Asymptotic for A058797
Alec Mihailovs
alec at mihailovs.com
Sun Aug 21 07:47:10 CEST 2005
"Franklin T. Adams-Watters" <franktaw at netscape.net> wrote:
> The entry for A058797 states that it is asymptotic to c*n!, with c=.224...
> I conjecture that c = BesselJ(0,2) = Sum (-1)^k/(k!)^2 = 0.223890779...
> (A091681).
I found a formula for A007754:
a(n,k) = Pi*(BesselJ(n+k+1,2)*BesselY(k,2) - BesselY(n+k+1,2)*BesselJ(k,2))
In particular, for k=0, we get a formula for A058797,
a(n) = Pi (BesselJ(n + 1, 2)*BesselY(0, 2) - BesselY(n + 1, 2) BesselJ(0,
2))
Now, BesselJ(n+1,2) ~ 1/(n+1)! and BesselY(n+1,2) ~ -n!/Pi
That gives exactly that asymptotics, a(n) ~ BesselJ(0,2)*n!
Similarly, the asymptotics for the k-th column of A007754 is
a(n,k) ~ BesselJ(k,2)*(n+k)!
(for fixed k and n -> infinity).
Alec Mihailovs
http://math.tntech.edu/alec/
More information about the SeqFan
mailing list