Puzzle Description:

This is the (almost certainly) final iteration of the sierpinski triangle iterations. You have to use the others to solve it!!!

#1: Joe (infrapinklizzard) on Mar 25, 2011

Iteration #1 http://webpbn.com/index.cgi?id=12529
Iteration #2 http://webpbn.com/index.cgi?id=12531
Iteration #3 http://webpbn.com/index.cgi?id=12532
Iteration #4 http://webpbn.com/index.cgi?id=12534

I truly enjoyed this series--but I'm glad it is the last. I fear I would feel compelled to solve iteration 6 if it existed. This looks amazingly like the Sierpinski Triangle (http://www.zeuscat.com/andrew/chaos/sierpinski.html )someone sent us to in a former iteration.

So one of my side hobbies is studying programs that solve paint-by-number puzzles. I have a page where I test different solvers on puzzles from the webpbn database. (See http://webpbn.com/survey/)

This puzzle is the new champion hardest puzzle in the data set. I normally allow each solver half an hour to try to solve a puzzle. None of them were able to solve this. I'm going to try letting one of them run much longer to see if it ever gets a solution.

Nonogram solving is one of many standard problems that AI researchers try to solve as a benchmark of the effectiveness of their methods, so my survey page has been getting occasional references in academic journals, and I'm in occasional communication with various researchers and programmers. It would be nice if the interesting example problems that I reference in that survey could be freely redistributed by researchers, maybe be printed in papers and so forth, but I can't generally let people do that because of the copyright restrictions. I wondering if Joshua would be willing to give permission for this puzzle to be so redistributed, under the stipulation that the attribution to him and copyright by him be retained? If you give such permission it will be so noted in the survey. Thanks.

My solver has been grinding on this for 48 hours without success. Pretty impressively bad.

Found to have multiple solutions by jan.

Someone finally built a solver that could solve this. It has multiple solutions.

Jan YES! Wow I'm late but of course you can use it and distribute an what not! At this point though, I fear it's too late haha.

Unfortunately, Jan won't be able to reply. :(

Goto next topic

You must register and log in to be able to participate in this discussion.