One strategy to prove the Prime number theorem involves removing some factors:

$$ \limsup_{x \to \infty} \underbrace{\frac{1}{x}\sum_{n \leq x} \mu(x)}_{\color{red}{A}}\leq \prod_{p \leq y} \left( 1 - \frac{1}{p}\right) \int_1^\infty \underbrace{ \frac{1}{t}\sum_{n \leq t} \mu(n)\mathbf{1}_y(n)}_{\color{blue}{B}} \times \frac{dt}{t} $$

The averages $A$ and $B$ are related since we are adding up the Möbius function on the subset of the integers whose factors $p \leq y$.

Here $\mathbf{1}_y(n) = 1$ if all the prime factors $p | n$ are less than $y$ and $0$ otherwise. Can't think of a better way of explaining this at the moment, it is called $v_y(n)$ in Daboussi's paper.

Proving that $\int_1^\infty \dots \ll \log y$ is rather convoluted. Are there any simpler ways to establish Daboussi's intermediate result?

  • 3
    $\begingroup$ I don't know. Have you seen the exposition of Daboussi's proof in Tenenbaum and Mendes France, The Prime Numbers and Their Distribution, published by the AMS as Volume 6 of the Student Mathematical Library? $\endgroup$ – Gerry Myerson Oct 1 '15 at 1:10
  • $\begingroup$ I think that while it is true that Daboussi’s proof is a nice application of what someone called ‘phipsiology’ (counting numbers without small/large prime factors), there is no way around the observation that Selberg’s elementary proof is simpler and easier to understand. $\endgroup$ – M Mueger Apr 21 at 8:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.