[seqfan] Re: Counting polyominoes with a given "sprawl"
Allan Wechsler
acwacw at gmail.com
Wed Oct 14 04:02:08 CEST 2020
Sean, you have the idea exactly. Your second table, for free polyominoes,
is identical to one I calculated laboriously by hand some years ago; I only
got through about row 7, though. I hadn't considered doing the same
calculation for fixed polyominoes, even though that table is more relevant
to the probability of finding a given polyomino by random accretion.
Since you have the code and the data, would you like the honor of
submitting the new sequences? So far I see four of them, for free and fixed
polyominoes, and for cell-perimeter and "sprawl". Actually, maybe 6, if it
were convenient to do a similar calculation for edge-perimeter.
On Tue, Oct 13, 2020 at 4:21 PM Sean A. Irvine <sairvin at gmail.com> wrote:
> Hi Allan,
>
> Here is what I compute for the perimeter polynomials of fixed polyominos
> (A001168). The row is the number of cells in the polynomials, the column
> is the size of the perimeter. So, for example, there are 20 polyominos with
> 5 cells and a perimeter of 9. The sum of each row is A001168(n). These
> numbers are with respect to A001168. I think A001168 makes more sense if
> you are trying to calculate probabilities for various arrangements. In
> either case, the sprawl should correspond to antidiagonal sums from this
> table.
>
> 1 [0, 0, 0, 0, 1]
> 2 [0, 0, 0, 0, 0, 0, 2]
> 3 [0, 0, 0, 0, 0, 0, 0, 4, 2]
> 4 [0, 0, 0, 0, 0, 0, 0, 0, 9, 8, 2]
> 5 [0, 0, 0, 0, 0, 0, 0, 0, 1, 20, 28, 12, 2]
> 6 [0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 54, 80, 60, 16, 2]
> 7 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 22, 136, 252, 228, 100, 20, 2]
> 8 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 80, 388, 777, 818, 480, 152, 24, 2]
> 9 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 28, 291, 1152, 2444, 2804, 2089, 856,
> 216, 28, 2]
> 10 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 154, 986, 3676, 7612, 9750, 8192,
> 4330, 1416, 292, 32, 2]
> 11 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 52, 644, 3530, 11772, 24472, 33336,
> 31202, 19532, 8130, 2180, 380, 36, 2]
> 12 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 9, 325, 2644, 12502, 38694, 79730,
> 114342, 115502, 83183, 41136, 14064, 3208, 480, 40, 2]
> 13 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 112, 1660, 10480, 44574, 129020,
> 264482, 391432, 423786, 337144, 193820, 79240, 22993, 4508, 592, 44, 2]
> 14 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 28, 828, 7508, 41408, 158532,
> 437186, 887404, 1347560, 1538558, 1331170, 859176, 410302, 142624, 35664,
> 6160, 716, 48, 2]
>
> Here are the numbers for free polyominos (A000105), the antidiagonals seem
> to match your values as far as you calculated:
>
> 1 [0, 0, 0, 0, 1]
> 2 [0, 0, 0, 0, 0, 0, 1]
> 3 [0, 0, 0, 0, 0, 0, 0, 1, 1]
> 4 [0, 0, 0, 0, 0, 0, 0, 0, 3, 1, 1]
> 5 [0, 0, 0, 0, 0, 0, 0, 0, 1, 3, 5, 2, 1]
> 6 [0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 11, 10, 10, 2, 1]
> 7 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 18, 37, 30, 15, 3, 1]
> 8 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 11, 58, 99, 112, 60, 23, 3, 1]
> 9 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 62, 234, 521, 607, 450, 176, 50, 4,
> 2]
> 10 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 25, 127, 485, 957, 1249, 1025, 562,
> 177, 42, 4, 1]
> 11 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 9, 83, 456, 1484, 3093, 4182, 3940,
> 2453, 1039, 275, 53, 5, 1]
> 12 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 44, 354, 1575, 4907, 9984,
> 14398, 14447, 10477, 5143, 1794, 401, 67, 5, 1]
>
> Sean.
>
>
>
> On Wed, 14 Oct 2020 at 03:04, Allan Wechsler <acwacw at gmail.com> wrote:
>
> > Sean, your interpretation that sprawl = cells + perimeter is correct.
> >
> > I cannot find a sequence "number of polyominoes with perimeter n" either.
> > Starting with n = 3 (the first case that isn't vulnerable to definitional
> > quibbling), the sequence should go 0,1,0,1,1,5 ... OEIS has 15 hits for
> > this, but none of them have any chance of being the desired sequence. At
> > the moment I don't have a good guess for n = 9.
> >
> > On Sun, Oct 11, 2020 at 3:36 PM Sean A. Irvine <sairvin at gmail.com>
> wrote:
> >
> > > Hi Allan,
> > >
> > > I haven't looked too closely at exactly what you write, but there are a
> > > number of existing sequences which attempt to capture this general idea
> > in
> > > a variety of ways.
> > >
> > > One way is to compute the "perimeter" of the polyomino. I think the
> > > perimeter is very close to your sprawl, but does not include the cells
> of
> > > the polyomino itself. This definitely has a probabilistic
> interpretation
> > > in physics. An example of this is in A003203.
> > >
> > > There are also various sequence measuring diameter etc.
> > >
> > > Sean.
> > >
> > >
> > >
> > > On Mon, 12 Oct 2020 at 08:26, Allan Wechsler <acwacw at gmail.com> wrote:
> > >
> > > > The classic A000105 counts the number of polyominoes with a given
> > number
> > > of
> > > > cells.
> > > >
> > > > Define the "sprawl" of a polyomino to be the number of cells either
> in
> > > the
> > > > polyomino or edge-adjacent to it, when the polyomino is drawn on a
> > piece
> > > of
> > > > graph paper.
> > > >
> > > > For example, the R-pentomino has five cells, and is adjacent to nine
> > > more,
> > > > so it has a sprawl of 14.
> > > >
> > > > The sprawl is important when you are trying to calculate how likely
> it
> > is
> > > > to find a given polyomino on a field of black and white cells,
> randomly
> > > > colored with equal probability. (There are other terms, but the
> sprawl
> > is
> > > > important.)
> > > >
> > > > How many polyominoes are there with a sprawl n? Starting with n = 3,
> I
> > am
> > > > pretty sure that this sequence starts 0,0,1,0,0,1,0,1,1,3,2, and if
> > this
> > > > data is right, then the sequence is not yet archived in OEIS.
> > > >
> > > > It's a very obvious idea, and so I will be less surprised if I simply
> > > > counted wrong. Can anyone confirm these numbers?
> > > >
> > > > For n = 0, 1, or 2 there are definitional problems which permit
> > argument
> > > > about how the sequence gets going, but from n = 3 onward things seem
> > > fairly
> > > > clear.
> > > >
> > > > --
> > > > Seqfan Mailing list - http://list.seqfan.eu/
> > > >
> > >
> > > --
> > > Seqfan Mailing list - http://list.seqfan.eu/
> > >
> >
> > --
> > Seqfan Mailing list - http://list.seqfan.eu/
> >
>
> --
> Seqfan Mailing list - http://list.seqfan.eu/
>
More information about the SeqFan
mailing list