The New Sudoku Players' Forum

[ChPicard]

Hi

I read the 17 given sudokus from Gordon Royle, a 40 000 sudoku file. After reducing the sudokus with rotation, translation and relabelling, I didn't found two equivalent sudokus.

Well done

[Adak]

Hi

I read the 17 given sudokus from Gordon Royle, a 40 000 sudoku file. After reducing the sudokus with rotation, translation and relabelling, I didn't found two equivalent sudokus.

Well done

Howdy,

Would you tell me what you mean by "translation" and "relabelling", please.

I thought the unique test was rotation and symmetries of reflexion.

Thanks

[Pat]

Would you tell me what you mean by "translation" and "relabelling", please.

I thought the unique test was rotation and symmetries of reflexion.

equivalence of puzzles

[ChPicard]

Would you tell me what you mean by "translation" and "relabelling", please.

Before I discovered canonicalize function in the sudoku program from Glenn Fowler, I developped the same idea but with other steps than described in this forum by Lummox JR in his article sent on Wed Mar 29, 2006 12:36 am.

First I worked with the puzzle grid and not the solution. I ordered the grid such a way than the box with the bigger number of clues is in the box 1, the next in box 2 ... Rotation 90 degrees is permitted.
After I permuted rows to have the bigger number of clues in the first row ...

After I relabelled to have the first digit 1 the second 2 ...

And with this technic, I verified than in the Gordon Royle's list, there are not 2 identical grids.

Jean-Pierre

[tarek]

groyle list of 17s has already been tested for that, so testing it with your program may show a false -ve.....

Try the top1465 puzzle list with your program ......... does it have isomorphs that already has been tested & it does have isomorphs.

tarek