2014 dxdy logo

Научный форум dxdy

Математика, Физика, Computer Science, Machine Learning, LaTeX, Механика и Техника, Химия,
Биология и Медицина, Экономика и Финансовая Математика, Гуманитарные науки




Начать новую тему Ответить на тему
 
 Hard Puzzles Contests
Сообщение07.08.2015, 13:12 


20/01/13
62
I would like to announce a new contests site:
http://hardpuzzles.free.fr/

The first contest is about solving Triple Klondike with a minimum number of moves.

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение19.08.2015, 09:37 
Аватара пользователя


01/06/12
1016
Adelaide, Australia
Wow cool! I only saw this by accident. Did you post it on Al Zimmerman forum?

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение25.08.2015, 23:05 


20/01/13
62
I only posted on AZPC_OffTopic, since I don't want to spam AZPC's mailing list.

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение01.09.2015, 20:29 


24/11/10
48
dimkadimon в сообщении #1046182 писал(а):
Wow cool! I only saw this by accident. Did you post it on Al Zimmerman forum?


Are you going to take part in this contest?

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение02.09.2015, 05:40 
Аватара пользователя


01/06/12
1016
Adelaide, Australia
Vitaly12 в сообщении #1049782 писал(а):
Are you going to take part in this contest?


I want to, but no time for it.

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение03.09.2015, 13:32 


24/11/10
48
dimkadimon в сообщении #1049860 писал(а):
Vitaly12 в сообщении #1049782 писал(а):
Are you going to take part in this contest?


I want to, but no time for it.


Maybe i'll give it a shot. What is different in this contest is that a theoretical best result is known.
The question is how close to it one can come.

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение03.09.2015, 15:07 


20/01/13
62
Vitaly12 в сообщении #1050126 писал(а):
Maybe i'll give it a shot. What is different in this contest is that a theoretical best result is known.
The question is how close to it one can come.


My current best result is 251 moves on deal 11.
We need 156 moves to remove all cards, and 65 moves to display all cards from the stock (at the top left).
So 251-65-156=30 moves remain.
Is it possible to explore all the positions to prove its optimality ?

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение03.09.2015, 21:42 


24/11/10
48
Цитата:
My current best result is 251 moves on deal 11.
We need 156 moves to remove all cards, and 65 moves to display all cards from the stock (at the top left).
So 251-65-156=30 moves remain.
Is it possible to explore all the positions to prove its optimality ?


Great result! Solving deals manually brought me something like 350 moves per deal. :-(
Have you used "heuristic strategy" from the paper
that you gave a link on it: "Solitaire: Man Versus Machine"?
Proving optimality looks to me is like giving it a full tree search - hardly a real option :o

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение04.09.2015, 00:00 


20/01/13
62
Vitaly12 в сообщении #1050276 писал(а):
Great result! Solving deals manually brought me something like 350 moves per deal. :-(
Have you used "heuristic strategy" from the paper
that you gave a link on it: "Solitaire: Man Versus Machine"?
Proving optimality looks to me is like giving it a full tree search - hardly a real option :o


Yes, I get around 350 moves manually too. The best human solvers get around 320 moves.

I remember reading an article describing some better heuristics. Use scholar.google.com and search for klondike or spider solitaire.

Personally, I use an heuristic, but it's completely different (and simpler too !).
I spent a lot of time tuning it though...

A full tree search doesn't seem impossible, because a lot of pruning is possible.

In fact, I still have several possible improvements, like a better moves generator. For example, I can move cards simply by swapping 2 bytes.
But right now, I'll take some rest, since I used a lot of time on this contest :roll:

 Профиль  
                  
 
 Hard Puzzles Contests
Сообщение24.09.2015, 00:08 


24/11/10
48
Прошла половина срока. Конкурс интересный и нелегкий. Жаль, что такая низкая активность :-(
Соревнуется, если я не ошибаюсь, лишь первый десяток участников...

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение24.09.2015, 17:40 


20/01/13
62
Vitaly12 в сообщении #1056136 писал(а):
Прошла половина срока. Конкурс интересный и нелегкий. Жаль, что такая низкая активность :-(
Соревнуется, если я не ошибаюсь, лишь первый десяток участников...


It's a beautiful problem, but difficult to get a working program in a few lines.

I'll add a voting system soon, so you'll be able to decide the next contest.

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение01.01.2018, 15:23 
Аватара пользователя


01/06/12
1016
Adelaide, Australia
Are there going to be more problems on this site?

 Профиль  
                  
 
 Re: Hard Puzzles Contests
Сообщение01.02.2018, 12:50 


20/01/13
62
Sorry for the slow reply.

At this moment, I'm too busy with real life (marriage and personal projects), but I hope to start a contest at the end of this year.

I'll probably launch a voting process to decide which problem will be chosen.
If you have some interesting computing ideas, I'm collecting them !

 Профиль  
                  
Показать сообщения за:  Поле сортировки  
Начать новую тему Ответить на тему  [ Сообщений: 13 ] 

Модераторы: Karan, Toucan, PAV, maxal, Супермодераторы



Кто сейчас на конференции

Сейчас этот форум просматривают: нет зарегистрированных пользователей


Вы не можете начинать темы
Вы не можете отвечать на сообщения
Вы не можете редактировать свои сообщения
Вы не можете удалять свои сообщения
Вы не можете добавлять вложения

Найти:
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group