Talk:Solve a Hidato puzzle: Difference between revisions

m
 
(16 intermediate revisions by 4 users not shown)
Line 223:
Putting g6 into the model and resolving again solves the problem by hand.
 
It is apparent that the number of equations increases rapidly with probemproblem size. A method exists which makes this increase linear see http://en.wikipedia.org/wiki/Tseitin-Transformation.
 
--[[User:Nigel Galloway|Nigel Galloway]] ([[User talk:Nigel Galloway|talk]]) 14:12, 30 December 2013 (UTC)
 
: if a1 = 0, the equation b2 + c2 + d2 = a1 means "Non of a's neighbors is 2" which is not necessarily true. so this solution does not work. --[[User:Ak|Ak]] ([[User talk:Ak|talk]]) 17:10, 21 March 2014 (UTC)
 
==extra credit suggestion==
 
How about adding an extra credit for this task: support Numbrix puzzles as well as Hidato puzzles. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 04:50, 18 December 2013 (UTC)
 
The REXX programming example only needed a couple of '''if''' statements to solve Numbrix puzzles in addition to Hidato puzzles. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 04:50, 18 December 2013 (UTC)
 
::Two points. First I think it would be better to add a related task Solve Numbrix Puzzles. I know what Numbrix is but only from external servlets and the press. Explaining it to them what be's satisfaction would further complicate this task. Secondly I had an extra credit for showing that the same logic could solve the Knight's Tour, but them what be made me remove it. The Knight's Tour is exactly the same as Hidato, only the Knight has a different view of which cells are adjacent. Numbrix has only 4 adjacent cells whereas Knight's Tour and Hidato have 8.--[[User:Nigel Galloway|Nigel Galloway]] ([[User talk:Nigel Galloway|talk]]) 13:35, 30 December 2013 (UTC)
 
::: Numbrix puzzles (in my thinking) are just a simpler Hidato puzzle, in the manner of having less possible moves (i.e., being more restrictive).   But I have no qualms of someone creating a Rosetta Code task just for Numbrix puzzles. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 16:20, 30 December 2013 (UTC)
 
::: As regarding the Knights Tour: in fact, I used the same code in the REXX example, I just tweaked/optimized the code a bit for speed. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 16:20, 30 December 2013 (UTC)
 
== solved it using constraint programming ==
 
Behold: http://hidoku-solver.appspot.com/
Anonymous user