Newsgroups: comp.ai.fuzzy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!purdue!news.bu.edu!bloom-beacon.mit.edu!news.mathworks.com!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!su-news-hub1.bbnplanet.com!arclight.uoregon.edu!news.ibm.net.il!news.biu.ac.il!news.huji.ac.il!wisipc.weizmann.ac.il!marganit!moddy
From: moddy@marganit.weizmann.ac.il (Teeni Moddy)
Subject: Re: fuzzy state machines?
Message-ID: <1996Dec10.115930.4350@wisipc.weizmann.ac.il>
Sender: news@wisipc.weizmann.ac.il (News User)
Organization: Weizmann Institute of Science, Computation Center.
X-Newsreader: TIN [version 1.2 PL0]
References: <32A83BF0.267E@metasphere.com> <19961207152400.KAA15875@ladder01.news.aol.com>
Date: Tue, 10 Dec 1996 11:59:30 GMT
Lines: 23

wsiler@aol.com wrote:

: It has been a long time since I worked with finite state automata, and I'm
: not familiar with the Moore machine nor with the "guard" feature. I'm more
: used to defining a finite state automaton as the tuple (input(s), current
: state, next state, outputs). I would think that the fuzzy counterpart of
: this would be (input(s), current state probabilities, next state
: probabilities, output probabilities). If the inputs are also fuzzy, we
: have (input probabilities, current state probabilities, next state
: probabilities, output probabilities). If you like, substitute
: possibilities or grades of membership for probabilities.

This is quite a problem, I think. If you use probabilities, you have natural
semantics, in which the state of the machine is its probability to be in any
distinct set, and the sum of all those probabilities would be 1.

Using fuzzy sets ( as you told me ) can have many interpretations as to what
the machine does to the membership grade of the 'old' state and of the 'new'
state.

In your case, I would prefare possibility to fuzzy.

-Moddy Teeni
