ffll-devel

[Ffll-devel] Possible example?

[chen <ch...@me...>]

Hi,

It is really nice to see the FFLL class.
but did someone has an example project for showing how to apply the FFLL
calss into
a uesful project, such as a MIMO fuzzy logic system can approximation a
nonliear
function/system or it can make some control tasks online or offline ?

Thanks in advance!

chen

Re: [Ffll-devel] Nested Rules Possible ?

[Michael Z. <Mic...@Lo...>]

I can't see anything in the FCL standard that prevents these type of =
nested rules, but FFLL doesn't deal with them

for things like:
=20
>> IF hydrogen IS high AND methane IS not_high THEN state IS overload;
>>  IF state IS overload AND filling_level IS full THEN feed IS stop;

I would create two FFLL models. One that has 'state' as output and =
another that uses that value as an input. In the example above, 2 models =
would be needed as the output variables are different ('state' and =
'feed')

Michael Z

AW: [Ffll-devel] Nested Rules Possible ?

[Ernst M. <mu...@aw...>]

Hello Michael,

if I use two models, the output has to be defuzzified and fuzzified again?

I need that the outptut variables of model 1 can be used without
defuzzification/fuzzification before they are used as input for model 2.

Greetings, Ernst

Re: [Ffll-devel] Nested Rules Possible ?

[Michael Z. <Mic...@Lo...>]

Yes and no... you don't have to "fuzzify" it, you can use singletons as =
the terms of the variable. So using your example:

IF hydrogen IS high AND methane IS not_high THEN state IS overload;
 IF state IS overload AND filling_level IS full THEN feed IS stop;

One of the input variables to the 2nd model would have singleton terms =
like "overload".=20

Michael Z
  ----- Original Message -----=20
  From: Ernst Murnleitner=20
  To: 'Michael Zarozinski' ; ffl...@li...=20
  Sent: Thursday, February 20, 2003 2:15 PM
  Subject: AW: [Ffll-devel] Nested Rules Possible ?


  Hello Michael,

  if I use two models, the output has to be defuzzified and fuzzified =
again?

  I need that the outptut variables of model 1 can be used without =
defuzzification/fuzzification before they are used as input for model 2.

  Greetings, Ernst