New challenge

Programs which generate, solve, and analyze Sudoku puzzles

Re: Explications

Postby gsf » Wed Aug 23, 2006 2:14 pm

Papy wrote:I have selected two grids in the Gordon files

here are the initial candidate (or pencilmark) grids for the two puzzles
commented with the number of cells with i candidates
e.g., 17 cells have 1 candidate (the clues), 3 cells have 2 candidates, etc.
Code: Select all
# 17.3.5.16.24.13.3.0.0

 24568     3     124689 |  2489    2689   24689  |  468      7     12458
 24678   12467   124678 |   5      2368   24678  |  3468     9     12348
 245678  245679  246789 |   1     23689   246789 |  3468   234568  23458
------------------------+------------------------+------------------------
  3456     8     13469  |   49      7     14569  |   2       34     349
  2467   124679  124679 |  2489   12689     3    |   5       48     4789
 23457   24579   23479  |  2489    2589   24589  |   1      348      6
------------------------+------------------------+------------------------
   1      247      5    |   6      2389    2789  | 34789    2348   234789
   9      2467   234678 |  2378   12358   12578  | 34678  1234568 1234578
 23678    267    23678  | 23789     4     125789 | 36789   123568 1235789

# 17.3.7.11.29.8.5.1.0

  12368   125678   135678  |    9      13568      4    |   568     13568    13568
 134689   1456789 13456789 |  1367     13568    13578  |  45689    13568      2
 1234689  1245689  1345689 |  1236     13568    12358  |    7      13568   1345689
---------------------------+---------------------------+---------------------------
   169      169      169   |    8      1569     12579  |    3        4      15679
    5      14689      2    |  13467    13469    1379   |   689     1678     16789
    7      14689   134689  |  12346   134569    12359  |  25689    12568    15689
---------------------------+---------------------------+---------------------------
  2489     24589    4589   |   34        7        6    |    1      2358     3458
  14689   146789   146789  |    5        2      1389   |   468     3678     34678
  12468      3     145678  |   14       148      18    |  24568      9      45678
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

No for signature

Postby Papy » Wed Aug 23, 2006 2:42 pm

Sorry you make a new mistake on the number of signature

In my demonstration if we take only the 4 analyses with 9
digit give a 36 digit number
Now each digit could take the value (for 17 clues) 5 values
0-1-2-3-4 more with more clues
on a 30 clues you will have 36 digits of 10 values!!!
not 22567 but
10*10*10*10....*10 36 times!

Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Postby coloin » Wed Aug 23, 2006 3:25 pm

I feel I ought to intervene here.

Papy - you are working on a classification system which gsf doesnt understand.

I dont understand it either, but I feel sure that it does not classify puzzles the way you say it does.

gsf is showing the different pencilmarks in two puzzles - which are therefore different puzzles.

C
coloin
 
Posts: 1637
Joined: 05 May 2005

Classification

Postby Papy » Wed Aug 23, 2006 5:05 pm

Hi

It will be more efficace that I end my works.
80% are done
The lasts 20 seem to be easy so mly theory and its demonstration
will be complet.

I have some english friends so they ask them to translate the communication I prepare.
Thanks GSF to look at my work.
:)

Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Victory!!!!

Postby Papy » Thu Aug 24, 2006 1:39 pm

The secrets of the Sudoku are mine!

I have write the first sudoku generator 17 clues (or I beleive it)
WITHOUT STARTING FROM A VALID GRID AND WITHOUT SOLVING IT
Only logic.
No grid in my soft

My idea were good.
My lsoftware create sudoku wich are not in the gordon files
If someone explain me how to canonized' grids I make the complement!


Release today...

Papy
(It's true I'm proud of me!!!)
Papy
 
Posts: 131
Joined: 15 August 2006

To add to the gordon file

Postby Papy » Sat Aug 26, 2006 8:45 pm

Sgf

here isa link to a 177 clues files
http://rapidshare.de/files/30869638/17Clues.log.html

Can you check them to see if then can be add to the Gordon collection
I make this 1000 soduko with my new generatoor without morph
(1 or two minutes)

I think realaese it tomorrow!!!
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: To add to the gordon file

Postby gsf » Sun Aug 27, 2006 5:44 am

Papy wrote:Sgf

here isa link to a 177 clues files
http://rapidshare.de/files/30869638/17Clues.log.html

Can you check them to see if then can be add to the Gordon collection
I make this 1000 soduko with my new generatoor without morph
(1 or two minutes)

I'm sorry to say 215 have no solution and the other 785 have multiple solutions
does your solver find the same?
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

1000

Postby Papy » Sun Aug 27, 2006 6:56 am

Could you send me the two files please.
The victory is not here but near!!!
I thonk thet my problem is in the value of the clue The position for are good
but I not detct corrctly the umpossible solutions.
Can you explain me how youy cannonize the soduko
Thanks a lot
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: 1000

Postby gsf » Sun Aug 27, 2006 1:21 pm

Papy wrote:Could you send me the two files please.

the first puzzle has no solution
Code: Select all
..................................899.1.........6.15...7..18...89....1...1..4.7.3

and the second puzzle has many solutions
Code: Select all
.................................1.7.12.........16..9.9..8....1..8..4.3...4..68.9
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Re: 1000

Postby gsf » Sun Aug 27, 2006 1:55 pm

Papy wrote:Can you explain me how youy cannonize the soduko

see http://forum.enjoysudoku.com/viewtopic.php?p=33737#p33737
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Gsf

Postby Papy » Sun Aug 27, 2006 6:46 pm

Hi,

A difficulty appear choising the value.
and my test for valid grid is not gog
I have to work again
I send you my first grid
But working on 17 clues is long...

Papy

I have reaf about connonize I not understand all. I Ask question. But you you have an application to determine the morph?
I don't found one on the net/
Papy
 
Posts: 131
Joined: 15 August 2006

Re: Gsf

Postby gsf » Sun Aug 27, 2006 7:22 pm

Papy wrote:But working on 17 clues is long...

that's what made me question your 50K in 1 min in the first place ...
Papy wrote:But you you have an application to determine the morph?
I don't found one on the net/

my solver posted on the programmer's forum
the %c format of the -f option lists the input puzzles in (a) canonical form
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

@gsf

Postby Papy » Fri Sep 01, 2006 6:40 am

In the Girdon files you have,for example, this two grids
[code]
_______81_3_2______________1_8_6__4____7__3__6________5__3__7___9____2______1____
_______81_3_2______________1_8_6__4____9__3__6________5__3__7___7____2______1____[
/code]

Only a pair 7-9 is inversed
How do you determine that they are not equal wihout compare clues one by one (with possible permutation)
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: 1000

Postby udosuk » Fri Sep 01, 2006 8:21 am

gsf wrote:...and the second puzzle has many solutions
Code: Select all
.................................1.7.12.........16..9.9..8....1..8..4.3...4..68.9

I tried to copy & paste this puzzle into Simple Sudoku, and the program ran like crazy for several minutes... Afraid that it would crash the system I halted the task manually... I wonder would others encounter the same problem?

Never thought a puzzle could have this much destructional power...:(

Papy, 2 puzzles with totally identical layout of clues, but just one different clue, could be completely essentially different puzzles... If you don't realise this you won't go anywhere meaningful with this project...:idea:
udosuk
 
Posts: 2698
Joined: 17 July 2005

Postby Papy » Fri Sep 01, 2006 10:40 am

Of course !!!!
But at this time I try to understand how worh software to detect morh (canonine) But I found no explication on the algorythme:
Have you to solve grids? How manage the values of the celle,? Are the block useful?
At this this time , for me the two sample are identic because I don't know how to see the permutation. (and in the sample the grid have not been checked)
If the value are permuted how to see the change?



For your crasy grid I regard.
It seems that the problem is in sample sudoku
It must use a bad recursive method(PERHAPS)

Papy
The Sudoku made by Papy: new terrorism for your computer:D
Papy
 
Posts: 131
Joined: 15 August 2006

PreviousNext

Return to Software