Newsgroups: comp.lang.prolog
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!portc01.blue.aol.com!newsxfer2.itd.umich.edu!howland.erols.net!surfnet.nl!swidir.switch.ch!in2p3.fr!univ-lyon1.fr!pasteur.fr!jussieu.fr!lcr.thomson-csf.fr!news.thomson-lcr.fr!news
From: Eric Jacopin <jacopin@thomson-lcr.fr>
Subject: Re: AI Planning with partially instantiated operators
X-Nntp-Posting-Host: mars
Content-Type: text/plain; charset=us-ascii
To: Jeffery <me505@umist.co.uk>
Message-ID: <3247D49E.794BDF32@thomson-lcr.fr>
Sender: news@news.thomson-lcr.fr
Content-Transfer-Encoding: 7bit
Cc: jacopin@laforia.ibp.fr
Organization: Thomson-CSF, Laboratoire Central de Recherches, Orsay, France
References: <R.32443CD0.3A3A@umist.co.uk>
Mime-Version: 1.0
Date: Tue, 24 Sep 1996 12:31:26 GMT
X-Mailer: Mozilla 2.02 (X11; I; SunOS 4.1.1-IP sun4c)
Lines: 40

Jeffery wrote:
> 
> Reposting article removed by rogue canceller.
> 
> I am working on an Artificial intelligence planning problem requiring
> partially instantiated operators.
> 
> My problem is how do I record the binding constraints. I am following
> the algorithm outline as presented in Weld's paper Introduction to Least
> Commitment Planning. Also need advice on the implementation of varsets.
> 
> Jeffrey
> email me505@ist.co.umist.ac.uk

1) the ftp:// site (mosaic or netscape 3.0) is an AI planning web page.
   There, you'll find a prolog (sources) planner (but designed for the 
   mac --- Open Prolog).

2) If you speak about bindings constraints, it's because you diving into
   unification of a precondition with a postcondition. You may not wish to 
   do this in Prolog and, instead, use Prolog's unification mechanism.

3) Though using Prolog, you may still wish to use binding constraints.
   You then just need to =.. a predicate (i.e. a precondition) to get the 
   variables from it, then check whether you got them already, then record
   trivial neq constraints in a list, such as [neq(X,Y),neq(Y,I)]. Then,
   you got to decide when to go through this list to check that indeed, X is
   not equal (neq) to Y and Y is not equal to I. If by any chance X is equal to 
   Y, then, trivially force you planner to backtrack.

4) You may wish to contact Sam Steel (sam@essex.ac.uk). One of his student
   reimplement UCPOP in prolog.

   Have planned fun!

--
Eric Jacopin              Thomson-CSF/Laboratoire Central de Recherche
tel +33 1 69 33 92 90     Domaine de Corbeville, 91404 ORSAY CEDEX
fax +33 1 69 33 08 65     FRANCE
ftp://ftp.ibp.fr/ibp/softs/laforia/pweak/Pweak.html
