Lower bounds for sums of Barban-Davenport-Halberstam type. (Q4032164)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 158492
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Lower bounds for sums of Barban-Davenport-Halberstam type. |
scientific article; zbMATH DE number 158492 |
Statements
Lower bounds for sums of Barban-Davenport-Halberstam type. (English)
0 references
12 May 1993
0 references
Barban-Davenport-Halberstam theorem
0 references
circle method
0 references
large sieve inequality
0 references
lower bound
0 references
0.81043416
0 references
0.74031085
0 references
0.73310393
0 references
0 references
0.70900494
0 references
0.7064418
0 references
The Barban-Davenport-Halberstam theorem gives an upper bound for the sum \[ S(Q,x)=\sum_{q\leq Q}\sum_{{0<a\leq q\atop (a,q)=1}}\left(\psi(x;q,a)-{x\over \varphi(q)}\right)^ 2. \] Here \(\varphi(q)\) is Euler's function, \(\psi(x;q,a)\) is the Chebyshev function which sums the von Mangoldt function \(\Lambda(n)\) over the arithmetic progression \(n\equiv a\pmod q\), \(n\leq x\). The significant point of the theorem is that the bound holds when \(Q\) can be near \(x\), say \(x/\log^ A x\leq Q\leq x\) for some fixed \(A\). In 1970, \textit{H. L. Montgomery} [Mich. Math. J. 17, 33-39 (1970; Zbl 0209.348)] expanded the inner sum and applied a theorem of Lavrik to obtain an asymptotic formula for \(S(Q,x)\). This was followed by a series of papers on the topic by \textit{C. Hooley} [J. Lond. Math. Soc., II. Ser. 10, 249-256 (1975; Zbl 0304.10029)] who obtained a more precise formula and also generalized the result by replacing \(\psi(x;q,a)\) with sums formed from sequences satisfying certain conditions. More recently, \textit{E. Bombieri}, \textit{J. B. Friedlander} and \textit{H. Iwaniec} [Acta Math. 156, 203-251 (1986; Zbl 0588.10042)] applied the large sieve inequality to obtain an upper estimate for the sum \[ S\bigl((\beta_ n),Q,x\bigr)=\sum_{q\leq Q}\sum_{{0<a\leq q\atop (a,q)=1}}\Bigl|\sum_{{n\leq x\atop n\equiv a\pmod q}}\beta_ n-{1\over\varphi(q)}\sum_{{n\leq x\atop (n,q)=1}}\beta_ n\Bigr|^ 2, \] where the complex sequence \((\beta_ n)\) satisfies a certain `Siegel-Walfisz type' condition. However, an asymptotic formula for this generalized sum is now difficult to obtain, because extra properties will be required from \((\beta_ n)\) in order to apply the `dispersion method' type of argument used by Montgomery and Hooley. The double summation associated with the sum \(S\bigl((\beta_ n),Q,x\bigr)\) shows that the problem has some connection with Farey sequences. This observation led the author to consider the circle method, which he now used together with the large sieve inequality to give a lower bound for the sum. Roughly speaking, the results show that the sum \(S\bigl((\beta_ n),Q,x\bigr)\) has the order \(Q\sum_{n\in x} |\beta_ n|^ 2\), but the precise statements of the theorems are rather complicated to be stated in full.
0 references