Re: Powerloom, a first stab...

Gerard Toonstra <[email protected]> Fri, 26 Mar 2010 08:51:00 +0000
Newsgroups gmane.comp.ai.powerloom
Message-ID <[email protected]>
--===============1944975865==
Content-Type: multipart/alternative;
	boundary="_ec82223d-ab6a-404a-a84f-fe6e2b8e83d5_"

--_ec82223d-ab6a-404a-a84f-fe6e2b8e83d5_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable



Thanks Hans=2C this has certainly been helpful. =20

The justifications example certainly looks very complex to me and your seco=
nd point made
me realize I'd be better off maintaining my own tree in the host code. Then=
 I can choose which=20
branches to explore based on a number of different algorithms (breadth-firs=
t=2C iterative deepening=2C best-first=2C etc.).=20
I think I'll go down that route for now at least.


Before looking at powerloom=2C I looked at JSHOP2. JSHOP2 can solve plannin=
g problems=2C but it requires compilation
of the KB and the language is certainly not as rich as powerloom's for the =
expression of particular relations and so forth.
Is powerloom well suited for planning problems in general? I haven't found =
any demos or tutorials that
show how it's used for those kinds of problems.

Rgds=2C

Gerard


> From: [email protected]
> Date: Thu=2C 25 Mar 2010 13:18:24 -0800
> To: [email protected]
> CC: [email protected]
> Subject: Re: Powerloom=2C a first stab...

> (2) minimality: you can do a "retrieve all" and PowerLoom will try to
> generate all possible solutions.  If you have some kind of cost
> function=2C you'd compute that cost for each solution and then select
> the one with the smallest one (e.g.=2C make the cost an output argument
> and do a :sort-by :values in the retrieve).  However=2C if there is a
> potentially large number of solutions=2C you might want to do something
> smarter than that by cutting off partial solutions that can't improve
> on what you already have.  That=2C however=2C you would have to do in
> code.  You might also have more complex subsumption relationships
> between solutions which might require a somewhat more sophisticated
> approach.
>=20
> Hope that helps=2C
>=20
> Hans
>=20
> >>>>> Gerard Toonstra <[email protected]> writes:
>=20
> > Hi there=2C
>=20
> > I'm currently evaluating powerloom for research in the context of an ag=
ent platform.
>=20
> > I've read through all documentation of powerloom so far and I think it'=
s certainly powerful enough
> > to use it for our purposes=2C certainly for all the standard logic.=20
> > I do have a question however on producing solutions and alternatives. T=
he only recursion=20
> > example I found in Powerloom used transitivity rules (parent/ancestor e=
xample and the RCC8 KB).
> > Unfortunately=2C my problem cannot be used with transitivity.
>=20
> > The problem I have is:
>=20
> > So that's the question basically. Can powerloom produce a list of entit=
ies that together form the solution=2C
> > based on some constraints that are set on the production of this list a=
nd present each acceptable=2C but minimal=20
> > solution individually?  I'm not asking for a piece of code=2C just what=
 I need to look for to realize this.
>=20
> > Rgds=2C
>=20
> G>=20
>  		 	   		 =20
> > _________________________________________________________________
> > Hotmail is redefining busy with tools for the New Busy. Get more from y=
our inbox.
> > http://www.windowslive.com/campaign/thenewbusy?ocid=3DPID27925::T:WLMTA=
GL:ON:WL:en-US:WM_HMP:032010_2
 		 	   		 =20
_________________________________________________________________
The New Busy is not the old busy. Search=2C chat and e-mail from your inbox=
.
http://www.windowslive.com/campaign/thenewbusy?ocid=3DPID27925::T:WLMTAGL:O=
N:WL:en-US:WM_HMP:032010_3=

--_ec82223d-ab6a-404a-a84f-fe6e2b8e83d5_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<style><!--
.hmmessage P
{
margin:0px=3B
padding:0px
}
body.hmmessage
{
font-size: 10pt=3B
font-family:Verdana
}
--></style>
</head>
<body class=3D'hmmessage'>
<br>Thanks Hans=2C this has certainly been helpful.&nbsp=3B <br><br>The jus=
tifications example certainly looks very complex to me and your second poin=
t made<br>me realize I'd be better off maintaining my own tree in the host =
code. Then I can choose which <br>branches to explore based on a number of =
different algorithms (breadth-first=2C iterative deepening=2C best-first=2C=
 etc.). <br>I think I'll go down that route for now at least.<br><br><br>Be=
fore looking at powerloom=2C I looked at JSHOP2. JSHOP2 can solve planning =
problems=2C but it requires compilation<br>of the KB and the language is ce=
rtainly not as rich as powerloom's for the expression of particular relatio=
ns and so forth.<br>Is powerloom well suited for planning problems in gener=
al? I haven't found any demos or tutorials that<br>show how it's used for t=
hose kinds of problems.<br><br>Rgds=2C<br><br>Gerard<br><br><br>&gt=3B From=
: [email protected]<br>&gt=3B Date: Thu=2C 25 Mar 2010 13:18:24 -0800<br>&gt=3B =
To: [email protected]<br>&gt=3B CC: [email protected]<br>&gt=3B Su=
bject: Re: Powerloom=2C a first stab...<br><br>&gt=3B (2) minimality: you c=
an do a "retrieve all" and PowerLoom will try to<br>&gt=3B generate all pos=
sible solutions.  If you have some kind of cost<br>&gt=3B function=2C you'd=
 compute that cost for each solution and then select<br>&gt=3B the one with=
 the smallest one (e.g.=2C make the cost an output argument<br>&gt=3B and d=
o a :sort-by :values in the retrieve).  However=2C if there is a<br>&gt=3B =
potentially large number of solutions=2C you might want to do something<br>=
&gt=3B smarter than that by cutting off partial solutions that can't improv=
e<br>&gt=3B on what you already have.  That=2C however=2C you would have to=
 do in<br>&gt=3B code.  You might also have more complex subsumption relati=
onships<br>&gt=3B between solutions which might require a somewhat more sop=
histicated<br>&gt=3B approach.<br>&gt=3B <br>&gt=3B Hope that helps=2C<br>&=
gt=3B <br>&gt=3B Hans<br>&gt=3B <br>&gt=3B &gt=3B&gt=3B&gt=3B&gt=3B&gt=3B G=
erard Toonstra &[email protected]&gt=3B writes:<br>&gt=3B <br>&gt=
=3B &gt=3B Hi there=2C<br>&gt=3B <br>&gt=3B &gt=3B I'm currently evaluating=
 powerloom for research in the context of an agent platform.<br>&gt=3B <br>=
&gt=3B &gt=3B I've read through all documentation of powerloom so far and I=
 think it's certainly powerful enough<br>&gt=3B &gt=3B to use it for our pu=
rposes=2C certainly for all the standard logic. <br>&gt=3B &gt=3B I do have=
 a question however on producing solutions and alternatives. The only recur=
sion <br>&gt=3B &gt=3B example I found in Powerloom used transitivity rules=
 (parent/ancestor example and the RCC8 KB).<br>&gt=3B &gt=3B Unfortunately=
=2C my problem cannot be used with transitivity.<br>&gt=3B <br>&gt=3B &gt=
=3B The problem I have is:<br>&gt=3B <br>&gt=3B &gt=3B So that's the questi=
on basically. Can powerloom produce a list of entities that together form t=
he solution=2C<br>&gt=3B &gt=3B based on some constraints that are set on t=
he production of this list and present each acceptable=2C but minimal <br>&=
gt=3B &gt=3B solution individually?  I'm not asking for a piece of code=2C =
just what I need to look for to realize this.<br>&gt=3B <br>&gt=3B &gt=3B R=
gds=2C<br>&gt=3B <br>&gt=3B G&gt=3B <br>&gt=3B  		 	   		  <br>&gt=3B &gt=
=3B _________________________________________________________________<br>&g=
t=3B &gt=3B Hotmail is redefining busy with tools for the New Busy. Get mor=
e from your inbox.<br>&gt=3B &gt=3B http://www.windowslive.com/campaign/the=
newbusy?ocid=3DPID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_2<br> 		 	   	=
	  <br /><hr />The New Busy is not the old busy. Search=2C chat and e-mail =
from your inbox. <a href=3D'http://www.windowslive.com/campaign/thenewbusy?=
ocid=3DPID27925::T:WLMTAGL:ON:WL:en-US:WM_HMP:032010_3' target=3D'_new'>Get=
 started.</a></body>
</html>=

--_ec82223d-ab6a-404a-a84f-fe6e2b8e83d5_--

--===============1944975865==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
powerloom-forum mailing list
[email protected]
http://mailman.isi.edu/mailman/listinfo/powerloom-forum

--===============1944975865==--