[seqfan] Re: Tatami

Richard J. Mathar mathar at mpia-hd.mpg.de
Thu Mar 24 17:06:42 CET 2016

The formula in http://list.seqfan.eu/pipermail/seqfan/2016-February/016160.html
seems to be correct. I computed (by brute force enumeration with a C++ program)
the arrangements of tatami tilings of the 3 X (2k+1) floor, k>=0,
with one monomer at an arbitrary position and 3k+1 dimers. Arrangements
which are symmetry-related by rotations or flips are counted with
multiplicity. The numerical result is

2, 10, 18, 38, 72, 136, 250, 454, 814, 1446, 2548...

This seems to have the g.f. -2*(x-1)*(2*x^2+4*x+1) ) / (x^2+x-1)^2  .

Then we have in the more general setup the number of tatami tilings of 
the (2m+1) X(2k+1) floor with one monomer at an arbitrary position
and 2*(2*m*k+m+k) dimers.  The result is symmetric in m <-> k and it 
suffices to tabulate the counts as a triangle m>=0, 0<=k<=m:

2  10
3  18 10
4  38  8 10
5  72 18  4 x
6 136 24  x x x
7 250 32  x x x x

The second column (or row) is the sequence already shown above.
The first column are the integers because in the 1 X (2k+1) case
the monomer can be placed basically at any position which leaves
an even number of free spaces at both sides for the dimers.


More information about the SeqFan mailing list