simplification of sums with binomial coefficients



Hi,

the default routines cant handle this case. Here's how you do it with
Zeilberger. You have

sm[k] = sum(r/k*binom(n,r)*binom(m,k-r)/binom(n+m,k),r,-inf,inf)

so you can find the recurrence for sm[k]:

(%i1) load(zeilberger)$
(%i2) load(solve_rec)$
(%i3) r/k*binom(n,r)*binom(m,k-r)/binom(n+m,k);
(%o3) (binom(m,k-r)*binom(n,r)*r)/(k*binom(n+m,k))
(%i4) summand_to_rec(%, r, k);
(%o4) sm[k+1]-sm[k]

which tells you that the sum is constant. Now

(%i5) sum(''%o3, r, 0, k), k=1;
(%o5) n/(n+m)

tells you that it is equal to n/(n+m).

Note that summand_to_rec is a wrapper for Zeilberger, which gives the
recurrence and some more data:

(%i6) Zeilberger(%o3, r, k);
(%o6) [[((r-1)*(r+m-k))/((n+m-k)*(r-k-1)),[-1,1]]]

If you get a more complicated recurrence you can try solving it with solve_rec.

Andrej

On 4/9/06, az <zoav1 at uic.nnov.ru> wrote:
> Hello,
> How do I make Maxima simplify
> sum(r/k*binom(n,r)*binom(m,k-r)/binom(n+m,k),r,0,k) ?
>
> simpsum doesn't seem to help. Maybe somebody can give an example of the
> Zeilberger
> share package usage in this situation?
>
> --
> Andrei Zorine
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>