Talk:Solve a Hidato puzzle: Difference between revisions
→Any good general algorithm?
(→Any good general algorithm?: c'est la vie) |
|||
Line 73:
I added a larger example in the C code, which is meant to make brute force search miserable. The C code can solve it, but it takes quite long; I expect D to behave similarly. The Perl/Python/Tcl code should be able to solve it, if only taking forever, since they all use the same exhaustive search method (Tcl is forced into such a situation, but anyway). The Mathprog code gives up on it on my machine, rather quickly, which can be said to be the bright side.
:You do not indicate how glpk gives up on your machine, I am guessing numerical instability. Evil Case 2 makes the problem much larger. I have added examples of using Mathprog suitable for this large example. Note the model is the same, I just used a basis recommended for large multi indexed problems with a nested structure. In all three cases the solution then takes less than 1 second.
Note that that example is trivial for a human. Does anyone have any clever way to deal with such cases? --[[User:Ledrug|Ledrug]] 00:59, 3 May 2012 (UTC)
| |||