Re: Basic Question

Pramit 'Jake' Sarma (psarma@che.iitb.ernet.in)
Thu, 1 Jul 1999 04:54:57 +0200 (MET DST)

I am not sure if there is all that much info in books, at present.
There are N ways of defuzzification, inference modelling, etc. within
the same sub-structure. I have never seen any reasoning behind why
any one chose one structure over the other, other than "it worked".
No why's, usually not even simulations.

As far as the Mamdani-type (M) or the Takagi-Sugeno (TS), the
differences are broader, of course. Both follow the Universal Function
Approximator (UFM), but that is a non-specific test, and will not
tell you which is better. Broadly, the output is fuzzy for M, *then*
defuzzed, while in TS it is directly crisp, usually via a linear
combi map. The TS structure, then, is a specialisation (more or less)
of the M. In general, ceteris paribus, the M should do better to equal.
The principal advantage of TS is the ease of incorporating (usu. linear)
crisp mathematical modelling tricks, such as LTI state-space models,
parameter-estimation, and so on.

Full-fledged "model-free" or linguistic estimation in M is *lost* partly
in TS. Or, roughly, [M] - [TS] - crisp, if anyone cares to parametrise
this meta-structure!

On Tue, 29 Jun 1999, Romero Ardoy Ravier wrote:

> >>>>>>>>>>>>>>>>>>>>>>>>>>>.
> Mamdani type FLC can be model free, but difficult to optimize, TS type
> FLC is somewhat not model free, but easy to automatically optimize.
> >>>>>>>>>>>>>>>>>>>>>>>>>>
>
> Hi everybody. Mr Song wrote this sentence in an email responding a
> basic question, and as I see it, I kind of feel like raising a simple
> question myself. I know the, let's say, structural differences between the
> Mamdani type of controller and the Takagi-Sugeno type, but I would like to
> know something else, of the sort of the differences in performanc;, why,
> when and where should we use one instead of the other; the limitations they
> present; the advantages, etc... The books I have read haven't told me much
> about it.
>
> Thanks, best regards,
>
> Javi.
>
>
> ############################################################################
> 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