Re: Basic Question

Pramit 'Jake' Sarma (psarma@che.iitb.ernet.in)
Sat, 26 Jun 1999 15:35:06 +0200 (MET DST)

The distinguishing feature of Fuzzy Rules are that they are based on Fuzzy
Linguistic Variables (FLV). Take for example, the crisp variable
Temp[erature] - typically the FLV is generated by a Fuzzy Grid/Partition
(or Quantisation) (FG) which is

(A) often smaller than the crisp partitions inherent in a crisp rule, eg.
with T crisp:
IF (T > T_min1) AND (T < T_max1) THEN (Current = 2.5 V): an action valid
for the non-overlapped interval [T_min1,T_max1];

(B) The FLV is almost always overlap-partitioned into fuzzy term-sets.
of an intuitive *linguistic* nature such as {..., ME[dium],HI[gh]}.
Also missing in crisp rule-bases is the "Linguistic Hedge" or the
possibility to modify these terms eg. Very HI = HI^2, or mu_HI(T)^2, again
highly intuitive.

(C) Then the fuzzy inference engine (FIE) processes all the inputs using
fuzzy logical operations which *appear* similar to the crisp IE, but have
a more compact [fuzzy] rule-base (FRB). Modifications to the FRB are easy
to do, and simple to understand, far oftener than the Crisp RB (CRB).

(D) Then it has been shown that a fuzzy model structure has the property
of a Universal Approximator (UA) : given a smooth function F(T), one can
always construct a fuzzy model F_fuz(T) ~~ F(T), for all T in
[T_min,T_max]. This property does not exist for a CRB model.

There exist good GUI's that directly generate a Fuzzy Model [in C++ say]
to test in a real/simulated case.

In Fuzzy Logic Control (FLC), which is my area of interest here, one does
not actually look at the FRB of an FLC to compare with a crisp expert
system. The "expert system" role, in fact the "logic" part of the fuzzy is
suppressed in FLC - which is treated as an input/output *controller*
block typically replacing a [linear] PID controller. The UA property makes
the FLC a *nonlinear* controller, in general.

Pramit

On Thu, 17 Jun 1999, Peter Hirsch wrote:

> I understand that fuzzy logic allows the implementation of rules to control
> a situation. (I refer to Fuzzy Tech's application note about a boiler
> temperature control system for a home heating system in which 405 rules were
> identified. (www.fuzzytech.com/e_a_esa.htm) I also have read that this sort
> of rule based logic is difficult to implement with conventional logic and
> code. Could somebody explain this a little bit more? I don't understand
> where the difficulty lies.
>
>
>
> ############################################################################
> 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
>

============================================================================
Pramit "Jake" Sarma
[Home] [IIT-B]

e-mail: jake_n_jazz@yahoo.com e-mail: psarma@che.iitb.ernet.in

Process Systems and Control
The PROCISS Group
|
{Mathematics/Physics}-<Applied NonLinear Control>-{Industrial World}
|
Process Control, Identification & Simulation Systems
============================================================================

############################################################################
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