Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!newsfeed.rice.edu!bcm.tmc.edu!news.msfc.nasa.gov!newsfeed.internetmci.com!EU.net!uknet!dcl-cs!bath.ac.uk!enssep
From: enssep@bath.ac.uk (S E Potter)
Subject: Configuration problems and unknown values
Message-ID: <DGL60M.DvL.B.mary@bath.ac.uk>
Organization: School of Mechanical Engineering, University of Bath, UK
Date: Tue, 17 Oct 1995 09:21:54 GMT
Lines: 22

Hello,
I'm working on trying to implement a system for the configuration
of hydraulics circuits using ANNs; my intention is for a network
to translate user specifications for the functionality of such a
circuit into the actual components required to achieve it.

However, the set of possible specification inputs is large and not
all will be of interest to a user for a particular problem; is there
any way of representing these `don't care' inputs in such a manner 
so as not to hinder the network's prediction ability?


Also, if anyone knows of any related work, using ANNs for 
configuration problems, I would be grateful to hear about it....


Thanks,

Steve Potter

University of Bath, UK

