Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.forth > #10251

Re: Sumbrero puzzle solver

From mhx@iae.nl (Marcel Hendrix)
Subject Re: Sumbrero puzzle solver
Newsgroups comp.lang.forth
Message-ID <58671515008435@frunobulax.edu> (permalink)
Date 2012-03-20 20:36 +0200
References <4f66df09$0$6622$9b4e6d93@newsspool2.arcor-online.net>
Organization Wanadoo

Show all headers | View raw


"A. K." <akk@nospam.org> writes Re: Sumbrero puzzle solver

> On 18.03.2012 23:00, Marcel Hendrix wrote:
>> David Kuehling<dvdkhlng@gmx.de>  writes Re: Sumbrero puzzle solver

>>>>>>>> "Marcel" == Marcel Hendrix<mhx@iae.nl>  writes:

>>> David Kuehling<dvdkhlng@gmx.de>  writes Re: Sumbrero puzzle solver
>>>>>>>>>> "Marcel" == Marcel Hendrix<mhx@iae.nl>  writes:

>>>>>> Here is a general solution to the so-called Sumbrero puzzle.  Some
>>>>>> examples can be found on http://www.omegajunior.net/code/sumbrero/.
[..]
>>>> There are only about five million possibilities to test for.  I found
>>>> it takes a maximum of 20 ms to find the hardest solution (1 million
>>>> tries). There is no reason to go for a better algorithm here.

>>> Not unless you want to solve many, many (say 10^8) puzzles :)

>>> Although you're right, I find it somewhat unsatisfying that a computer
>>> needs a million tries for a problem easily solved by a human.

> It does NOT !!

What doesn't? Is this reaction aimed at David or to me?

I have published solutions to two kinds of sumbrero's now. Both reduce
the problem much more than a constraint solver would/can, then solve
the constraints rather clumsily. Whatever, I am quite sure the solution
is much (400 times or more?) faster than, say, minion can do it.

Even so, I think it is no shame to use brute force for a problem that
yields to it, It gives much faster development than first writing a 
Prolog engine.

> Reducing the solution space is a core topic in constraint programming.
> see f.ex.
> http://kti.ms.mff.cuni.cz/~bartak/constraints/index.html

I learned a few interesting things there. Unfortunately/amazingly these
codes all seem to be proprietary.

> How do you think that your car navigation system works? Or DNA 
> sequencing? Or flight schedulers for airports? Absolutely impossible by 
> brute force.

I am sure our brains didn't sit around while Bartak developed their 
foundation.

> For further playing, the other one should take a bit longer to run:
> http://www.omegajunior.net/code/sumbrero/sixbythree.html

I have not yet found a sixbythree in a newspaper. At some point the
algorithm should degenerate.

-marcel

Back to comp.lang.forth | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

Sumbrero puzzle solver mhx@iae.nl (Marcel Hendrix) - 2012-03-10 18:24 +0200
  Re: Sumbrero puzzle solver David Kuehling <dvdkhlng@gmx.de> - 2012-03-11 00:30 +0100
    Re: Sumbrero puzzle solver mhx@iae.nl (Marcel Hendrix) - 2012-03-11 01:25 +0200
      Re: Sumbrero puzzle solver David Kuehling <dvdkhlng@gmx.de> - 2012-03-11 23:27 +0100
        Re: Sumbrero puzzle solver mhx@iae.nl (Marcel Hendrix) - 2012-03-19 00:00 +0200
          Re: Sumbrero puzzle solver "A. K." <akk@nospam.org> - 2012-03-19 08:24 +0100
            Re: Sumbrero puzzle solver mhx@iae.nl (Marcel Hendrix) - 2012-03-20 20:36 +0200
              Re: Sumbrero puzzle solver mhx@iae.nl (Marcel Hendrix) - 2012-03-20 23:12 +0200
              Re: Sumbrero puzzle solver "A. K." <akk@nospam.org> - 2012-03-21 12:06 +0100
  Re: Sumbrero puzzle solver "A. K." <minforth@arcor.de> - 2012-03-11 12:10 +0100
    Re: Sumbrero puzzle solver "A. K." <akk@nospam.org> - 2012-03-12 21:51 +0100
  Re: Sumbrero puzzle solver Albert van der Horst <albert@spenarnc.xs4all.nl> - 2012-03-11 19:22 +0000
    Re: Sumbrero puzzle solver David Kuehling <dvdkhlng@gmx.de> - 2012-03-11 22:35 +0100
    Re: Sumbrero puzzle solver C G Montgomery <cgm@physics.utoledo.edu> - 2012-03-11 18:23 -0400

csiph-web