View unanswered posts | View active topics It is currently Tue Mar 19, 2024 6:29 am



← Back to the Calcudoku puzzle page




Reply to topic  [ 15 posts ]  Go to page Previous  1, 2
 Using 'Unique Solution' to help solve 
Author Message

Posted on: Mon May 13, 2013 12:07 am




Posts: 106
Joined: Sat May 14, 2011 1:08 am
Post Re: Using 'Unique Solution' to help solve
mce wrote:
And it's even impossible not to use this if one ends up spotting that a certain combination of numbers cannot be right as otherwise the solution would not be unique. It's not something one can "unlearn on demand".


Huh? Just because you see that you could eliminate some combination of numbers because then the solution wouldn't be unique doesn't mean that you have to use that information in trying to find your solution. In fact if you start with the premise that the solution might not be unique you will ignore this sort of thing.


Profile

Posted on: Tue May 14, 2013 5:33 am




Posts: 4
Joined: Fri May 10, 2013 7:39 am
Post Re: Using 'Unique Solution' to help solve
mce wrote:
tommyp0123 wrote:
picklepep wrote:
I disagree. The unique solution is a criteria and a stricture to these puzzles. It is a given much like that of there only being one iteration of a number in a row or column. I don't believe in self-imposed limits when we are solving these puzzles. To do so would limit my ability to sharpen my skill, or gain insight in the way the numbers work together.

I can think of a few puzzles where if I hadn't used this method the only alternative would have been the old 'trial and error'. Now if I had an alternative between using two methods, one of which was plugging in numbers, and the other which required me to think, I'm going to use the one that sharpens my acumen.


I respect that. It's been my experience that I have always found a way to proceed using technique and not brute-force.


I would not call "making use of the uniqueness of the solution" brute force. It's just a piece of info one has from the start. And it's even impossible not to use this if one ends up spotting that a certain combination of numbers cannot be right as otherwise the solution would not be unique. It's not something one can "unlearn on demand".


I interpreted trial and error to mean brute-force, i.e. plugging in numbers. It's true you can't 'unlearn' but you can prove for squares (isn't that what solving is - proving that a certain square must be a certain number) without using that information. It may point you in a particular direction though.


Profile

Posted on: Wed May 15, 2013 9:37 am




Posts: 84
Joined: Fri May 13, 2011 9:37 am
Post Re: Using 'Unique Solution' to help solve
In my opinion an unique solution is a matter of trust. My system does some simple logic and then uses trial and error / backtracking. The first solution found always happens to be the good one.
As this is not satisfying, I made a simple module that evaluates all candidates. No-good candidates are sorted out. If then a solution is found, it is certainly the only one.
Btw, for the more difficult puzzles, this module reduces the time needed for trial and error.


Profile

Posted on: Sat May 18, 2013 1:47 pm




Posts: 13
Joined: Sun Mar 04, 2012 6:46 pm
Post Re: Using 'Unique Solution' to help solve
mparisi wrote:
mce wrote:
And it's even impossible not to use this if one ends up spotting that a certain combination of numbers cannot be right as otherwise the solution would not be unique. It's not something one can "unlearn on demand".


Huh? Just because you see that you could eliminate some combination of numbers because then the solution wouldn't be unique doesn't mean that you have to use that information in trying to find your solution. In fact if you start with the premise that the solution might not be unique you will ignore this sort of thing.


Of course one does not "absolutely have to use the info", as there is plenty of structure in these puzzles and there always is another way to conclude. But as I wrote, you cannot "unlearn" it either and I see no fun in spending a lot of time in finding an alternative explanation for what I already know. I prefer to discover things that I do not yet know. For those who disagree, here's a new challenge: take a 12x12 and describe every possible path to the solution according to any set of rules you like. Just don't forget to eat, sleep, and live in between. [wink]

I also don't see what's wrong with using it anyway. After all, we all also use the fact that any number can only occur once in a row/column. It's a given. So is the uniqueness.


Profile
User avatar

Posted on: Sat May 18, 2013 2:30 pm




Posts: 3295
Joined: Thu May 12, 2011 11:58 pm
Post Re: Using 'Unique Solution' to help solve
And very related, the "no-op" puzzles: viewtopic.php?f=16&t=440

E.g. the cage with "2" and 2 cells has to be 2-:
2+ obviously doesn't work, and 2x and 2: don't
work because the solution is unique (also wrt.
the operators (!))


Profile
Display posts from previous:  Sort by  
Reply to topic   [ 15 posts ]  Go to page Previous  1, 2

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
cron
All forum contents © Patrick Min, and by the post authors.

Forum software phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by STSoftware.