Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!news.duq.edu!newsgate.duke.edu!agate!nntpfeed.doc.ic.ac.uk!sunsite.doc.ic.ac.uk!warwick!niss!bath.ac.uk!dcl-cs!usenet
From: Jonny Carlos da Silva <jonny@comp.lancs.ac.uk>
Subject: Re: Expert System methodologies
Content-Type: text/plain; charset=us-ascii
To: K Flear <100551.1414@CompuServe.COM>
Message-ID: <3300887C.3E54@comp.lancs.ac.uk>
Sender: usenet@comp.lancs.ac.uk
Nntp-Posting-Host: engc45.comp.lancs.ac.uk
Content-Transfer-Encoding: 7bit
Organization: EDC- Lancaster University
References: <5dkpfn$6vq$1@mhadf.production.compuserve.com>
Mime-Version: 1.0
Date: Tue, 11 Feb 1997 14:55:56 GMT
X-Mailer: Mozilla 3.0 (Win95; I)
Lines: 94

K Flear wrote:
> 
> I am currently researching methodologies used in the development
> of expert systems, in particular the knowledge elicitation
> process.  I am interested in what, if any, particular
> methodologies are used and which aspects of the methodology prove
> most successful/ unsuccessful - especially -
> 
> which questioning strategies are employed plus issues such as
> 
> Is rapid prototyping used/ found to be useful.
> 
> are Shells  used - usefulness/shortcomings.
> 
> I would be grateful for any feedback on the above.  If there is
> someone who would be better placed to provide feedback, please
> forward this message.
> 
> Thanks very much  for your help.
> 
> 
> Karen Flear
> 
> --
> 

Hi Karen,

Presently, I am working at the Engineering Design Centre- Lancaster
University, as a Visiting Scholar from the Federal University of Santa
Catarina, Brazil, in which I am also lecturer at Mech. Eng.
Department. 

This work corresponds to my doctorate degree thesis, which consits of
the development of an Expert System for Fluid Power Systems Design
focusing in the Concurrent Engineering Aspects.

The system is being developed in CLIPS, it has several modules, and it
is running in Windows (CLIPS 6.0, standard interface) and UNIX (through
a MetaCard Interface).

Regarding your question about the shells usefulness/shortcomings, I have
worked with some shells (CLIPS and KAPPA-PC) and I can see that without
them that expert system development process would be greatly enlarged.
Because features such as, Inference Engine, Knowledge Representation,
Forward/Backward Chaining, etc. are already implemented in the shells.
Therefore, once a shell choice is made, the knowledge engineering can
focus more on the domain issues, rather than in the implementation
details.

Concerning the rapid prototype aspect, I have been using the incremental
procedure in my research for it helps to clarify as early as possible
the key points in the development process, such as:
-How are the user inputs defined?
-How can the system be maintained and expanded?
-Is the chosen knowledge representation technique really suitable for
the specific domain? 
-It helps to explain in a better way the expert system advantages and
objectives to the domain expert, who not always have this background.

I hope these aspect help your research.
Following, there are some good references in AI.

Giarratano, Joseph and Riley Gary
Expert Systems Principles and Programming
1994, PWS Publishing Company
ISBN 0-534-93744-6

Gonzalez, Avelino J. and Dankel, Douglas D.
The Engineering of Knowledge-Based Systems
Theory and Practice
1993, Prentice-Hall, Inc..
ISBN 0-13-276940-9

Waterman, D.A.- 
A Guide to Expert Systems, 
Addison-Wesley Publishing Company, 1986.

Rich, E. and Knight, K.
Artificial Intelligence, 
McGraw-Hill Book Company, 1991.

I have just started the process of interacting of the industrial
experts, and the knowledge elicitation is NOW MY KEY ISSUE. I am
thinking in using questionary plus classical interviews. If you have any
suggestion, comment or question about this work feel free to contact me.

Cheers,

Jonny

Engineering Desing Centre- Lancaster University- Lancaster LA1 4YR- UK

Homepage http://www.grante.ufsc.br/jonny/jonny.html
