Talk:Knight's tour: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
Line 1:
Wow, this was found fast. I was still prepping my first couple implementations.:)
--[[User:Markjreed|Markjreed]] 02:09, 30 May 2011 (UTC)
Added my original perl solution and sample output; moved out of draft status.
--[[User:Markjreed|Markjreed]] 02:48, 30 May 2011 (UTC)
== References ==
Line 45:
:I didn't really have time to measure its effectiveness though. --[[User:Paddy3118|Paddy3118]] 13:27, 30 May 2011 (UTC)
:The Perl solution is deterministic, and produces complete tours for all 64 start squares. So it must be something in the ordering of the move list. --[[User:Markjreed|Markjreed]] 01:54, 31 May 2011 (UTC)
:Try a 7x7 board. --[[User:Paddy3118|Paddy3118]] 03:24, 31 May 2011 (UTC)
:: I will. At this point there is something wrong with this solution. I will have to come back to it later (possibly this weekend). This was implemented from the WP description which talks about sets. I think the heart of the problem lies in the used of unordered sets everywhere and I will have to walk through some comparative examples side by side. This wouldn't be the first time I've been burned by a vague WP algorithm description. I'll have to borrow one of the other solutions and add code to show me the order squares get presented. --[[User:Dgamey|Dgamey]] 09:48, 31 May 2011 (UTC)
:: Confirmed that I sometimes get incomplete tours on 7x7 with Warnsdorff using the move list order in the Perl code. Leaving the any-size-board generalization out of code on page until I handle that case. --[[User:Markjreed|Markjreed]] 12:47, 31 May 2011 (UTC)
== C++ formatting ==
| |||