0
$\begingroup$

let $\forall n $ integer $p_n(t)=\frac{1}{n!}(t^n(1-t)^n)^{(n)} $ i 'm looking for an explicit constant $0<c<1$ ( very good small c ) independant of $n$, and a constant $b$ ( non explicit) independant of $n$ such that $ I_n:=| \int_{0}^1 \int_{0}^1 \frac{P_n(x)P_n(y)}{1-xy}dxdy| \leq b c^n$ or such that $(I_n)^\frac{1}{n} $ equivalent to $ c $

remarque 0: this integral is similar to Beukers integral

remarque 1: after integrating n times for $ y $ we get $I_n =| \int_{0}^1 \int_{0}^1 \frac{P_n(x)x^ny^n(1-y)^n}{(1-xy)^{n+1}}dxdy| $

remarque 2: $ I_n= \displaystyle \sum_{k=0}^{+\infty} \int_{0}^1 P_n(x)P_n(y)x^ky^k dx dy= \displaystyle \sum_{k=n}^{+\infty} (\int_{0}^1 P_n(x)x^k dx )^2 = \displaystyle \sum_{k=n}^{+\infty} (\int_{0}^1 x^n(1-x)^n \frac{(-1)^n}{n!} k(k-1)..(k-n+1) x^{k-n} dx )^2 $ \ $ =\displaystyle \sum_{k=n}^{+\infty} {{\Big{(}} \frac{k(k-1)..(k-n+1)}{(k+1)(k+2)..(n+k+1)}\Big{)}}^2 $ so we can may be use the hypergeometric term to find c

remarque 3: i'm not sure but i think that $ ( \int_{0}^1 \int_{0}^1 \frac{P_n(x)P_n(y)}{1-xy}dxdy)_{ n \geq 0} $ satisfy a recurrence relation (as Apery number), i don't want that method to find $c$ using poincarre theorem

$\endgroup$
3
  • $\begingroup$ Hasn't Fedor Petrov convinced you that the decay is merely polynomial? Or are you just ignoring the answers unless they confirm your conjectures? Voting to close. $\endgroup$
    – fedja
    Commented May 13, 2019 at 4:23
  • $\begingroup$ i'm sorry for mr Fedor Petrov, he doesn't gave the answer that i was expecting ( i thought that my old question was understable because i said as beukers integral), and upper bound or equivalent to beukers integral are given by $c^n$ and not $O(1/n^2)$), so i'm looking another answer and this time i think that the question is clear.. $\endgroup$
    – mamiladi
    Commented May 13, 2019 at 4:34
  • 1
    $\begingroup$ Yes, the question is clear. It amounts to "I was shown that the decay is polynomial (Fedor Petrov gave you both the upper and the lower bound) but I don't like it, so, could someone, please find an exponential bound for me?". That is not how mathematics works. We are neither law, nor religion: another lawyer or priest won't tell you anything different. Fedor's answer was final. $\endgroup$
    – fedja
    Commented May 13, 2019 at 4:56

0

Browse other questions tagged .