3
$\begingroup$

The question here about estimating positive moments of the divisor function on sets of nonzero density $A\subset \{1,2,\ldots,x\}$ was answered giving $$ S_a(x):=\sum_{n \in A} d(n)^a \geq |A|(\ln x)^{a(\ln 2-\delta)}\quad (1) $$ for any $\delta>0,$ arbitrarily small.

This relies on the fact that after removing rare large values the divisor function in $\{1,2,\ldots x\}$ is essentially of size $(\ln x)^{\ln 2}$ on average.

What if I now "clip" the divisor function at level $m=(\ln n)^{2 \ln 2}$? So consider lower bounding the sum $$ S_{a,m}(x):=\sum_{n \in A} \left\{\max(d(n),m)\right\}^a $$

It seems to me that this should make no difference to the lower bound in (1), but I am not sure how to make this rigorous.

$\endgroup$

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.