It is interest to solve some of the derived problems:
- given a particular division, generate the next (previous) division
- given a particular division, find its index in the lexicographically
order array of all divisions

For particular values of x and y, some interesting cases are
of reasonable size. e.g.

   (y%2) ness y=: 14
135135

So the case of 7 equal-sized subsets of i.14 requires */4 14 135135
= 7567560 bytes.  Even (8,16) should be possible:

   (4*y)*(y%2) ness y=: 16
1.2973e8



----- Original Message -----
From: John Randall <[EMAIL PROTECTED]>
Date: Thursday, July 24, 2008 7:28
Subject: RE: [Jprogramming] equal-size partitions
To: Programming forum <[email protected]>

> Here's a partial solution: it does not factor out permutations of the
> pieces.  I have some doubts as to how large a problem of this type one
> can realistically solve, given that the multinomial coefficients get
> so large.
> ...

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to