Re: ...Contemplating unspecified constraint SOLVERS.
- From: "Dirk Mittler" <mdirk@xxxxxxxxxxxx>
- Date: Sat, 27 Aug 2005 10:44:57 -0400
It was never my intention to complain. {:-)} And I thank you for pointing
out that in constraint solvers, "labelling" is an important step. I will
look up the subject in more detail then, when I find the time. For now I
*imagine* that labelling could satisfy this need of assigning to the
attributed variables, after constraints have been specified. This could be
why in one Prolog example which I've seen, the 'label' command is given last
over several variables to tell the compiler to solve. It will intrigue me
then to learn how this command really works.
Dirk
"A.L." <alewando@xxxxxxxxxx> wrote in message
news:l5t0h158k927v29o1asgv3rni43eei9v1r@xxxxxxxxxx
> On Fri, 26 Aug 2005 22:29:40 -0400, "Dirk Mittler"
> <mdirk@xxxxxxxxxxxx> wrote:
>
>>
>>"A.L." <alewando_won_@xxxxxxxxxxxxxxx> wrote in message
>>news:9v7pg1pr95s0q0dh6nl95fpqm70ra2fkpb@xxxxxxxxxx
>>
>>>
>>> Could you please summarize in one SHORT paragraph what is your
>>> point?...
>>>
>>> A.L.
>>
>>You are focussing on the fact that an earlier exercise of mine, which
>>people
>>would have a hard time even to find on the bulletin board by now, was in
>>fact not functional, and that yet I spent time on it. But I believe that
>>one
>>would need to understand precisely the text which you have omitted here,
>>in
>>order to begin to write one's own constraint *solvers* . As opposed to
>>just
>>using a CLP(R) solving system that came with the Prolog compiler.
>
> It is not one paragraph, and this is not short. In addition, despite
> reading your post multiple times, I have no idea what is your point.
>
> As I understand (and I may be wrong...) one is that Bratko doesn't
> explain how to write your own constraint solver. This is not a
> rocket science to write simple (probably not efficient) constraint
> solver. Taka a book about constraint programming (Krzysztof Apt
> dicusses this in his book in details, although the code he provides
> is not in Prolog) or implement Prolog constraint solver that is
> posted on Roman Bartak's page:
>
> http://kti.ms.mff.cuni.cz/~bartak/html/constraints.html
>
> See the section
>
> "CSP (Constraint Satisfaction Problems) (1996)
>
> The program implements a skeleton for labeling, the main part of CSP
> solvers. You need to include a plug-in module containing a
> constraint solver to get a specific labeling procedure. Two examples
> of plug-in modules are enclosed.
>
> PROLOG source code of labeling kernel
> plug-in module for constraints over integers (backtracking)
> plug-in module for constraints over integers (forward checking)
> Word Puzzle Solver
>
> This is a program, based on CSP, for solving word puzzles. Example
> of usage is enclosed.
>
> PROLOG source code of Word Puzzle Solver
> word puzzle example.."
>
> In addition, there are a lot of publications about implementing
> constraint solvers in Prolog or extending Prolog for constraint
> solving. See papers by Daniel Diaz who did this for Gnu Prolog. Go
> to SICS page and look for SICStus Prolog. There are publications by
> Mats Carlsson who implemented constraint solver for SICStus Prolog.
>
> Use google. Stop complaining.
>
> A.L.
.
- Follow-Ups:
- Re: ...Contemplating unspecified constraint SOLVERS.
- From: A . L .
- Re: ...Contemplating unspecified constraint SOLVERS.
- References:
- Re: I now realize WHY my 'consistency check' was pointless: CLP.
- From: A . L .
- Re: ...Contemplating unspecified constraint SOLVERS.
- From: Dirk Mittler
- Re: ...Contemplating unspecified constraint SOLVERS.
- From: A . L .
- Re: I now realize WHY my 'consistency check' was pointless: CLP.
- Prev by Date: Starting with Prolog
- Next by Date: Native (not cygwin) build of Mercury for WinDoze
- Previous by thread: Re: ...Contemplating unspecified constraint SOLVERS.
- Next by thread: Re: ...Contemplating unspecified constraint SOLVERS.
- Index(es):
Relevant Pages
|