ENG  RUSTimus Online Judge
Online Judge
Problems
Authors
Online contests
About Online Judge
Frequently asked questions
Site news
Webboard
Links
Problem set
Submit solution
Judge status
Guide
Register
Update your info
Authors ranklist
Current contest
Scheduled contests
Past contests
Rules
back to board

Discussion of Problem 1589. Sokoban

Problem 1589 "Sokoban". New tests added
Posted by Vladimir Yakovlev (USU) 14 Jul 2011 15:18
The problem is rejudged, one author have lost AC. Thanks to Erop [USU]
Re: Problem 1589 "Sokoban". New tests added
Posted by Dest 27 Aug 2011 20:32
Yeah, thanks to Erop [USU] :).
I wish to add some tests too.
I’ve sent them to "timus_support@acm.timus.ru". Did you receive them?
Re: Problem 1589 "Sokoban". New tests added
Posted by Oracle[Lviv NU] 5 Sep 2011 01:05
If it is possible, can I see test #92 (or some similar)? I've spent a lot of time getting "Accepted", but still no idea, what kind of test is it.
Thanks.
P.S. my mail: Oracle@acm.lviv.ua

Edited by author 05.09.2011 11:56

Edited by author 05.09.2011 11:56
Re: Problem 1589 "Sokoban". New tests added
Posted by Dest 5 Sep 2011 13:12
What do you mean under "kind of test"? Walls, boxes... as usually.
I think that it's very small to have a "kind".
It has about 110 pushes in push-optimal solution.
But, seems that pushes not a primary factor of hardiness.
Here is the test (very easy) that has 100 pushes in push-optimal solution:
########
#._.*__#
##$_*__#
##__*_*#
##_*__.#
#_$$**$#
#@.____#
########

I suppose that it's hard because of huge amount of states passes through without pruning, and estimation function can't do anything in first 2/3 of search tree because of boxes are placed at goals and removed from them many times.

Here is another test (hard. But, maybe, your solver cracks it instantly) with only 73 pushes:
########
#._.@__#
#$$$*$*#
#._$___#
#.$.*_*#
#_$.$__#
#._.*__#
########
Re: Problem 1589 "Sokoban". New tests added
Posted by Oracle[Lviv NU] 6 Sep 2011 01:26
About kind: some boxes should be moved from one "room" to another, a lot of boxes on board, etc.

Really, first test is quite easy (about 2 seconds for my algo on my laptop)... And second is real hell))) On my laptop it takes about 10 seconds to solve it. And solution is found only at the end of the search (all of about 800000 possible states should be tested).
My solution do not give push-optimal solution nor move-optimal solution (and I do not use any estimation function at all), but nevertheless it is also hard for my algo too.

Thanks a lot!