| New puzzle generator |
Brian
Kwon-Tom Obsessive
Puzzles: 4760
Best Total: 9m 6s | Posted - 2009.05.01 17:11:38
Quote: Originally Posted by v_e_e_n_c_a Does anybody know if there exists a way how to convert puzzles from archive to Loopy string without rewriting it? |
|
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.05.01 17:38:38
Quote: Originally Posted by brian
What I did was view page source, copy the info ".2..3..22..002..2312.....12...........2....223.2...2..221...222..1231.212.212...311.2...01...32.1..2", replace '.' with 'a', and add "10x10:" to the front.
|
Thanks, it works fine.. |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.05.25 09:04:18
Another two symmetrical puzzles - should be little bit harder
First
Second |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.06.07 14:42:25
Added new feature - puzzles are harder now and still without backtrack.. Puzzles of size 10x10 are generated in about 0,25 sec and 20x14 in about 1,5 sec.. Please post me some feedbacks about current difficulty of puzzles..
Here are some:
10x10 No. 1
10x10 No. 2
10x10 No. 3
10x10 No. 4
20x14 No. 1
Puzzle 2 hyperlink didn't work correctly - erased - only loopy string can be found later (down)
20x14 No 3
20x14 No 4
Or their loopy strings:
10x10:323a2b12b12d2a22e1b2b01a2c2b2a2b2a33a2b11a2a2a212b2c2a32221c21b111c32b1a23a
10x10:3212c22d2a12a22a21221a2a33d1a2b21c2a2222a2a2b232b22a2b2a221b3e221b3a3a3b12a3b2
10x10:3a112i22a11222a132213c21a1b1322b1c2a2a1a2b21a2a2b3b3b23a3a1c2e3d3a3b1a12
10x10:222c1a12232f21d3a2c1a111b222c21c2a222c1a32a3b2a1a2d2a123c3a2a3a212231f
20x14:33a1a2b12b2c21e2b2322c31a3113a1d2b3f211a2b20a12c2a12d23a3a32b11c2232b1a1b2a3a1c22a1131a3a3g1311a2a2a2b32b2a22f3b2a131a3c2a3b1b1a1a1b2c2c311a2c2132c2313e23a3d02221e1b3a3c3e0a11a2e3a23a222c21a1a322a
20x14:2a22b21a2a1a12a2d2a1d2122c21a21c1112a1a2a2a2a21b2112a1d1a22212222a2b1122a2b1a2b1c1b112a2d222a22c2c1b2a1c1a12a11a2a21a221b1c2b2121a1a1212212b2a212c2121d22a22b2d11a2a12a22a1a21a111d1b21221d22a12a2122b1b21111d12c2a1b2222a2aa
20x14:3b222d122222b2a2f13a1b2a223b2b2b222a1b2a3b2b312122c12d13a32b212212d2c1d1c2a31322b2d132a1b2b3d10a13b0a22a2b311a1a3a12b2c1122c21d2a3a21a3b2d22e1g1223b12a2a223a1b3a21b3b31e1b2e13a2a1b3a233c332a2a1b2b
Last edited by v_e_e_n_c_a - 2009.06.07 15:01:45 |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.06.12 12:57:37
One more feature added to puzzle generation / solving.. Should be little bit harder to solve.. Time needed to generate puzzle 20x14 of this difficulty is in avarage 1,6 sec.. I posted one example to archive.. Your comments or solutions are wellcome.. Thanks.. |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.08.16 12:30:24
I introduced some advanced features to my program - edges interactions and advanced avoiding small loops..
Puzzles should be more difficult, but i still don't use any backtrack, so it is really fast..
Please add comments on difficulty..
This is probably the last thing I made in the solver..
Puzzle 1
Puzzle 2 |
MondSemmel
Kwon-Tom Obsessive
Puzzles: 6159
Best Total: 7m 47s | Posted - 2009.08.16 15:26:08
I liked both puzzles (1m33s and 1m30s for the first and second puzzle respectively). Two puzzles of any difficulty setting are probably not a sufficiently large sample size, but based on just these two, I'd compare them to some of the slightly-harder-than-medium-difficulty Friday puzzles?
Solving them did not require any trial & error whatsoever but did otherwise require some quite advanced techniques.
Great job! |
Jankonyex
Kwon-Tom Obsessive
Puzzles: 5669
Best Total: 9m 35s | Posted - 2009.08.17 03:37:06
Quote: Originally Posted by v_e_e_n_c_a |
solutions
loopy |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.08.17 06:17:01
Thanks for comments.. Here are also some randomly generated larger puzzles.. Size 20x14 (weekend puzzle-sized)
Puzzle 3
Puzzle 4
Puzzle 5
Last edited by v_e_e_n_c_a - 2009.08.17 06:17:40 |
MondSemmel
Kwon-Tom Obsessive
Puzzles: 6159
Best Total: 7m 47s | Posted - 2009.08.17 07:19:08
I don't use the highlander technique on non-official or non-BotM puzzles, which made these puzzles much harder than they otherwise would have been - I had solving times of 7m13s, 12m57s and 6m51s for the 3rd, 4th and 5th puzzle respectively. That being said, I think they weren't really more difficult than, say, official weekend puzzles which take me about 4m to 4m30s to solve - there were lots of highlander opportunities I intentionally didn't use.
Still, they were quite difficult  . Thanks for the puzzles, I enjoyed solving them. |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.08.17 07:47:39
You can use highlander, cause I can ensure the uniqness of my puzzles..  I wanted to write very fast generator, which can generate also harder puzzles.. So they can't be much harder than the puzzles on this site - simply because I don't use backtrack in the solver (no trying possibilities).. Puzzles on this site are generated with backtrack and probably the time to generate them is much slower.. I am writing this program for players who wants to play relatively hard puzzles, but they don't want to wait for them too long..
Times for generating 20x14 puzzles of this difficulty is 3 - 3.5 sec.. For 10x10 it is about 0.5 sec and beasts are generated in about 1.1 minute.. I can post also some beasts if you like or some more puzzles of other size..
I am now working on GUI with lot of usefull functions for solving puzzles..
Thanks again for comments..
2 MondSemmel - if you want, i can post beta version of my program, when i will finish it - for trying it and testing - if you are intrested in.. (But it will be only for Windows)
Last edited by v_e_e_n_c_a - 2009.08.17 07:49:14 |
MondSemmel
Kwon-Tom Obsessive
Puzzles: 6159
Best Total: 7m 47s | Posted - 2009.08.17 07:56:20
Quote: Originally Posted by v_e_e_n_c_a 2 MondSemmel - if you want, i can post beta version of my program, when i will finish it - for trying it and testing - if you are intrested in.. (But it will be only for Windows) |
I'd love to check it out! I'm not sure how much time I'll be able to spend to test it in the near term (I'll soon be away at GamesCom for the remainder of this week) but I'm certainly going to try creating weekend puzzles or beasts with it . |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.08.20 10:31:13
As I said before my solver don't use backtrack and from this reason it is very fast.. I tried to compare also difficulty with other slither link apps..
E.g.: for this puzzle (randomly generated with my program with advanced logic switched on)
10x10:c11b2b2a3a1212a22a1b1233d2a2d3b0a12a21a2a1e2b11c2b31b2c33b1a23a1c2a32c2a
Loopy : only four crosses were found (around 0)
LoopDeLoop : four crosses found (around 0) and one cross between 1 & 1 up.. Nothing else without trying possibilities.. (Rated as SI10 - Depth:6 and S - UnFin:82%) |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.11.04 10:07:52
I added a lot of features to my solver and it is much harder, still with polynomial time (no backtrack). Check it out. There are 5 randomly generated puzzles
P1
P2
P3
P4
P5
Last edited by v_e_e_n_c_a - 2009.11.04 10:18:09 |
Tilps
Kwon-Tom Obsessive
Puzzles: 6504
Best Total: 20m 22s | Posted - 2009.11.04 21:47:29
Unless you invoke full solver, my solver is polynomial time. It is just a rather high polynomial...
Backtracking is only non-polynomial if it is an unlimited search.
That being said, how many patterns have you had to program in to your solver? I had a look at that sample problem from august and I can see a few patterns present, but I can't see any progress from application of systematic approaches without trials.
I define a pattern as something which can be used consistently but the original derivation of that pattern requires trials. |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.11.05 06:48:58
Quote: Originally Posted by tilps Unless you invoke full solver, my solver is polynomial time. It is just a rather high polynomial...
Backtracking is only non-polynomial if it is an unlimited search.
That being said, how many patterns have you had to program in to your solver? I had a look at that sample problem from august and I can see a few patterns present, but I can't see any progress from application of systematic approaches without trials.
I define a pattern as something which can be used consistently but the original derivation of that pattern requires trials. |
Today my program contains a lot of features (about 15-20 such as advanced loop avoiding, path deductions, advanced coloring and so on). The most of these features uses very, very small, but expressive set of rules (patterns).
To the time complexity. I don't use trying possibilities as I said before. It means that when I remove clue in time of generating puzzle, time for solving puzzle doesn't depends on the count of removed clues. This is the most visible for example in the beast-sized puzzles. Your program start very quickly, but then it terrible slows down (when more clues are removed). In my app all of them last about 0,1 sec.
So for example, following 40x30 puzzle were generated in less than 2 minutes.
Rated FIC+EOM-Depth:8 and SI - UnFin:59% and S - UnFin:65% by LoopDeLoop
40x30:222b2b23a32b1123b2c1a22b1a32d3a12a3a2a1a1c3a2c20a1a12b22a2h1c1b1a3a320f1b2e22b113a33a3a2c3a23a2b31a1a22322a2313a322h2312a22b2b2c212b2a212a1c20a1a1a1b3b212a3a2a1c3a3122g222a21b12a1f11a2f0a2b11a3b2b22a3a1c1e2c3b32a3a3a3c113c22c2d2a22b22a1e2f1e1b31c1231b2a23d1233b2a3b3b1a1a012c3a2d2a322a10b2d22a322a1a323d21c3a12a22a321b3a1a2a11b2c22c12a33121b3b2e3a21f0a122a32121a321a2a1a3a1c3c3b2a1a2b1c1c1d122b1i1b111a3a2b3a1e112d2a22b10a23a3c1b111a3a2a2b12a2a2a13a2a12a221a12c2e1d11a22a21f1b3d23c3b33a212c23c3c0b22c11d11c3b12b1a1a21222d3a31221213a12a3a2a3a32c22b212c2a212d2c2a2c2c23f232d22a0c33a2a2b1e11a3e10f23c12b1a2233b0a2c2212b3a132a3a30a1a22b20d12a2b32a2b32a3c2b2c233b2a21b11a1a11b2a3a1a2d2a32213e0d11c3m123a2d12c2a2b2d1a13b2b31e1b2a1d21a21132a2d22a2c123b22d3d3a2e22b212a1b212c1322b3b2113d31b2a123a3a2a2b1a1h3a3b122a131a2b21d2a23a22a2232b321a |
Tilps
Kwon-Tom Obsessive
Puzzles: 6504
Best Total: 20m 22s | Posted - 2009.11.05 07:41:29
For certain your approach is going to be faster, I was just clarifying terminology.
I suspect that a lot of the slowness of my puzzle generator is to do with failing to recognise puzzles which have multiple solutions sooner rather than later...
But I don't like programming fixed patterns. One reason is because my solver has to work on arbitrary grid shapes which makes it much harder to define patterns generically, but really they kind of strike me as wrong. |
v_e_e_n_c_a
Kwon-Tom Obsessive
Puzzles: 2071
Best Total: 32m 53s | Posted - 2009.11.05 08:06:27
These 'patterns' aren't patterns in terms like 33 -> 3 lines & 2 crosses.. It is much more closer to relations (you can look at it as a coded locks and other info). So it implies that there is no privilegued pattern (like 33).
I didn't wrote these rules manuly, this small set was generated by computer..
Last edited by v_e_e_n_c_a - 2009.11.05 08:07:46 |
Para
Kwon-Tom Obsessive
Puzzles: 1923
Best Total: 19m 28s | Posted - 2009.11.05 23:10:53
The puzzles were fun, not too hard. Especially the lack of clues in some makes it easier to figure out where to look to make progress. I didn't use uniqueness to solve them, using that would make it easier. |
mrtheman
Kwon-Tom Obsessive
Puzzles: 653
Best Total: 17m 29s | Posted - 2009.11.15 06:19:58
Hmm, some interesting reading here. A question though; do you think it is actually worth implementing all those advanced deducing rules into a program? I have also written a slither link generator/solver which does do guessing, and although it may be slower, it certainly generates puzzles much harder than the ones you posted here.
Also, with adding all these patterns, aren't you also making it think more like a human, which results in the puzzles generated being actually easier for humans to solve?
I do get that it's just good fun to write a generater that is very advanced, though I'm not sure this is the way you will be creating the best puzzles..
That's just my view though, other opinions are much appreciated.
Last edited by mrtheman - 2009.11.15 06:20:18 |
