[seqfan] Re: A possible new Keyword
Max Alekseyev
maxale at gmail.com
Thu Sep 24 20:03:24 CEST 2015
Notice that there is also a way to enter coefficients of polynomials in
infinite number of variables without referring to partitions.
This is what I call prime factorization order:
For n with the prime factorization n = p1^k1 * p2^k2 * ..., where p1=2,
p2=3, ... are the primes in their natural order (so only a finite number of
ki are nonzero),
we let a(n) be the coefficient of u[1]^k1 * u[2]^k2 * ....
E.g., multinomial coefficients C(k1+k2+...; k1, k2, ...) arranged this way
give http://oeis.org/A008480
Regards,
Max
On Thu, Sep 24, 2015 at 11:47 AM, Juan Arias de Reyna <arias at us.es> wrote:
>
> Frequently we consider sequences of polynomials in an infinite number of
> variables.
>
> For example the fourth polynomial in a given sequence I find is
>
> > - u[1]^4 + 4 u[1]^2 u[2] - 2 u[2]^2 - 4 u[1] u[3] + 4 u[4]
>
> There is a coefficient for each partition of n = 4. In this case each
> coefficient is
> related to a partition of 4
>
> 4 = 1 + 1 + 1 + 1 coef -1
> = 1 + 1 + 2 coef 4
> = 1 + 3 coef -4
> = 2 + 2 coef -2
> = 4 coef 4
>
>
> To include the sequence of coefficients we want to give a canonical order
> to
> all possible partitions.
> I will consider useful that all partition of the number n precede all
> partitions of n+1.
>
> But how to order the partitions of a number n?
> Perhaps writing the summands in increasing order and then by lexicografic
> order as I have
> written above the partitions of 4.
>
> Maybe somebody here in OEIS has a better election.
> I will like to fix a particular order,
> then in all cases that a sequence of
> this type is included in OEIS this same order should be used. Also a new
> keyword as the
> actual tabf and tabl should be added to the Keywords file, to indicate
> this type of sequence.
>
>
> Somebody know sequences of this type already included in OEIS?
> Has this problem appeared before?
>
> Best regards,
> Juan Arias de Reyna
>
> _______________________________________________
>
> Seqfan Mailing list - http://list.seqfan.eu/
>
More information about the SeqFan
mailing list