All about flooble | fun stuff | Get a free chatterbox | Free JavaScript | Avatars    
perplexus dot info
Discussion Forums
Login: Password:Remember me: Sign up! | Forgot password

Forums > General Discussion
This is a forum for discussing anything and everything.
Bryan
2004-08-19 15:43:36
Using a computer to solve puzzles

I am always surprised and a little disappointed to see yet another person "solve" a puzzle by writing a computer program. It seems to me that, unless the puzzle is an algorithm, using a computer misses the point. Instead of solving the puzzle as stated, the programmer is instead solving a different puzzle, namely: "write a program in the language of your choice to crunch numbers using brute force."

Sure, the program changes from puzzle to puzzle, but the programmer has still managed to homogenize them all, attacking each with one overarching solution, rather than using their noggin to think in new ways.

What do others think?

nikki
2004-08-19 17:01:33
Explanations

Personally, I don't have a strong opinion about using a computer program to solve a non-algorithm problem. I think it is better if the person can explain the basic thought process they used when making their program. For example, if there were some relationships that they noticed that simplified their program, that would be interesting to hear about.

I think in general, I am not anti-computer programs, anti-spreadsheet solutions, I am simply Pro-Explanation. Some things I do find disappointing are when people go online to find the complete answer, or if they give the answer without the solution which is their thought process (ex: Whodunit). Again, this falls inline with not explaining a thought process.

It doesn't bother me if someone goes online to find a known geometric relationship, or a known equation for a partial sum, because they are just using that information as tools in their own original thought process.

Those are my thoughts =)

Gamer
2004-08-20 10:01:33
Re: Using a computer to solve puzzles

Computers are just Brute Force. So any actual solving is better and more rewarding than just cranking out the solution.

Some people might like to see the solution first, then figure out how to get there, I don't know.

nikki
2004-08-20 11:43:27
Define Brute Force

I always think of the definition of Brute Force as being "I will blindly try every scenario until I find one that solves the problem."

But if a person goes "Well, I can narrow it down to one of these possibilities (that is much less than all the possibilities) using the explained logic, and then I can go through each of them to see which works," is that Brute Force? I don't think so, but others might.

If someone used logic to narrow down the possibilities (and explained it) and then used a computer to check each of the remaining possibilities, is that Brute Force? Is that an acceptable use of a computer program.

I probably sound like I'm trying to keep computer programs in =) I actually haven't used one to solve any Flooble problems myself (though I've used Excel), so I am not defending computer solutions out of personal habits =)

Gamer
2004-08-20 13:37:54
Re: Using a computer to solve puzzles

I think the way I define brute force is if you do it (by hand, not by computer) it should take less time than to figure out with explanation. For example, if you have 3 possibilities left, it may take less time to just check 2 rather than explain them away. If you have 300, it isn't going to be quick to check all of them though.

As for using brute force when it's not quicker, the less the better.

SilverKnight
2004-08-20 13:56:49
Re: Using a computer to solve puzzles

I would agree with Gamer's description... but it is imprecise in that the definition is a matter of degree. How many possibilities do you have to have... or how much time must one spend before it is "brute force"?

And if someone is faster than another "by hand", then is there a different cut-off before it becomes "brute force".

Allow me to offer another description:
Brute force: strictly defining the "solution domain" (perhaps after narrowing it down somewhat) and iterating through all the points therein, evaluating each as a possible solution.

As for that being "good" or "bad" well... that's an opinion, a value judgement.
_______________________________________________________

To address Bryan's initial point:

Sometimes the best problems are those that don't lend themselves to brute force. Maybe that's why some geometry and probability problems seem to be well received on Perplexus. Of course, I've seen a few algorithm/computer only problems that were well received as well.

nikki
2004-08-20 15:28:20
Number Pyramid

Ok, so would my solution to Number Pyramid count as brute force? It sounds like it might.

Brian Smith
2004-08-20 22:51:48
Re: Using a computer to solve puzzles

If someone used logic to narrow down the possibilities (and explained it) and then used a computer to check each of the remaining possibilities, is that Brute Force? Is that an acceptable use of a computer program.

Actually this scenario happens in professional mathematics. For example, the proof for the Four Color Theorem used a computer algorithm to check a few hundred discreet cases.

nikki
2004-08-25 10:43:38
Number Pyramid

I'm still wondering if you guys would count my solution to Number Pyramid as using Brute Force.

I don't think it does, but that might because I have a slightly different definition of Brute Force than you guys.

Cory Taylor
2004-08-25 12:01:46
Re: Using a computer to solve puzzles

Without checking the validity of the solution, the method seems logical, rather than brute force-ical. In this question, a brute force solution would have checked hundreds of possibilities, your method reduced the set of possibilities substantially before any outcome had to be tested.

Bruce Brantley
2004-08-28 14:43:57
Re: Using a computer to solve puzzles

Back to the question of using a computer. I think people like puzzles because they are fun. I prefer to think to a logical conclusion or try to approach from different angles to come to a solution. But what if I enjoyed writing computer programs? What if the challenge for me is to figure out how to write a program that will solve the puzzle? I don't see anything wrong with it. That is the purpose of the puzzle anyway. I say to each his or her own. I don't read other peoples' postings anyway until I have either come to a conclusion or determined that I can't come to a conclusion. So it's not like a spoiler. Keep em coming Charlie. I can't understand them, but they look cool scrolling down the screen.

Mike Graham
2004-09-06 23:19:48
Re: Using a computer to solve puzzles

Brian Smith said:
Actually this scenario happens in professional mathematics. For example, the proof for the Four Color Theorem used a computer algorithm to check a few hundred discreet cases.

I'd be shocked if topologists ever got away with that... a ew hundred cases does not show anything valuable in theoretical math. "It's intuitive, it works a few hundred times, and no one has provided a counterexample" is not a proof.

SilverKnight
2004-09-07 01:15:26
Re: Using a computer to solve puzzles

No Mike...

What Brian said is accurate (except that he used the homonym discreet instead of discrete). And the reason why it worked was that they were able to show that ALL POSSIBLE scenarios were represented by a few hundred discrete cases. Or, in other words, they were able to show that any possible scenario was equivalent to one of the discrete cases at issue.

Therefore, once this had been established, one needed only to iterate (brute force) through those discrete cases to complete the proof.

SilverKnight
2004-09-07 01:16:37
Re: Using a computer to solve puzzles

(Perhaps I should have used "homophone" rather than "homonym". ;-)

Mike Graham
2004-09-07 13:03:11
Re: Using a computer to solve puzzles

Thank you, I knew that, I feel like an idiot now.

Copyright © 2002 - 2024 by Animus Pactum Consulting. All rights reserved. Privacy Information