Decisionnal system with crisp outputs

From: Herve Postec (hpostec@club-internet.fr)
Date: Sat Feb 23 2002 - 21:04:30 MET

  • Next message: Masoud Nikravesh: "BISC: FLINT 2001 Now is Online (Both Text and Lectures)"

    Hello

    My problem is : how to defuzzyfy in the context of a decisionnal system with
    crisp output set ?

    2 (simple) examples :

    1/ An alarm system for a factory which gives a signal (say a light switching
    on for instance) when the work load (WL specified in time units) becomes
    "inferior" to the available time (AT) to realize the work.

    I consider an alarm level (AL) which is a fuzzy value depending on WL and
    AT.

    I realized a (very) little set of fuzzy rules which modelizes : AL = f (WK,
    AT) :

    IF WK is high AND AT is low THEN AL is high
    >...

    The resulting three dimensionnal curve is rather near the result I aim at.

    BUT the question is : for which value of AL do I switch the light on or of ?

    With a simple test like : AL < N => switch off and AL > N => switch on, I'm
    afraid I loose interest of fuzzy inference.

    Is there a graceful way to solve this problem ?

    Do I raise the problem in a wrong way ?

    2/ Same situation with a mushroom identification system.

    Operator inputs a set of characteristics and the system delivers the list
    of "best" matching mushrooms.

    The ouput of the inference process is a fuzzy set of mushrooms.

    In this set, some mushrooms have "high" membership value and other have low.

    I'd like to filter the solution set in order to keep the really "best"
    matching elements

    Is it the right way to get an alpha-cut of the solution set ?

    I'll be insterested by any answer to theses questions or links to any source
    describing examples

    I apologize for my bad english.

    Thanks a lot.

    --
    

    Hervé POSTEC hpostec@club-internet.fr

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



    This archive was generated by hypermail 2b30 : Sat Feb 23 2002 - 21:30:34 MET