Consider the sum $$\sum_{n=3}^N(\mu(n))^22^{\nu(n)}=\sum_{\rho_1=3}^N2+\sum_{\rho_1\rho_2=15}^N4+\sum_{\rho_1\rho_2\rho_3=105}^N8+\dots$$ where for each $i$ we have $\rho_i$ in the set of primes and $\ell=\prod_{i=1}^m\rho_i$ ranges over the set of squarefree values having $m=\nu(\ell)$ factors.

The simple upper bound I have come up with so far is


where $k$ is the largest integer such that $p_k\#\le N$. I believe that this simple upper bound is many orders of magnitude larger than the value of the sum for large $N$, but I don't know how to improve it.

Is there anything known about the value of this sum or a better upper bound on it?

  • 1
    $\begingroup$ This may be helpful: oeis.org/A069201 $\endgroup$ – Matthew Conroy Feb 1 '18 at 0:23
  • $\begingroup$ @MatthewConroy: $O(n\log n)$ sounds about right, thank you for the find! $\endgroup$ – abiessu Feb 1 '18 at 1:05

The comment by Matthew Conroy (https://math.stackexchange.com/users/2937/matthew-conroy) supplied the answer to my question, and also reminded me of the usual notation for the quantity which I labeled $\nu(n)$ in my question. This usual notation is $\omega(n)$ (http://oeis.org/wiki/Omega(n),_number_of_prime_factors_of_n_(with_multiplicity)) and represents the number of distinct prime factors of a number $n$.

The sequence listed at http://oeis.org/A069201 lists the asymptotic formula as $Cn\log n+O(n)$, where $C$ is the constant given by

$$C=\prod_p\left(1-\frac 1p\right)^2\left(1+\frac 2p\right).$$

| cite | improve this answer | |

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.