[seqfan] Re: strangeness in A008290 : Triangle T(n, k) of rencontres numbers
Max Alekseyev
maxale at gmail.com
Sun Sep 8 20:05:30 CEST 2013
Here is one possible approach:
O.g.f. for the n-th row of A008290(n,k) is
n!*\sum_{i=0}^n (-1)^i/i!*(1-x)^i.
So
\sum_{k=1}^n k^w * A008290(n,k) ) = (x*d/dx)^w n!*\sum_{i=0}^n
(-1)^i/i!*(1-x)^i at x=1.
At the same time, (x*d/dx)^w (1-x)^i at x=1 equals S(w,i)*(-1)^i*i!,
where S(,) are Stirling numbers of 2nd kind.
Therefore,
\sum_{k=1}^n k^w * A008290(n,k) ) = n!*\sum_{i=0}^n S(w,i) = n! * B(w).
P.S. There may be a simpler way to prove the same, e.g., from formula
(8) at http://mathworld.wolfram.com/BellNumber.html
Regards,
Max
On Sun, Sep 8, 2013 at 1:32 PM, Wouter Meeussen
<wouter.meeussen at telenet.be> wrote:
> call for help/explanation/insight : sum(k=1 ..n; k^w A008290(n,k) ) =
> B(w) n! for n>=w
>
> with B(w) = 1,2,5,15,52,203,877,4140,21147,115975,578570,... (A000110 : Bell
> or exponential numbers)
> why should this be?
>
> For those who wonder what happens for n<w, look at
> sum(k=1 ..n; k^w/n! A008290(n,k) ) for different values of n (within rows)
> and k (successive rows)
>
>
> 0 1 1 1 1 1 1 1 1 1 1 1 1
> 0 1 2 2 2 2 2 2 2 2 2 2 2
> 0 1 4 5 5 5 5 5 5 5 5 5 5
> 0 1 8 14 15 15 15 15 15 15 15 15 15
> 0 1 16 41 51 52 52 52 52 52 52 52 52
> 0 1 32 122 187 202 203 203 203 203 203 203
> 203
> 0 1 64 365 715 855 876 877 877 877 877 877
> 877
> 0 1 128 1094 2795 3845 4111 4139 4140 4140 4140
> 4140 4140
> 0 1 256 3281 11051 18002 20648 21110 21146 21147
> 21147 21147 21147
> 0 1 512 9842 43947 86472 109299 115179 115929 115974
> 115975 115975 115975
> 0 1 1024 29525 175275 422005 601492 665479 677359
> 678514 678569 678570 678570
> 0 1 2048 88574 700075 2079475 3403127 4030523 4189550
> 4211825 4213530 4213596 4213597
>
> If we dig a bit deeper, and look at the deficits of these rows versus their
> ultimate limit, we get a new surprise:
>
> 1 0 0 0 0 0 0 0
> 2 1 0 0 0 0 0 0
> 5 4 1 0 0 0 0 0
> 15 14 7 1 0 0 0 0
> 52 51 36 11 1 0 0 0
> 203 202 171 81 16 1 0 0
> 877 876 813 512 162 22 1 0
> 4140 4139 4012 3046 1345 295 29 1
>
> with row sums = A005493 : a(n) = number of partitions of [n+1] with a
> distinguished block.
> which calculates as Sum(k=1..n; k StirlingS2[n, k] ) for n=1 .. 16
>
> It is A137650 : Triangle read by rows, A008277 * A000012.
> calculated as
> Sum(j=0..k; Binomial[k, j] BellB[n - k + j] ) for n=0..7 and k=0..n
>
> So, a closed formula for sum(k=1 ..n; k^w A008290(n,k) ) is up for grabs.
> No?
> Is this new?
>
> Wouter.
>
> --------------------------- for my record --------------------------
> Table[Tr/@Table[k^w /n!
> Binomial[n,k]If[n-k==0,1,Round[(n-k)!/E]],{n,0,16},{k,0,n}],{w,16}];
> Table[ BellB[w]-Tr/@Table[k^w /n!
> Binomial[n,k]If[n-k==0,1,Round[(n-k)!/E]],{n,0,15},{k,0,n}],{w,16}];
> Table[ Sum[ StirlingS2[n,n-j],{j,0,n-k}],{n,7},{k, n}];
>
>
>
>
>
> _______________________________________________
>
> Seqfan Mailing list - http://list.seqfan.eu/
More information about the SeqFan
mailing list