Re: Bivalent vs Fuzzy

Pete Moss (uhjg9787r56ser65@ix.netcom.com)
Sat, 28 Mar 1998 22:06:20 +0100 (MET)

Dan Samber wrote:
>
> Hi!
>
> I'm currently BURIED under fuzzy logic books & articles that
> tell me of all the PHILOSOPHY behind fuzzy logic and even how
> to use it... all of which is great.
> What I'm looking for is (ideally) SEVERAL examples of how fuzzy
> logic is used in real applications AS COMPARED AND CONTRASTED WITH
> TRADITIONAL SOLUTIONS. Something that would hold my hand & walk me
> through the steps of design using both bivalent and fuzzy logic
> & demonstrate how fuzzy logic can yield benefits.
> I'm surprised that this is so difficult to find. I have both Kosko's
> and McNeil's books.... but I feel I MUST have this kind of compare &
> contrast presentation to REALLY get the "AHA!".
>

I've also been looking for something like that for several years. Everything
I've seen (so far) could be handled quite nicely with Bayesian probability,
decision theory, or control theory. Perhaps my mind has just been poisoned with
too much Western empiricism. I never even knew "Western" was a dirty word.
I've read the Tao Te Ching a dozen times, and freely confess I find a lot more
in common between David Hume and Lao Tsu than between the latter and the fuzzicists.
I haven't given up though. If there's something to get, maybe someday I'll
get it.

While we're searching, take a peek at another generalization of yes/no logic
to the interval zero/one.

http://omega.albany.edu:8008/JaynesBook.html

That's the happiest find I've come across in a long time! It's a book that
E. T. Jaynes has been working on, apparently, for forty years. In chapter
one it shows how different logic rules with (possibly) uncertain antecedents,
predicates, and conclusions correspond to different applications of the Bayes
theorem. Fun stuff.

Best regards,
Pete