View previous topic :: View next topic |
Author |
Message |
Pat
Joined: 23 Feb 2010 Posts: 207
|
Posted: Sun Nov 08, 2015 9:33 am Post subject: |
|
|
nataraj wrote: |
In the end, I settled for this cat: {28 cells given}
Code: |
X . . . . . . . X
X X . . . . . X X
X . X . . . X . X
X . . . X . . . X
. X X . . . X X .
. . . . X . . . .
X . . X X X . . X
. X . . . . . X .
. . . X X X . . .
|
Again, it gave me a puzzle every 1-14 tries,
with an almost even mix of Mild and Hard.
No Easy, no Very Hard.
|
slow, so i stopped after just 23 puzzles —
Code: |
2.......979.....154.3...6.23...1...8.29...45.....4....8..654..1.1.....6....321...
2.......979.....184.8...7.35...4...2.16...49.....6....3..654..1.4.....3....321...
3.......557.....812.1...3.41...8...7.34...56.....4....8..654..2.2.....1....321...
6.......124.....793.7...2.58...6...3.59...68.....7....9..654..7.2.....3....321...
6.......558.....292.9...4.19...3...2.38...17.....4....7..654..8.5.....6....321...
8.......114.....259.2...7.45...8...6.28...51.....4....2..654..9.6.....4....321...
8.......352.....749.6...1.56...9...8.17...59.....4....2..654..7.6.....3....321...
8.......475.....323.9...8.75...8...1.86...39.....6....1..654..9.4.....5....321...
2.......547.....831.8...9.76...3...1.83...67.....9....8..654..9.2.....1....321...
3.......186.....257.2...8.92...7...8.98...74.....6....1..654..3.2.....1....321...
3.......241.....859.2...4.65...8...3.71...84.....9....1..654..8.3.....2....321...
3.......949.....167.8...5.35...8...7.27...86.....6....1..654..8.4.....3....321...
4.......559.....271.3...9.49...1...6.38...57.....6....2..654..3.5.....1....321...
5.......736.....127.8...4.64...9...5.56...29.....6....1..654..3.3.....5....321...
5.......781.....534.7...9.22...3...5.38...76.....9....7..654..9.5.....2....321...
7.......662.....419.8...7.53...1...8.59...46.....4....2..654..9.6.....2....321...
9.......725.....434.3...5.97...6...2.45...78.....4....3..654..1.1.....3....321...
5.......771.....596.4...3.21...8...4.47...83.....6....2..654..3.5.....2....321...
1.......426.....354.9...6.77...3...2.18...37.....9....9..654..3.2.....5....321...
8.......363.....975.9...6.21...7...6.48...23.....4....3..654..1.1.....6....321...
9.......618.....736.3...2.82...9...5.79...82.....4....8..654..7.5.....6....321...
6.......415.....678.7...9.17...6...9.69...35.....8....9..654..3.1.....4....321...
2.......578.....465.6...9.11...3...7.47...32.....4....9..654..2.1.....6....321...
|
— listed roughly from easiest to toughest;
would you still say, "No Easy, no Very Hard"?
EDIT:
plenty more are quickly available by mutating any of the above;
e.g.4.......578.....395.6...8.11...3...7.47...32.....4....9..654..2.1.....6....821...
3.......578.....465.6...9.11...9...7.97...82.....4....9..654..2.1.....5....321...
— what ratings would you assign them? |
|
Back to top |
|
|
Pat
Joined: 23 Feb 2010 Posts: 207
|
|
Back to top |
|
|
Pat
Joined: 23 Feb 2010 Posts: 207
|
Posted: Sun Nov 08, 2015 1:18 pm Post subject: what rating would you assign these 2 VH puzzles? |
|
|
what rating would you assign these 2 VH puzzles?
gsf (2007.Jan.17) wrote: |
.473..8......1..5..132.8......1....5..15.26..7....3......4.516..9..6......4..127. |
in Extreme Su Doku, Wayne Gould wrote: |
.9.....6.5...2...4..17.49..9...5...3.5.4.2.9.4...1...8..62.53..3...9...6.7.....5. |
|
|
Back to top |
|
|
Pat
Joined: 23 Feb 2010 Posts: 207
|
Posted: Sun Nov 08, 2015 1:35 pm Post subject: re: puzzles for a given Layout |
|
|
|
|
Back to top |
|
|
nataraj
Joined: 03 Aug 2007 Posts: 1048 Location: near Vienna, Austria
|
Posted: Tue Nov 10, 2015 9:16 am Post subject: Pattern Based Sudoku Generation - Final Report |
|
|
Pattern Based Sudoku Generation - Final Report
Finally, found some time to take the new pattern based strategies from the test bench and integrate them into the Sudoku Farm. As a (for my needs) realistic test case, I went through the whole process of seed/grow/harvest/select in order to produce a "Special Edition": All grades from Easy to Advanced (I even included 2 "Extreme" - puzzles the Sudoku Helper cannot solve)
Pattern: The CAT (see above). 28 clues, horizontal symmetry.
Creation and preliminary selection:
I let it run for 2 minutes and stopped. The result was 32 puzzles:
Code: |
10.11.2015, 08:27:23 - AUTO on
Tue, 10 Nov 2015 07:27:50 GMT
EASY 5
MILD 10
HARD 12
VERY HARD 2
ADVANCED 1
EXTREME 2
UNKNOWN 0
COLOR SUDOKU 0
SQUIGGLY 0
X-SUDOKU 0
10.11.2015, 08:29:38 - AUTO off.
|
A decent mix. I selected one of the two very hards, found the advanced satisfactory and let the computer select the Easy to Hard.
Printed the result and was VERY disappointed. While the cat pattern is recognizable in the original pattern, it disappears when printed as a sudoku.
The (thin and thick) grid lines obscure the pattern, the (relatively) small numbers compared to the distance between them - it all adds up to this
conclusion:
Unless there is a very simple, easily recognizeable pattern AND a good reason to use a pattern in the first place, the benefit-cost ratio is just way too low to warrant the extra effort.
The Special Cat Edition can be downloaded here:
http://sudoku.saueregger.at/201511Cat%20Special.pdf
All puzzles can be found in the Sudok-o-Rama by entering their number (calendar icon, paste number, hit do it), e.g. 1635289030 and 1124913204 for the two "Extreme"s.
If anyone has a pen and paper method to solve these two, I'd much appreciate your help!
---------------------------------
It was a cool project, though. Liked it a lot.
Again thanks, Pat, for your guidance!
The Sudoku Farm now has a potentially useful feature that might come in handy once in a while. For special occasions (Valentine's Day, why not?).
Not for everyday use.
Helmut
______________________________________
Grids for the Extremes
1124913204:
Code: |
+-----------+-----------+-----------+
| 5 . . | . . . | . . 4 |
| 8 7 . | . . . | . 5 9 |
| 2 . 1 | . . . | 6 . 3 |
+-----------+-----------+-----------+
| 3 . . | . 2 . | . . 7 |
| . 6 5 | . . . | 2 9 . |
| . . . | . 7 . | . . . |
+-----------+-----------+-----------+
| 4 . . | 9 8 2 | . . 6 |
| . 8 . | . . . | . 4 . |
| . . . | 1 4 3 | . . . |
+-----------+-----------+-----------+
|
1635289030
Code: |
+-----------+-----------+-----------+
| 3 . . | . . . | . . 1 |
| 9 5 . | . . . | . 2 4 |
| 8 . 2 | . . . | 6 . 7 |
+-----------+-----------+-----------+
| 5 . . | . 3 . | . . 8 |
| . 8 9 | . . . | 7 6 . |
| . . . | . 1 . | . . . |
+-----------+-----------+-----------+
| 6 . . | 3 9 4 | . . 2 |
| . 3 . | . . . | . 9 . |
| . . . | 2 6 8 | . . . |
+-----------+-----------+-----------+
|
|
|
Back to top |
|
|
Pat
Joined: 23 Feb 2010 Posts: 207
|
Posted: Tue Nov 10, 2015 12:01 pm Post subject: re: "cat" — 28 cells given |
|
|
Pat wrote: |
4.......578.....395.6...8.11...3...7.47...32.....4....9..654..2.1.....6....821...
3.......578.....465.6...9.11...9...7.97...82.....4....9..654..2.1.....5....321...
| the 1st i suppose you'd call Mild
re the 2nd, in view of your earlier statement — nataraj wrote: | For VH I limit the number of empty cells during "fill"
(when PMs are written down)
to <7 per house. | — apparently you'd discard it,
while i'd call it a nice VH [ play ] |
|
Back to top |
|
|
tt1987
Joined: 20 Jan 2016 Posts: 1
|
Posted: Wed Jan 20, 2016 5:20 am Post subject: |
|
|
Quote: | Frequencies of graded Sudokus in test run
The number of sudokus requested was:
- 61 each for Easy, Mild, Hard
- 44 Very Hard
- 17 Advanced
The "Farm" creates at least 50% more than what is needed, so it stopped when 67 Very Hard were generated.
This is the result:
Code:
3.11.2015 15:05:49 - AUTO on
EASY 274
MILD 254
HARD 376
VERY HARD 67
ADVANCED 47
EXTREME 142
3.11.2015 15:17:51 - AUTO off.
TOTAL puzzles: 1160
What I do then, is look at the individual VH and ADV: which of the puzzles look interesting (not too many tedious steps, not too few hard steps). Don't want ALL m- or w-wings (sirloin steak every day??? ah, how dull!) or only xy-wings every day (I dimly remember having read comments about all VH puzzles being solved by xy wings on this very site ...)
BTW the statistics above do not include all the positions that were rejected during the "watering and pruning" phase because they produced too many solutions or none at all.
To give you a feeling about how many tries are needed to achieve one (as yet ungraded) sudoku, I include the production history of one VH Sudoku:
Code:
Tue, 03 Nov 2015 14:07:31 GMT
Seed:1246252372 Init: 1
C1: 8 clues, 6 tries, 1/2 ms.
C2: 26 clues, 1 solutions, 22(2/2/2/1/1/2/1/5/6) tries, 15/17 ms.
C3: 26 (+0 extra) clues, 1 solutions, 0/17 ms.
R1: 25 clues, 2 positions. 8/25 ms.
Statistics of steps taken: 5/90/5/4/1/1/0/0/0/0. time:146 ms.
It starts with a random seed.
First phase (germination): 6 tries to build a position of 8 clues
Second phase (growth): add clues until we have a unique solution. 22 tries. Optional (C3) add redundant clues to make the puzzle less tedious.
Third phase(pruning): remove clues to make the puzzle more interesting.
I did not record the tries during that phase.
All in all I'd give a conservative estimate of 25 positions tried for one ungraded sudoku.
The whole process (12 minutes real time) created some 25.000 positions to arrive at 44 "very hard"s. Maybe my carrot example was not fitting well enough - seems more like sprinkling a handful of stardust on magic ground to harvest a few moonstones Wink
By far the longest phase in the whole process is the grading phase: as you can see it needs 119 milliseconds - compared to the mere 25 milliseconds for all the tries before.
But without the grading phase all the rest would we in vain ..
Helmut |
nataraj I have done but failed |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|