An almost all result on q₁ q₂ c q

. Then we apply it to a question of Fujii and Kitaoka as well as generalize it to more variables. At the end, we will present a new way to attack the above congruence equation question through higher moments.

Review text

This paper is refer to following famous congruence equation question: Question 1: Given any \(\varepsilon>0\). Is it true that for any modulus \(q\geq 1\) and any integer \(c\) with \((c,q)=1\), the congruence equation \(q_1q_2\equiv c \pmod q\) is solvable for some \(1\leq q_1,\,q_2\ll_\varepsilon q^{1/2+\varepsilon}\)? The author proved that for any modulus \(q\geq 1\) and any integers \(1\leq L_1,\dots ,L_t\leq q\) and \((c,q)=1\) \[ \begin{multlined} \sum_{a_1=1}^{q}\dots \sum_{a_t=1}^{q} \bigg| \sum_{q_1\in(a_1,a_1+L_1]\atop{(q_1,\,q)=1}} \dots \sum_{q_t\in(a_t,a_t+L_t]\atop{(q_t,\,q)=1\atop{q_1q_2\dots q_t\equiv c \pmod q}}} 1 -\, \frac{L_1\dots L_t}{q^t}\,\varphi(q)^{t-1} \bigg| \\ \leq C_q^2(t+1)t^{2\omega(q)}d(q)^{t-1}q^{t-1}L_1\dots L_t\end{multlined}\tag{1} \] where \(C_q=1\) if \(q\) is odd and \(C_q=2^{t+1}/2\) if \(q\) is even. As corollary the author obtained that for any \(\varepsilon>0\) there exists a constant \(C_{\varepsilon,\,t}>0\) such that \[ q_1\dots q_t\equiv c \pmod q \] is solvable in \(q_i\in (a_i,\, a_i+C_{\varepsilon,\,t}q^{1/t+\varepsilon}),\, i=1,\dots,t\) for all but \(o(q^t)\) \(t\)-tuples \(a_1,\dots,a_t\). The exponent \(1/t +\varepsilon\) is best possible. In the case \(t=2\) the author obtains an answer to Question 1 in an almost all sense. Also as corollary of the main result for prime \(q\) the author recovers \textit{M. Z. Garaev}'s \(3/4\) result [Acta Arith. 124, No. 1, 27--39 (2006; Zbl 1158.11002)]. Moreover, it opens up a new way to attack Question 1 through investigating higher moment analogues of (1). The author applies the main result to a question of \textit{A. Fujii} and \textit{Y. Kitaoka} [Nagoya Math. J. 147, 137--146 (1997; Zbl 0917.11051)]. More precisely, he proves the following result in an almost all sense: If \(C_{(x,y)}(r) \) denotes the compact disc with center \((x,y)\) and radius \(r\), \(\varepsilon>0\) is arbitrary then there exists \(C_{\varepsilon}>0\) such that for all \(q\) \[ \bigcup_{x,\,y=1\atop{xy\bmod q}}^{q-1}C_{(x,y)}(r) \] cover the square \([0,\,q]\times[0,\,q]\) except for an area of \(o(q^2)\).

Recommendations

Cites work

Cited in

(3)

This page was built for publication: An almost all result on \(q_1 q_2 \equiv c\pmod q\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q621771)