Namespaces
Variants
Actions

Difference between revisions of "Page theorem"

From Encyclopedia of Mathematics
Jump to: navigation, search
(TeX part done)
(TeX done)
 
Line 1: Line 1:
{{TEX|part}}
+
{{TEX|done}}{{MSC|11M06|11N13}}
  
 
Page's theorem on the zeros of Dirichlet $L$-functions.
 
Page's theorem on the zeros of Dirichlet $L$-functions.
Line 10: Line 10:
  
 
c) for complex $\chi$  modulo $d$,
 
c) for complex $\chi$  modulo $d$,
$$
+
\begin{equation}\label{1}
 
L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_3}{\log d}\,,\ |t| \le 5\,;
 
L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_3}{\log d}\,,\ |t| \le 5\,;
$$
+
\end{equation}
  
 
d) for real primitive $\chi$  modulo $d$,
 
d) for real primitive $\chi$  modulo $d$,
$$
+
\begin{equation}\label{2}
 
L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_4}{\sqrt{d}\log^2 d}\,;
 
L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_4}{\sqrt{d}\log^2 d}\,;
$$
+
\end{equation}
  
 
e) for $2 \le d \le D$ there exists at most one $d=d_0$, $d_0 \ge (\log^2 D)/(\log\log^8 D)$ and at most one real primitive $\psi$ modulo $d$ for which $L(s,\psi$ can have a real zero $\beta_1 > 1- c_6/\log D$, where $\beta_1$ is a simple zero; and for all $\beta$ such that $L(\beta,\psi) =0$, $\beta > 1 - c_6/\log D$ with a real $\psi$ modulo $d$, one has $d \equiv 0 \pmod {d_0}$.
 
e) for $2 \le d \le D$ there exists at most one $d=d_0$, $d_0 \ge (\log^2 D)/(\log\log^8 D)$ and at most one real primitive $\psi$ modulo $d$ for which $L(s,\psi$ can have a real zero $\beta_1 > 1- c_6/\log D$, where $\beta_1$ is a simple zero; and for all $\beta$ such that $L(\beta,\psi) =0$, $\beta > 1 - c_6/\log D$ with a real $\psi$ modulo $d$, one has $d \equiv 0 \pmod {d_0}$.
  
Page's theorem on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107037.png" />, the number of prime numbers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107038.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107039.png" /> (<img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107040.png" />) for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107041.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107042.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107043.png" /> are relatively prime numbers. With the symbols and conditions of Section 1, on account of a)–c) and e) one has
+
Page's theorem on $\pi(x;d,l)$, the number of prime numbers $p \le x$, $p \equiv l \pmod d$ for $0 < l \le d$, where $l$ and $d$ are relatively prime numbers. With the symbols and conditions of Section 1, on account of a)–c) and e) one has
 
+
$$
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107044.png" /></td> </tr></table>
+
\pi(x;d,l) = \frac{\mathrm{li}(x)}{\phi(d)} - E \frac{\chi(l)}{\phi(d)}\sum_{n \le x} \frac{n^{\beta_1 - 1}}{\log n} + O\left({x \exp\left({-c_7 \sqrt{\log x}}\right)}\right) \ ,
 
+
$$
where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107045.png" /> or <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107046.png" /> in accordance with whether <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107047.png" /> exists or not for a given <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107048.png" />; because of , for any <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107049.png" /> one has for a given <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107050.png" />,
 
 
 
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107051.png" /></td> <td valign="top" style="width:5%;text-align:right;">(*)</td></tr></table>
 
  
This result is the only one (1983) that is effective in the sense that if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107052.png" /> is given, then one can state numerical values of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107053.png" /> and the constant appearing in the symbol <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107054.png" />. Replacement of the bound in
+
where $E=1$ or $0$ in accordance with whether $\beta_1$ exists or not for a given $d$; because of (2), for any $d \le (\log x)^{1-\delta}$ one has for a given $\delta>0$,
 +
\begin{equation}\label{3}
 +
\pi(x;d,l) = \frac{\mathrm{li}(x)}{\phi(d)}  + O\left({x \exp(-c_8 \sqrt{\log x})}\right) \ .
 +
\end{equation}
  
by the Siegel bound: <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107055.png" /> for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107056.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107057.png" />, extends the range of (*) to essentially larger <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107058.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107059.png" /> for any fixed <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107060.png" />, but the effectiveness of the bound in (*) is lost, since for a given <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107061.png" /> it is impossible to estimate <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107062.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/p/p071/p071070/p07107063.png" />.
+
This result is the only one (1983) that is effective in the sense that if $\delta$ is given, then one can state numerical values of $c_8$ and the constant appearing in the symbol $O$. Replacement of the bound in (2)
 +
by the Siegel bound: $L(\sigma,\chi) \ne 0$ for $\sigma > 1-c(\epsilon)d^{-\epsilon}$, $\epsilon > 0$, extends the range of (*) to essentially larger $d$, $d \le (\log x)^A$ for any fixed $A$, but the effectiveness of the bound in (3) is lost, since for a given $\epsilon > 0$ it is impossible to estimate $c_8(\epsilon)$ and $O_\epsilon$.
  
 
A. Page established these theorems in [[#References|[1]]].
 
A. Page established these theorems in [[#References|[1]]].
  
 
====References====
 
====References====
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  A. Page,  "On the number of primes in an arithmetic progression"  ''Proc. London Math. Soc. Ser. 2'' , '''39''' :  2  (1935)  pp. 116–141</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  A.A. Karatsuba,  "Fundamentals of analytic number theory" , Moscow  (1975)  (In Russian)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top">  K. Prachar,  "Primzahlverteilung" , Springer  (1957)</TD></TR></table>
+
<table>
 +
<TR><TD valign="top">[1]</TD> <TD valign="top">  A. Page,  "On the number of primes in an arithmetic progression"  ''Proc. London Math. Soc. Ser. 2'' , '''39''' :  2  (1935)  pp. 116–141</TD></TR>
 +
<TR><TD valign="top">[2]</TD> <TD valign="top">  A.A. Karatsuba,  "Fundamentals of analytic number theory" , Moscow  (1975)  (In Russian)</TD></TR>
 +
<TR><TD valign="top">[3]</TD> <TD valign="top">  K. Prachar,  "Primzahlverteilung" , Springer  (1957)</TD></TR>
 +
</table>

Latest revision as of 15:13, 3 July 2020

2020 Mathematics Subject Classification: Primary: 11M06 Secondary: 11N13 [MSN][ZBL]

Page's theorem on the zeros of Dirichlet $L$-functions.

Let $L(s,\chi)$ be a Dirichlet L-function, $s = \sigma + i t$, with $\chi$ a Dirichlet character modulo $d$, $d \ge 3$. There are absolute positive constants $c_1,\ldots,c_8$ such that

a) $L(s,\chi) \ne 0$ for $\sigma > 1 - c_1/\log(dt)$, $t \ge 3$;

b) $L(s,\chi) \ne 0$ for $\sigma > 1 - c_2/\log(d)$, $0 < t < 5$;

c) for complex $\chi$ modulo $d$, \begin{equation}\label{1} L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_3}{\log d}\,,\ |t| \le 5\,; \end{equation}

d) for real primitive $\chi$ modulo $d$, \begin{equation}\label{2} L(s,\chi) \ne 0\ \ \text{for}\ \ \sigma > 1 - \frac{c_4}{\sqrt{d}\log^2 d}\,; \end{equation}

e) for $2 \le d \le D$ there exists at most one $d=d_0$, $d_0 \ge (\log^2 D)/(\log\log^8 D)$ and at most one real primitive $\psi$ modulo $d$ for which $L(s,\psi$ can have a real zero $\beta_1 > 1- c_6/\log D$, where $\beta_1$ is a simple zero; and for all $\beta$ such that $L(\beta,\psi) =0$, $\beta > 1 - c_6/\log D$ with a real $\psi$ modulo $d$, one has $d \equiv 0 \pmod {d_0}$.

Page's theorem on $\pi(x;d,l)$, the number of prime numbers $p \le x$, $p \equiv l \pmod d$ for $0 < l \le d$, where $l$ and $d$ are relatively prime numbers. With the symbols and conditions of Section 1, on account of a)–c) and e) one has $$ \pi(x;d,l) = \frac{\mathrm{li}(x)}{\phi(d)} - E \frac{\chi(l)}{\phi(d)}\sum_{n \le x} \frac{n^{\beta_1 - 1}}{\log n} + O\left({x \exp\left({-c_7 \sqrt{\log x}}\right)}\right) \ , $$

where $E=1$ or $0$ in accordance with whether $\beta_1$ exists or not for a given $d$; because of (2), for any $d \le (\log x)^{1-\delta}$ one has for a given $\delta>0$, \begin{equation}\label{3} \pi(x;d,l) = \frac{\mathrm{li}(x)}{\phi(d)} + O\left({x \exp(-c_8 \sqrt{\log x})}\right) \ . \end{equation}

This result is the only one (1983) that is effective in the sense that if $\delta$ is given, then one can state numerical values of $c_8$ and the constant appearing in the symbol $O$. Replacement of the bound in (2) by the Siegel bound: $L(\sigma,\chi) \ne 0$ for $\sigma > 1-c(\epsilon)d^{-\epsilon}$, $\epsilon > 0$, extends the range of (*) to essentially larger $d$, $d \le (\log x)^A$ for any fixed $A$, but the effectiveness of the bound in (3) is lost, since for a given $\epsilon > 0$ it is impossible to estimate $c_8(\epsilon)$ and $O_\epsilon$.

A. Page established these theorems in [1].

References

[1] A. Page, "On the number of primes in an arithmetic progression" Proc. London Math. Soc. Ser. 2 , 39 : 2 (1935) pp. 116–141
[2] A.A. Karatsuba, "Fundamentals of analytic number theory" , Moscow (1975) (In Russian)
[3] K. Prachar, "Primzahlverteilung" , Springer (1957)
How to Cite This Entry:
Page theorem. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Page_theorem&oldid=50873
This article was adapted from an original article by A.F. Lavrik (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article