cp's OEIS Frontend

This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.

A256535 The largest number of T-tetrominoes that fit within an n X n square.

Original entry on oeis.org

0, 0, 1, 4, 5, 8, 11, 16, 19, 24, 29, 36, 41, 48, 55, 64, 71, 80, 89, 100, 109, 120, 131, 144, 155, 168, 181, 196, 209, 224, 239, 256, 271, 288, 305, 324, 341, 360, 379, 400, 419, 440, 461, 484, 505, 528, 551, 576, 599, 624, 649, 676, 701, 728, 755, 784, 811
Offset: 1

Views

Author

Jack W Grahl, Sep 15 2015

Keywords

Comments

No T-tetromino fits in a 1 X 1 or 2 X 2 square: a(1)=a(2)=0. A single T-tetromino can be placed in a 3 X 3 square, and must occupy the center square. Four T-tetrominos fit within a 4 X 4 square with no spaces left over, in a rotationally symmetric tiling: a(4)=4.
For n = 4m, it is obvious that a(n) = n^2/4, by repeating the construction for n=4. For n = 4m + 2, it can be shown, using a chessboard coloring, that a(n) < n^2/4. By tiling an L-shaped strip of width 2 in a manner that can be indefinitely extended, one can show that a(n) = n^2/4 - 1.
Shuxin Zhan proved that it is not possible to tile a square of side 4m+1 or 4m+3 with T-tetrominos and a single monomino. Thus there must be at least 5 empty squares in any partial tiling by T-tetrominos. This bound is achieved for tilings in 5 X 5, 7 X 7, 9 X 9 and 11 X 11 squares. Robert Hochberg proved that for n > 11, there must be either 5 or 9 empty squares. He conjectured that 5 is always enough.
Jack W Grahl proved that, for squares, 5 monominos are always sufficient. This means that the sequence is given by n^2/4, (n^2-1)/4-1, n^2/4-1, (n^2-1)/4-1, for n = 4m, n = 4m+1, n = 4m+2 and n = 4m+3, respectively (which the exception of a(1) = 0), and generating function x^3*(-1-2*x+2*x^2-2*x^3+x^4) / ( (1+x)*(x^2+1)*(x-1)^3 ). - Jack W Grahl, Jul 25 2018

Examples

			The optimal tiling for a 4 X 4 square is:
   AAAB
   DABB
   DDCB
   DCCC
This forms the building block of a solution for all n a multiple of four.
For n=7 a solution is given by:
   ABBBCCC
   AABD/CE
   A/DDDEE
       F/E
       FFG
       FGG
       //G
with 5 empty squares, and the 4 X 4 square in the lower left filled in as above.
For n=6, a tiling of the excess after removing a 4 X 4 square shows us how optimal solutions can be generated for any even number that is not a multiple of 4:
   //ABBB
   /AAABC
       CC
       DC
       DD
       D/
The pairs A&B and C&D can be extended in the manner of a frieze. A nice solution for 9 X 9 does not include tilings of smaller even squares:
   ABBBCDDDE
   AABFCCDEE
   AGFFCHHHE
   GGGFI/HJ/
   KKKIIIJJJ
   /KL//MNNN
   OLLLPMMNQ
   OORPPMSQQ
   ORRRPSSSQ
		

Programs

  • Mathematica
    Delete[Flatten[ Table[{4n^2, 4n^2 + 2n - 1, 4n^2 + 4n, 4n^2 + 6n + 1}, {n, 0, 14}]], 2] (* or *)
    CoefficientList[ Series[1 + (x^4 - 2x^3 - 2x + 1)/((x - 1)^3 (x^3 + x^2 + x + 1)), {x, 0, 58}], x] (* Robert G. Wilson v, Jul 25 2018 *)
    LinearRecurrence[{2,-1,0,1,-2,1},{0,0,1,4,5,8,11},60] (* Harvey P. Dale, Aug 11 2024 *)
  • PARI
    concat([0,0], Vec(x^3*(1 + 2*x - 2*x^2 + 2*x^3 - x^4) / ((1 - x)^3*(1 + x)*(1 + x^2)) + O(x^60))) \\ Colin Barker, May 24 2019

Formula

From Jack W Grahl, Jul 25 2018: (Start)
a(4m) = 4m^2;
a(4m+1) = 4m^2 + 2m - 1;
a(4m+2) = 4m^2 + 4m;
a(4m+3) = 4m^2 + 6m + 1.
(End)
From Colin Barker, May 24 2019: (Start)
G.f.: x^3*(1 + 2*x - 2*x^2 + 2*x^3 - x^4) / ((1 - x)^3*(1 + x)*(1 + x^2)).
a(n) = (-7 + 3*(-1)^n + 2*(-i)^n + 2*i^n + 2*n^2) / 8 for n>1, where i=sqrt(-1).
a(n) = 2*a(n-1) - a(n-2) + a(n-4) - 2*a(n-5) + a(n-6) for n>7.
(End)