A recurrence - Triangle
Paul D. Hanna
pauldhanna at juno.com
Thu Jun 1 00:27:42 CEST 2006
Brendan, Seqfans,
One of the nice formulas that Max gave can obviously form a triangle:
> ss(n)=n!*sum(k=0,n,binomial(-x-2,k)/(n-k)!)
> For example, for n=0..10 we get the following values:
> ? vector(11,n,ss(n-1))
> %1 = [1, -x - 1, x^2 + 3*x + 3, -x^3 - 6*x^2 - 14*x - 11, x^4 +
> 10*x^3
> + 41*x^2 + 76*x + 53, -x^5 - 15*x^4 - 95*x^3 - 305*x^2 - 489*x -
> 309,
If row n of a triangle T(n,k) = coefficient of x^(n-k) in ss(n) as given
above,
we obtain:
1;
-1,-1;
3,3,1;
-11,-14,-6,-1;
53,76,41,10,1;
-309,-489,-305,-95,-15,-1;
2119,3623,2524,925,190,21,1;
-16687,-30414,-23121,-9569,-2345,-343,-28,-1; ...
which is not in the OEIS.
This triangle T can be expressed by
T = C*S*D
where
C = Pascal's triangle
S = A049444 ("Generalized Stirling number triangle of first kind")
D = diagonal matrix [1,-1,1,-1,...].
Just another alternate formula.
FYI,
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.seqfan.eu/pipermail/seqfan/attachments/20060531/96ad8981/attachment.htm>
More information about the SeqFan
mailing list