A generalization on the difference between an integer and its inverse modulo \(q\) (Q925945): Difference between revisions

Let \(q>2\) and \(c\) be two integers with \((q,c)=1\). For integers \(k\geq1\) and \(n\geq 1\), define \[ M(n,k,c,q)= \mathop{\mathop{{\sum}'}_{a_1=1}^q\cdots \mathop{{\sum}'}_{a_n=1}^q}_{a_1\cdots a_nb\equiv c\pmod q}(a_1\cdots a_n-b)^{2k} \] and \[ \begin{multlined} E(n,1,c,q)=M(n,1,c,q)+\frac{\varphi^n(q)q^{n+1}}{2^n} -\left(\frac{\varphi(q)q^2}{3}+\frac{q}{6}\prod_{p|q}(1-p)\right)^n \\ -\left(\frac{\varphi^n(q)q^2}{3}+\frac{\varphi^{n-1}(q)}{6}\prod(1-p)\right).\end{multlined} \] In this paper, by using the properties of the Gauss sums, primitive characters and the mean value of Dirichlet \(L\)-functions, the authors study the hybrid mean value of \(E(n,1,c,q)\) and the hyper-Kloosterman sums \(K(c,n+1,q)\), the asymptotic property of the mean square value \(\sum_{c=1}^pE^2(n,1,c,p)\), and give two mean value formulas.