From: earlcox@earlcoxreports.com   
      
   Thank God, Bill! I thought it was just me.   
      
   After Richard's hand slapping, I've been reading more and commenting less,   
   but this problem definition is very confusing. To begin with, of course,   
   "about 1" should be a fuzzy number centered around "1", thus the membership   
   function should probably start at "0" and extend across "1" to "2" giving   
   About(1) = bellcurve{1,1), where the first parameter is the center and the   
   second parameter is the expectancy or width of the line from the center to   
   the zero membership point (see The Fuzzy Systems Handbook). Anyway, I'm   
   rambling.  The rest of the problem statement is very odd. I hope this brief   
   discussion will encourage bizon to (1) write a paragraph or two explaining   
   what he is actually trying to do and then (2) reformulate his problem so it   
   make a bit more sense.   
      
   Earl   
      
      
   "William Siler"  wrote in message   
   news:49b9df3d.0404301301.3e1da12a@posting.google.com...   
   > "bizon"  wrote in message   
   news:...   
   >   
   >   
   > I have no idea what you are trying to do here. None of this makes much   
   > sense to me.   
   >   
   > > I have some troubles with understanding one thing. Suppose we have:   
   > >   
   > > fuzzy set:   
   > >   
   > > "about 1" [1 1.01 1.02...2]   
   >   
   > What in the world does that mean? Ordinarily, "about 1" would be a   
   > fuzzy number.   
   >   
   > > three input membership functions defined as:   
   > > A1:   
   > > x         1     1.5     2   
   > > u(x)    1        0.5    0   
   > >   
   > > A2:   
   > > x         1     2     3   
   > > u(x)    0      1     0   
   > >   
   > > A3:   
   > > x         2     2.5     3   
   > > u(x)    0        0.5    1   
   >   
   > Why not write these as   
   >   
   >  A1:   
   >  x         1     2   
   >  u(x)    1       0   
   >   
   >  A2:   
   >  x         1     2     3   
   >  u(x)    0      1     0   
   >   
   >  A3:   
   >  x         2     3   
   >  u(x)    0       1   
   >   
   > Those are easier to understanc.   
   >   
   > > three output membership functions defined as:   
   > >   
   > > B1:   
   > >   
   > > y         1     2.5     4   
   > > u(y)    1        0.5    0   
   > >   
   > > B2:   
   > >   
   > > y         1     4     9   
   > > u(y)    0      1     0   
   > >   
   > > B3:   
   > >   
   > > y         4     6.5     9   
   > > u(y)    0        0.5    1   
   >   
   > I will rewrite these as   
   >   
   > > B1:   
   > >   
   > > y         1     4   
   > > u(y)    1       0   
   > >   
   > > B2:   
   > >   
   > > y         1     4     9   
   > > u(y)    0      1     0   
   > >   
   > > B3:   
   > >   
   > > y         4     9   
   > > u(y)    0       1   
   > >   
   >   
   >   
   > > and one of the rules is:   
   > >   
   > > if (x=="okolo 1") then y="okolo 1"   
   >   
   > Where did this rule come from?   
   >   
   > > let x=1.5 so u(x)=0.5 according with membership function A1   
   > >   
   > > let implication Mamdani is: u(x,y)=min(u(x),u(y))   
   > >   
   > > so how evaluate value of u(y) when x=1.5??   
   > >   
   > > Please about any help,   
   >   
   > It is impossible to help without knowing what you are trying to do. We   
   > have several apparently unconnected things. First, there is a   
   > "definition"?? of about 1. Next we have some membership functions,   
   > apparently unrelated to the "about 1" concept. Then we have a rule,   
   > apparently unrelated to anything else. Finally, there is a definition   
   > of "Mamdani inplication" which has nothing to do with the rule.   
   >   
   > Please say what you are trying to do.   
   >   
   > William Siler   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)