first trial

回答: 嗬嗬,downloaded explaner, it's crunching.....品茗992011-10-23 14:35:44

Tried the first one. The way it reasons seems to confirm my reasoning. It tries some hypothesis and tries to reject it by leading to a conflict, therefore narrowing the possible choices.

If this is the only way, then it's not much different from trial and error. At most, it's more systematic and heuristic, although it classifies the rules, for example, contradiction forcing chain, region forcing chain, double forcing chain, etc.

It's impossible for a person to solve in this way even he knows the logic. It's probably better using direct trial and error, for example, with the help of a tool, like the sudoku helper or my program. In that way, you can start with the position with the least number of possibilities. It eventually becomes a search on a graph. For the example sudoku, the least possibilities is 3.

所有跟帖: 

回複:first trial -hqw2000- 給 hqw2000 發送悄悄話 (1268 bytes) () 10/23/2011 postreply 20:17:43

sort of -品茗99- 給 品茗99 發送悄悄話 (1370 bytes) () 10/24/2011 postreply 18:05:02

In case you are interested -hqw2000- 給 hqw2000 發送悄悄話 (680 bytes) () 10/24/2011 postreply 18:23:06

Very nice work! -品茗99- 給 品茗99 發送悄悄話 (306 bytes) () 10/26/2011 postreply 17:41:30

謝謝,不想證明什麽了。估計改用c++實現的話,可以快一百倍。 -hqw2000- 給 hqw2000 發送悄悄話 (0 bytes) () 10/28/2011 postreply 18:50:59

您的軟件放在網上了嗎。用什麽寫的。 -wxcfan123- 給 wxcfan123 發送悄悄話 (0 bytes) () 10/29/2011 postreply 16:21:41

用python 實現的,不打算放在網上。 -hqw2000- 給 hqw2000 發送悄悄話 (281 bytes) () 10/29/2011 postreply 17:13:27

請您先登陸,再發跟帖!