Re: A simple question

David A. Fuess (fuess@llnl.gov)
Sat, 6 Jun 1998 21:36:25 +0200 (MET DST)

Will Dwinnell <76743.1740@CompuServe.COM> wrote:

>"1. What is the big advantage in using fuzzy logic to build a
>controller, ..."
>
>Keep in mind that fuzzy logic is not one technique: there are
>fuzzy regressions, fuzzy clustering systems and so on. My
>understanding is that fuzzy controllers are usually fuzzy rule
>sets arranged in a matrix- really a look-up table with
>interpolation. The advantage usually touted for fuzzy rule

A FLC can be reduced to a lookup table over a response surface, but so
can a PID with appropriate granulation of inputs. This is a technique
I've sometimes applied in implementing a FLC on a small
microcontroller where the processor is slow but the desired response
is fast. In fact, I don't even look up the result, I calculate it's
address! However, in general FLCs do active rule interpretation thus
allowing for the possibility of updating the rules after the
controller is installed or even determining the rules in real-time
through one of the rule generating techniques.

>systems is that they are easier to construct and easier to
>understand. I suspect that this is more true for fuzzy rule
>bases not confined to the giant matrix format.
>
>"...and
>2. what fuzzy logic controlers can do that a Neural Net can't?"
>
>Assuming a sufficiently broad selection of each, theoretically
>nothing.

The original question needs some attention I think. All too often we
seem to get hung up on "can" some technique do this and such? The real
question should be, which technique is most appropriate given the a
priori knowledge about the process, the controller, available data,
maintenance requirements, developer experience, etc. Yes, both
techniques, along with a host of others, can do the same task. But
which one should do it in this instance. (Related question: If you can
dig a trench with a spoon, why would you use a back hoe?)

############################################################################
This message was posted through the fuzzy mailing list.
(1) To subscribe to this mailing list, send a message body of
"SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at
(2) To unsubscribe from this mailing list, send a message body of
"UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com"
to listproc@dbai.tuwien.ac.at
(3) To reach the human who maintains the list, send mail to
fuzzy-owner@dbai.tuwien.ac.at
(4) WWW access and other information on Fuzzy Sets and Logic see
http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info
(5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html