[seqfan] Chessboard polyominoes

Joseph S. Myers jsm at polyomino.org.uk
Sun Nov 28 02:51:12 CET 2010


Continuing my computations of various sequences related to various forms 
of generalised polyominoes, I tried to compute the sequences for 
"chessboard polyominoes" - A001933 and A001071.

The main reference for both those sequences (whose terms are unchanged 
from the 1973 book) is "W. F. Lunnon, personal communication" so I am just 
going by the names of the sequences to interpret what they mean.  The 
natural interpretation would be polyominoes with n squares cut from a 
chessboard (i.e. an infinite chessboard pattern), where for A001933 
rotating and turning over the polyomino is not considered to result in a 
different shape (and the reverse of the chessboard is supposed to have the 
same pattern) and for A001071 rotating is not considered to result in a 
different shape but two polyominoes related only by a reflection are 
considered different.

Unfortunately this doesn't reproduce the terms given for A001071.  For 
A001933 things are fine; I get 2, 1, 4, 7, 24, 62, 216, 710, 2570, 9215, 
34146, 126853, 477182, 1802673, 6853152, 26153758, 100215818 which agrees 
with the given terms and adds four more.  But for my interpretation of 
A001071 I get 2, 1, 4, 10, 36, 110, 392, 1371, 5000, 18251, 67792, 253040, 
952540, 3602846, 13699554, 52298057, 200406388, which disagrees on the 
even-numbered terms only from a(6) onwards.

I wonder in this case if the sequence is actually meant to be counting 
something else.  In particular, perhaps two one-sided chessboard 
polyominoes are also (in A001071) considered the same if the shapes of the 
two polyominoes (ignoring the colours) are the same or related by a 
rotation or translation, but the colourings are only related by a 
reflection.  This yields the value of a(6) in the sequence, since the 
relevant cases are (in ASCII art)

XO
O
X
OX

X
OX
XO
O

where a reflection swaps the colours and I think it also reproduces a(8).  
If Neil still has Lunnon's letter it might help elucidate what was meant.  
The rule I suggest above is outside the range of rules my polyform 
enumeration code can handle at present, but it would be interesting if 
anyone can implement something that reproduces the terms given in the 
sequence.

As for a sequence for the terms I computed for what I consider the natural 
interpretation of "one-sided chessboard polyominoes", the definition of 
A121198 is rather confusingly explained - that sequence is marked "obsc", 
but as best I understand it I think it is equivalent to what I computed as 
one-sided chessboard polyominoes, and so I think that A-number (with a 
rewritten description) will be the natural home for my terms.

-- 
Joseph S. Myers
jsm at polyomino.org.uk




More information about the SeqFan mailing list