Re: -- A sequence or real numbers
- From: "Mariano Suárez-Alvarez" <mariano.suarezalvarez@xxxxxxxxx>
- Date: Thu, 14 Feb 2008 18:00:50 +0000 (UTC)
On Feb 13, 5:30 pm, c...@xxxxxxxxxxxxxxx wrote:
Does anyone know, or at least have a clue on the following question ?
Let (b_n)_n be the sequence defined by :
b_1=1/4 , b_2=1/16 and
b_{n+1}=b_1 b_n+b_2 b_{n-1}+ ... +b_k b_{n-k+1}+ ... +b_1 b_n
Ex: b_3=b_1 b_2+b_2 b_1=1/32
b_4=b_1 b_3+b_2 b_2+b_3 b_1=5/256 etc.
Finally, let
X_n=b_n . n^{3/2}
A computer evaluation hints to the fact that (X_n) is decreasing to a
limit approx equal to 1.141...
Proof (?)
Note: the initial values (i.e. b_1,b_2) seems to be essential. Any
change, and (X_n)_n appears
to behave chaotically. [again, computer says so ]
Any information is warmly welcome.
Thanks
cosv
Define a (formal) series putting
f = sum_{n >= 1} b_n t^n
One can easily see this actually has a positive convergence
radius. Your recurrence relation means that
f^2 = f - 1/4
Solving this tells us that
f = (1 - sqrt(1 - t)) / 2
Therefore you can compute the b_n exactly by
figuring out the derivatives of f:
b_n = f^{(n)}(0) / n!
The derivatives at zero have a simple form,
which will let you find what you want.
-- m
.
- References:
- -- A sequence or real numbers
- From: cosv
- -- A sequence or real numbers
- Prev by Date: Re: -- A sequence or real numbers
- Next by Date: HPCNCS-08 Draft paper submission deadline is just few days from now
- Previous by thread: Re: -- A sequence or real numbers
- Next by thread: Re: -- A sequence or real numbers
- Index(es):
Relevant Pages
|