =(\alpha_a)\) and \(\boldsymbol \beta= (\beta_b)\) are quite general sequences. Accordingly, they prove that for every \(\varepsilon >0\) there exist \(\delta_0 >0\) such that the following inequality holds \[ U_\chi (\boldsymbol \alpha, \boldsymbol \beta, N ) \ll N^{4-\delta_0}, \] uniformly for bounded weights \(\boldsymbol \alpha\) and \(\boldsymbol \beta\), for all \(a\) and \(b \geq 1\) and uniformly for \(p^{1/2} > N\geq p^{1/8 + \varepsilon}\). The second sum that they study is \[ T_\chi( A,B,C, \mathcal{D};\boldsymbol \alpha):=\sum_{1 \le a \le A} \sum_{1\le b \le B} \sum_{1\le c\le C} \left| \,\sum_{d \in \mathcal{D}} \alpha_d \chi(ab-cd)\, \right|, \] where \(A\), \(B\), \(C\) and \(D\) are integers \( \geq 1\), \(\mathcal{D}\subseteq \mathbb F_p^*\) has cardinality \(D\), and where \(\chi\) is a non-principal character modulo \(p\). For this sum the authors prove that for every integer \(\nu \geq 1\), there exists a constant \(C(\nu)\) such that, for all prime \(p\), for all non-principal character \(\chi\) modulo \(p\), for all positive integers \(A\), \(B\), \(C\) and \(D\) satisfying \(ABC\text{ and } D < p\), for all set \(\mathcal{D} \subseteq \mathbb F_p^*\) of cardinality \(D\), and for all arbitrary complex weights \(\boldsymbol \alpha\) satisfying \(|\alpha_d| \leq 1\) if \(d\in \mathcal D\), one has \[ T_\chi(A,B,C, \mathcal{D};{\boldsymbol\alpha})\leq C(\nu) ABCD \left( \left(\frac{p}{ABC}\right)^{1/(2\nu)} D^{-1/2} + \left(\frac{p^{1/2}}{ABC}\right)^{1/(2\nu)}\right)(\log p)^{4/\nu}. \] The last sum that authors consider is the following one, which is more general than \(U_\chi (\boldsymbol \alpha, \boldsymbol \beta, N )\), \[ T_\chi( N):=\sum_{1 \le a \le N} \sum_{1\le b \le N} \sum_{1\le c\le N}\left| \sum_{1\le d\le N} \chi(ab-cd)\right|, \] where \(N\geq 1\). For, they prove that for any \(\varepsilon>0\), one has the inequality \[ T_\chi( N)\ll N^4\frac{\log \log p}{\log p}, \] for every prime \(p\), for every non-principal character \(\chi\) modulo \(p\), and for every \(p >N \ge p^{1/8 + \varepsilon}\).
    0 references
    0 references
    character sum
    0 references
    determinant
    0 references
    Burgess bound
    0 references

    Identifiers