|
|
 | | From: | Sascha Herpers | | Subject: | usability of imported functions in ampl | | Date: | Thu, 13 Jan 2005 15:05:17 +0100 |
|
|
 | Hi,
currently I try to model a problem in ampl using imported functions.
What I'm trying to do is, add a constraint to a clsp (capacitated lot
sizing problem).
The new constraint it of the form
myfunc(x1, x2, x3, ..., xn) <= targetvalue;
That is, myfunc takes the lotsizes and return a value, which is suppost
to be no greater than targetvalue.
Is that possible?
What I am a little worried about is, how can ampl (or cplex) optimize a
model with a function of which it does not know how it reacts when given
a set of input parameters? The return of myfunc could be stochastic. (It
isn't in my case, but it could be.)
Regards,
Sascha
|
|
|
