4
$\begingroup$

Let $$ \forall_{n=1\ 2\ \ldots}\quad I(n)\ :=\ \frac {(6\cdot n-3)!!}{(2\cdot n-1)!!\cdot(4\cdot n-3)!!} $$

Let $\ M(n)\ $ be the smallest natural number such that

$$ M(n)\cdot I(n)\ \in\ \mathbb Z $$

is an integer. What is the behavior of sequence $\ M(n)$?  As a minimum, the Prime Number Distribution Theorem tells us that $\ M(n)\rightarrow\infty.\ $ But I'd like to ask about more:

Q1   Is $\ M(n) < \left(\frac 43\right)^{3\cdot n}\ $ for all large $\ n?$

Q2   What is the least non-negative real $\ a\ $ such that for every $\ \epsilon > 0\ $ the inequality $\ M(n) < (a+\epsilon)^n\ $ holds for all large n.  (Actually, $\ a \ge \frac {4\cdot e^2}{27})$.

Q3   Would above constant $\ a\ $ be $\ a=\frac {4\cdot e^2}{27}$?   (Highly unlikely).

$\endgroup$
1
  • $\begingroup$ Just in case, let me mention that I just have fixed a silly typo (twice): now I have the correct Euler $e$ (i.e. $\exp(1)$) instead of accidental $\pi$. $\endgroup$ Commented Jan 16, 2017 at 18:57

1 Answer 1

12
$\begingroup$

We have $$ I(N)=\frac{(6N-2)!(N-1)!}{2(4N-2)!(3N-1)!}. $$ We count the exponent of a prime $p$ in $k!$ using a standard formula $[k/p]+[k/p^2]+\dots$. This gives the sum $w(p):=\sum_{s} h(n/p^s)$ for the exponent of $p$ in the number $I(n)\frac{6n-1}{4(4n-1)}$, where we denote $h(x)=[x]+[6x]-[3x]-[4x]$. $h$ takes the values $\pm 1$ and 0. And $h(x)=-1$ only when $x$ lies between $k+3/4$ and $k+5/6$ for some integer $k$. Your number $M(n)$ equals $\prod_p p^{\min(-w(p),0)}$. The powers $p^s$ for $s\geqslant 2$ do not affect the asymptotics (as usual).

Thus asymptotically $M(n)$ behaves as a product of primes $p$ for which $n/p$ lies in $(k+3/4, k+5/6)$ for integer $k$. By the PNT the logarithm of this product behaves as $$n\cdot \sum_{k=0}^\infty \left(\frac1{k+3/4}-\frac1{k+5/6}\right)=n\left(\frac{\sqrt{3}-1}2\pi-\log\frac{27}4\right)=0.195\ldots \cdot n,$$ this answers all three questions.

$\endgroup$
1
  • $\begingroup$ I've recognized your answer. Isn't your constant $\ 0.195...\ $ the extra part over 1? I.e. shouldn't your constant, as asked in the **q.**, be $\ 1.195...$ ? $\endgroup$ Commented Jan 16, 2017 at 19:05

Your Answer

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

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