The Riemann zeta function and tuning/Appendix: Difference between revisions

Lériendil (talk | contribs)
mNo edit summary
Lériendil (talk | contribs)
derived euler product
Line 2: Line 2:


== 1. Euler product expression for the zeta function ==
== 1. Euler product expression for the zeta function ==
{{todo|derive this|inline=1}}
Starting with the definition of the zeta function:
 
<math>\displaystyle
\zeta(s) = \sum_{n = 1}^\infty \frac{1}{n^s}
</math>,
 
by the Fundamental Theorem of Arithmetic, and for <math>\mathrm{Re}(s) > 1</math>, where the series is absolutely convergent, it can be rearranged into a sum over prime factorizations:
 
<math>\displaystyle
\zeta(s) = \sum_{r_k = 0}^\infty \frac{1}{\left(\prod_{k = 1}^\infty {p_k}^{r_k} \right)^s} = \sum_{r_k = 0}^\infty \prod_{k = 1}^\infty \frac{1}{{p_k}^{s r_k}}
</math>,
 
where <math>p_k</math> is the ''k''-th prime.
 
The sum on the outside is really an infinite product of sum operators for each ''k'', and so:
 
<math>\displaystyle
\sum_{r_k = 0}^\infty \prod_{k = 1}^\infty \frac{1}{{p_k}^{s r_k}} = \prod_{k = 1}^\infty \sum_{r = 0}^\infty \frac{1}{\left({p_k}^s \right)^r}
</math>,
 
which is a geometric series, and given that <math>{p_k}^s</math> is greater than 1 for all ''k'' when <math>\mathrm{Re}(s) > 1</math>,
 
<math>\displaystyle
\zeta(s) = \prod_{k = 1}^\infty \frac{1}{1 - {p_k}^{-s}},
</math>
 
which is the Euler product formula we desired.


=== 1a. Dirichlet series for the von Mangoldt function ===
=== 1a. Dirichlet series for the von Mangoldt function ===
Line 10: Line 36:
<math>\displaystyle
<math>\displaystyle
\zeta(s) = \prod_{p \text{ prime}} \left(1 - p^{-s}\right)^{-1}
\zeta(s) = \prod_{p \text{ prime}} \left(1 - p^{-s}\right)^{-1}
</math>
</math>.


Take the natural logarithm to convert the product into a sum:
Take the natural logarithm to convert the product into a sum:
Line 16: Line 42:
<math>\displaystyle
<math>\displaystyle
\ln \zeta(s) = \sum_{p \text{ prime}} - \ln \left( 1 - p^{-s} \right)
\ln \zeta(s) = \sum_{p \text{ prime}} - \ln \left( 1 - p^{-s} \right)
</math>
</math>.


==== Logarithmic Series Expansion ====
==== Logarithmic Series Expansion ====
Line 24: Line 50:
<math>\displaystyle
<math>\displaystyle
-\ln(1 - x) = \sum_{k=1}^{\infty} \frac{x^k}{k}
-\ln(1 - x) = \sum_{k=1}^{\infty} \frac{x^k}{k}
</math>
</math>.


Substitute \( x = p^{-s} \).  
Substitute \( x = p^{-s} \).  
Line 31: Line 57:
<math>\displaystyle
<math>\displaystyle
-\ln\left(1 - p^{-s}\right) = \sum_{k=1}^{\infty} \frac{\left(p^{-s}\right)^k}{k}
-\ln\left(1 - p^{-s}\right) = \sum_{k=1}^{\infty} \frac{\left(p^{-s}\right)^k}{k}
</math>
</math>.


Simplify the expression.
Simplify the expression.
Line 39: Line 65:
\sum_{k=1}^{\infty} \frac{\left(p^{-s}\right)^k}{k}
\sum_{k=1}^{\infty} \frac{\left(p^{-s}\right)^k}{k}
= \sum_{k=1}^\infty \frac{1}{k} \left(p^k\right)^{-s}
= \sum_{k=1}^\infty \frac{1}{k} \left(p^k\right)^{-s}
</math>
</math>.


We thus have:
We thus have:
Line 56: Line 82:
= \sum_{p \text{ prime}} - \ln \left( 1 - p^{-s} \right)
= \sum_{p \text{ prime}} - \ln \left( 1 - p^{-s} \right)
= \sum_{p \text{ prime}} \sum_{k=1}^\infty \frac{1}{k} \left(p^k\right)^{-s}
= \sum_{p \text{ prime}} \sum_{k=1}^\infty \frac{1}{k} \left(p^k\right)^{-s}
</math>
</math>.


==== von Mangoldt function ====
==== von Mangoldt function ====
Line 68: Line 94:
0 & \text{otherwise}
0 & \text{otherwise}
\end{cases}
\end{cases}
</math>
</math>.


Reindex the double sum.
Reindex the double sum.
Line 77: Line 103:
<math>\displaystyle
<math>\displaystyle
\frac{\Lambda(n)}{\ln n} = \frac{\ln p}{k \ln p} = \frac{1}{k}
\frac{\Lambda(n)}{\ln n} = \frac{\ln p}{k \ln p} = \frac{1}{k}
</math>
</math>.


Therefore:
Therefore:
Line 83: Line 109:
<math>\displaystyle
<math>\displaystyle
\ln \zeta(s) = \sum_{n=2}^\infty \frac{\Lambda(n)}{\ln n} \cdot \frac{1}{n^s}
\ln \zeta(s) = \sum_{n=2}^\infty \frac{\Lambda(n)}{\ln n} \cdot \frac{1}{n^s}
</math>
</math>,


which is valid for \(\mathrm{Re}(s) > 1\), with the sum starting at \(n=2\) because \(\Lambda(1) = 0\).
which is valid for \(\mathrm{Re}(s) > 1\), with the sum starting at \(n=2\) because \(\Lambda(1) = 0\).