Re: learning algorithm for tuning fuzzy rules based on the gradient descent method

[EMAIL PROTECTED] (Brahim HAMADICHAREF) wrote in message news:<[EMAIL PROTECTED]>...
> Hi,
> 
> I do not have this paper but if you Google for "gradient
> descent method" you will find information about how this
> technique is used in other domains. Searching the IEEE
> Explorer it returned me IEEE Transaction papers.
> 
> I think that you should use a fraction of this error to modify
> the rules (usually definition of the fuzzy sets of your system)
> and this until this error get mimized (either limiting the
> number of iterations or relative error from one iteration
> to another very very small). That should be the basic concept.
> 
> Maybe this link can help you to understand better the way
> this technique can be used.
> http://www-math.cudenver.edu/~aknyazev/teaching/98/4660/fp/r2/
> 
> Hope this help.

> Brahim
> 
> > assume the input is X = [ 12,13,14,15]
> > and estimated output is y =[12.345, 13.215, 14.871, 15.654]
> > we know that the error occoured is the difference 
> > So 
> > we have,
> >    Y' = [0.345,0.215,0.871,0.654]
> > how do we go about doing this  using the above?
> > 
> > any help with this will be appreciated
> > thanks.

The idea of a Google search is excellent.

The gradient method, like most optimization techniques, tries to find
those values of a set of parameters that will minimize an error
function, starting from a set of initial guesses and using first
derivatives of the error function. In your case the parameters would
be those defining a set of parameterized membership functions. The
error term could be, for example, the sum of squares of differences
between actual and desired outputs. (I don't understand why you say
the output should equal the input. If this is the case, why would you
need a controller at all?) For computation, you will need a model of
the process being controlled to determine actual output. You should
probably use initially a simple setpoint model.

I don't understand why you say the output should equal the input. If
this is the case, why do you need a controller at all?

William Siler